Recently a Chinese media interviewed me and I talked about a few side projects I have done in the past. I talked about the Neuralrad Mammo Screening project and Neuralrad multiple brain mets SRS platform. More awareness on radiation therapy to the general public will greatly help the community and we believe Stereotactic Radiosurgery (SRS) will eventually replace majority of the whole brain radiation therapy (WBRT) in the next five years.
Super effort. I understand your reluctance to accept funding but if you ever change your mind on that be sure to publish it here on HN. If giving you more tools means more progress in this domain without the usual red tape then I'm all for giving you as much of a push as possible.
What's the best way to contact you? I've been fighting brain cancer for 7 years and work closely with a group of neuro-oncologists, researchers, non-profits, and investors in the space.
1. do you have a patreon or something where people can support you?
2. do you know how to obtain the nucleotide sequences of c-MYC?
3. have you seen machine learning applications that can predict protein function? specifically, the goal is to predict if protein X could participate in cleaving of DNA similar to CRISPR Cas9? false positives are okay, provided the negatives are accurate (i.e., guarantees protein cannot cleave DNA like Cas9).
thanks so much! would it be possible to contact you for other questions? or could you recommend a reddit sub for questions like this? researchgate doesn't seem responsive.
for instance, how to obtain the nucleotide sequences for TP53, or for the NHN/RuvC domains of CRISPR Cas9?
re other questions:
The crucial thing is: what are your goals? Bioinformatics and cancer biology in particular are rather vast fields. Questions and answers make sense if you can place what you just got in some framework. And getting there is non-trivial, see i.e. +800pp book:
https://www.amazon.com/Lewins-GENES-XII-Jocelyn-Krebs/dp/128...
is it possible to find nucleotide sequences from amplified c-MYC from cancerous tissue, either burkitt's lymphoma or TNBC?
ensembl is valuable, but it would be more efficient to hire someone to decode the different options. what type of researchers understand the different sequence options on an ensembl page: cancer biologist, geneticist, bioinformatics scientist?
thanks again for your help!
the goal is conducting some first-principles analysis of cancer, nothing fancy and very likely a waste of time.
WBRT is pretty brutal. Am I right in thinking you are focusing on multiple site treatment/palliative treatment of metastatic presentations? High site count also or sticking to say < 5?
Exactly, I'm working on a workflow platform for multiple brain metastasis stereotactic radiosurgery. This will greatly benefit patients with more than 5 BMs.
Metastatic disease is when a cancer spreads to multiple locations. This can make it difficult/impossible/impractical to treat effectively, especially surgically.
Whole brain radiotherapy works by killing everything a little bit in the hopes that the tumors die first (e.g. like chemo). There are good reasonswhy this tends to mostly-sort-of be true, but getting the balance right is hard and too much dose will definitely cause other problems.
SRS is a way of targeting radiation directly to locations to kill cells, with less effective dose (hence damage) to other parts of the brain.
It's all pretty harsh stuff, and you can die from the necrotic tissue caused by it, also.
Often with this kind of disease you know you aren't going to cure someone, but you can get rid of symptoms and make people more comfortable (palliative care).
Where did you learn to program on distributed Nvidia GPUs? The article implied you were self taught and learning to do this is quite challenging for various reasons.
Not least, Nvidia's documentation is not the best resource to learn from. This seems like quite a lot of work to understand ML and write custom CUDA code to get this to work. Do you have any insight about how you taught yourself these things and what tools you use?
Not the OP but I taught myself in a couple of weeks picking apart some of the sample CUDA code and reading some of the (excellent) pdfs on the architecture of the Nvidia range. At the time the GTX285 was hot stuff, the same code runs unchanged on a 1080ti and I would expect it to continue to work on even more modern incarnations. CUDA is pretty good as a platform to build on if you understand the basic idea behind the engine, and some low level experience will come in very handy. And ML on CUDA vs ML in C or some other language is typically a matter of shuttling the data and the results back and forth between main memory and the card as well as implementing the most time consuming portions of the algorithm you are using in a custom kernel, you can usually get 50% or so of the theoretical maximum speed with relatively little effort. Getting to full speed is going to be a lot harder, but then you could of course also add another card (or another three) get get an instant boost.
Usually you would - nowadays at least - use someone else's optimized kernel + ML library but if you wanted to roll your own that's doable.
I'm not your OP but I learnt all these things at univeristy during my BCompSci. Understanding ML algorithms came down to a lot of math / statistics units. I learnt about parallel computing during a dedicated unit called "Distributed and Parallel Computing"
Imagine for a moment if FB and Google took all their developers and resources and applied it to solving cancer, how long do you think it would take to make a dent in the problem space? Or is there a hard limit stopping us from progressive like Moores law for scientific understanding.
> Imagine for a moment if FB and Google took all their developers and resources...
It's a nice thought exercise, but of course not realistic. FAANGs wouldn't be able to attract or retain the talent if they were purely in scientific medical research (where do you think those fancy offices and fat paychecks comes from?).
Having said that, it doesn't have to be either/or. I strongly believe that companies can be successful in balancing between thriving commercial businesses and using its profits to advance research that can benefit the society at large. Case in point: Google's DeepMind has for years published groundbreaking work in various areas including breast cancer research [1], protein folding [2], energy savings and climate change [3], and more. Also Calico [4], or its investments in health startups via GV, and of course all the well known services to users.
What I'd like to see is more companies following a path where a portion of its profits is used to fund research in areas that can cause societal change for the better (which, in its turn, may eventually became new sources of revenue down the road - e.g. protein folding, self-driving cars).
(disclaimer: Googler here, but no relationship with any of the projects mentioned)
Sir, We've all been hearing that AI will revolutionize medical diagnostics for a long time, seen startups come & go in this space but seeing your work impact lives on the ground has convinced me that the secret missing ingredient was a self-less human who can build an accessible(non-commercial) solution for the masses.
A decade ago I did a voluntary work on a simple app for oral cancer detection(questionnaire, data-entry) with an oncologist, Who used it to for survey in the tribal regions of India and he used to say how lack of early detection is the number one reason for so many deaths(many die without knowing that they had cancer).
He's now settled in Germany, But I would still pass this story to him & perhaps his colleagues in India could make use of it.
Hi. Some great projects. What's more commendable is your dedication towards your projects and seeing them through to end- to the point that they are actually useful. This is what I truly admire.
I have a question for you. What is the tech stack that you use?
And if it is not too much: What resources did you use to learn Deep Learning?
> Recently a Chinese media interviewed me and I talked about a few side projects I have done in the past.
I apologize if this has been asked and answered before, but do you speak Mandarin, or was the interview in English?
Asking out of curiosity if it’s the former, and if so, how difficult was it to learn whilst also working on this and other things? And are there any resources or tips you might share that you found helpful?
According to the article, Coolwulf went to Nanjing University for undergrad, so pretty safe to assume he would do an interview with Chinese media in Mandarin. And since he likely grew up in China, it was probably very easy for him to learn Mandarin indeed!
Great work. I'm inspired seeing others use their skills to work on something important. So many of the smartest, most well educated engineers of my generation have put their talents to use doing things that are of little value to society. Thank you for demonstrating to us what else is possible.
Unless it's more expensive than existing treatments the medical industry will close the circles around you excluding you.
That's why not one startup has hacked healthcare in America, not one. No breakaway successes making pharma cheaper. Like those incubators in Bangladesh, for premature babies not startups that is, those did OK. Some pill startups yes, but again that's an expensivification of medicine. If you can make medicine more expensive, they welcome you in!
Jim Clark tried this, he was on a roll after Silicon Graphics and Netscape. Huge roll about as strong as Elon Musk as a serial entrepreneur. Then he targeted healthcare and couldn't do shit, just couldn't get anything to happen. He literally talked about getting "rid of all the assholes" by which he meant insurance and doctors and hospitals and middlemen and pharma and all the other "assholes" of that nature in his own words, but leave "only one asshole in the middle--us [paraphrased]." It's in a book. That book also talks about guys going on airplanes and chasing goats off cliffs, saying "Some people do this."
Well the real structure of medicine isn't designed around the human body, it's designed around cornering the market. Market dominance. So of course it has this immune system against cost reduction and efficiencies--efficiencies especially--and you do know it lobbies, don't you? And can bribe the FDA like the Sacklers did? Or lobby the FDA, and then bribe underneath so when people see favoritism they think it's the over-the-counter placebo causing a placebo effect without suspecting an additional more potent under-the-table dosage of money. In case the administration has built up a tolerance to the over-the-counter stuff.
>That's why not one startup has hacked healthcare in America, not one.
This post makes a lot of points, but in general I think they boil down to the above statement : the belief that large, complex systems are just run by stupid and/or malicious people and that a sufficiently clever 'hack' will fix all the problems. I think that is an attitude that is common on HN, but wrong.
Most big problems are not technology problems, they are People problems with a capital P. Technology problems can be fixed with 'one simple hack they don't want you to know about!!!' People problems are complex and messy and cause and effect can be intermingled vertically and horizontally with other seemingly unrelated factors as well as temporally with things that don't even exist yet or used to exist but don't anymore!
The way we fix these messy, complex People problems is by respecting that they are real problems, that the people acting on those systems are (mostly) reasonable people just doing what reasonable people do, and slogging through solutions a day at a time with the oldest technology around - political power. These problems resolve if you can get enough people to agree they need solved.
What about the NHS in the UK? They should be more aligned to wanting cheaper cancer diagnosis and also anything to help people.
If America has dysfunctional healthcare there is still the rest
of the world. Which might be good for Americans eventually as the tech will come across one way or another.
> If America has dysfunctional healthcare there is still the rest of the world. Which might be good for Americans eventually as the tech will come across one way or another.
Which raises the question - why does the vast majority of healthcare tech development come from the US? (I included "development" to get around the out-sourcing of testing to China.)
It is probably much much better for the wealthy and well insured, but on average not as good for all people. With that skew the high end is going to be better. That is my guess.
Modern treatments are hugely expensive to develop, and they also tend to be very specific (ie. only 1 in 5,000 people might get the exact right kind of brain tumour for your treatment to be an option).
With only about 1 million all-cause deaths per year in the UK, that means your treatment for a specific terminal brain tumour might only have 200 patients per year.
The 'we just saved 4 hours of clinician time by using fancy AI' just isn't worth it if it only saves 800 hours of clinician time per year, yet costs millions to develop.
The fix is to roll this out somewhere with more patients (eg. China) and where trials are cheaper (ie. China).
Use rugged AI like coolwulf. That's what's up. He hauled ass, be like coolwulf.
No don't be like coolwulf, be coolwulf. Be him, do what he did, make the algorithms from your brain protect the algorithms that make up your brain. You know they're symbiotic right?
Transistor and neuron fighting together against destruction. Symbiosis.
EDIT: coolwulf might as well be John von Neumann, the last line of defense against brain cancer and brain surgery, those neurons reaching beyond the skull to make a change in the World for some part of that World to then reach inside that skull and protect it. The chip as the final garrison of the biological human brain. Ghost in the Shell. Ghost retreating from the Shell and regrouping, no ruling out whether it can triumph when it returns to reconquer its damaged Shell.
Money moves too many times, like the pebbles under the shells. So that's a deflection so people don't see who's serving whom. Every time someone thinks they can remember what's under a shell before it's shuffled, they move it really fast, and like with feints. And of course this doesn't work for real, hands aren't fast enough, so you gotta layer it with like something to injure their eye with a flash while they're looking, say a laser to the pupil, or maybe...
You could make a little trapdoor for the pebble provided the guy will not ask which shell the pebble was actually in. So deflect all questions.
But the end outcome is servitude comes from all over the world to an American banker, an American doctor, and an American executive. There's a few others. But the long tail of the American middle class? That was only necessary because those bankers, doctors and executives would never themselves (or with their sons) man the nuclear ballistic submarines, dogfight jets, aircraft carriers, CIA undercover agents, and especially infantrymen--never never never infantrymen. I've only heard of a single lonely guy becoming an enlisted man in a high-bullets-flying environment, when he could have gone anywhere--from Philips Exeter in fact. Anecdotal.
Yeah undercover agents get wrapped up and tortured, but infantrymen are bullet sponges. Riflemen, holy shit. WW2 riflemen, no let's not talk about that. At least spies there's a whole song and dance to the whole thing, part two of 1984, and speaking as someone who's stood up to torture, there's not the immediacy of the infantry where the brain surgery is much messier. Well on second thought it's not that different from getting rocks thrown at you, like I have this year...
Listen. So this is what's up. There's these guys, they eat all this work and don't do anything. Doctors don't do anything. They only use 5% of what they learned, the rest being a masochistic filter the elders imposed on the youngers to keep them from bringing them down to human reality with competition. So morally a doctor should only study for 1 year, not 20 (if not less, because of retention). That's five percent! Don't argue with that part, argue with something else!
And then plus surgeons, spesh brain surgeons which is what every medical student "wants to be when he grows up" as elder doctors put it, so the thing there is younger surgeons have much better manual dexterity than older surgeons, you can't do it past your fifties generally. What does that mean? Learn it faster! But no, they have to say "I studied for 20 years to tell you my bullshit" like come on, that's masochism. You spent 1 year learning worthwhile things and 19 years in the medical school "slammer" learning to be a good bitch, that's the whole game in medicine, hazing. Because otherwise people would say hey just pump out more doctors then the doctors have to say "nonono, you can't because it's soooooo expensive to make a doctor and America loses money [pretends to lose money] on training doctors and you have to be the absolute best of the best of the best to learn how to multiply 3-digit numbers".
It's so so stupid, especially because of this: it's all multiple choice. All of it. It's all multiple choice. No essays or explanations, just filling in black circles. I have contempt for using that form of testing alone because I'm so good at it with no effort. Like not even doing the reading. Show up "naked", a good enough mathematician should be able to learn from the test itself, from NOTHING, NOWHERE, and NOBODY. Destroy the integrity of these measurements. Get the test to leak the answers. Like I could do it in Chinese [I don't read Kanji, I can only make out Oracle Bones from 2600 BC] and be measurably better than a monkey. I don't generally brag about it because I have no respect for it, like not having respect for torture. But if that's the underlying reason some doctor claims to be better than me? If such a doctor says he's better than me, then let us agree we are not equal.
While the US health care system as a whole is mess, I don't think it's intractable if you can attack it from the right directions. GoodRX did with pharmaceuticals, and Mark Cuban seems to be doing a good job with CostPlus Drugs. (Valisure is also pretty interesting, doing in-house validation of generics).
Marty Makary wrote a book a couple years back, The Price We Pay on the wicked knot of a problem that is US health care. For those mildly interested, there was a Peter Attia Podcast inteview a while back that covers the gist of it: https://peterattiamd.com/martymakary/
For those that want to get some color from some of the biggest problems from a clinician/practioner's perpective, I found some of these podcast episodes to be pretty great/eye opening: https://zdoggmd.com/podcasts/
Super impressed. Amateurs can and do play their parts in science --- there are numerous discoveries made in astronomy, mathematics, and definitely in computer applied sciences that sprang from the minds of amateurs.
As a “professional” programmer, I’m humbled by your accomplishments. I really must find ways to contribute more to the world. It seems there’s a lot of opportunities in AI to do it.
In the UK the NHS don't do screening for breast cancer for under 50s because it's believed that it would do more harm than good by leading to unnecessary treatment for cancers that would never have actually caused any harm, and even where no treatment is carried out it causes great distress. Though there are arguments that the age cut off is too high, and should be set at 40.
The above is with regard to a well-funded and regulated screening program that presumably has much better precision/recall than this website. I wonder what the cut off age is for this website before the diagnoses cause more harm than good? 60? 70?
This is getting lots of upvotes because it's confirmation bias for the large segment of HN readers who believe that problems would easily be solved by a small number of brilliant technologists, if only it weren't for governments and big organisations with all their rules and regulations.
A lot of people, including myself, don’t believe that central health authorities have the right to make that call.
Moreover, I personally don’t have confidence in their ability to make those kinds of decisions, and I believe the abysmal performance of the NHS supports my view.
I think central authorities absolutely must make that call. Who else is going to decide how to dole out a scarce social resource? Americanizing healthcare is obviously not a good choice given how much worse it does overall by basically every measure (unless you're rich and don't give a toss about other people).
I certainly agree that central authorities can be better. But that's kind of a truism.
What alternative options do you have in mind? Admittedly I'm short on alternative ideas.
I understand why people don't want money to completely decide who gets health care. But I don't think it should be a complete non-factor either. If I've worked hard and saved diligently for an emergency, and an emergency arrives, I feel I should be able to tell a doctor "will you treat me? I have money".
But if you insist upon perfect equality of health care access, I guess it can't be. Some central authority has to decide if anyone is allowed to treat you, regardless of what you've saved for the eventuality.
Medical tourism will always be a thing for the rich. But I personally do not see an argument for driving a social service, even if only slightly, with capitalism.
Perhaps that’s part of my developing thought on the matter: the public system paid by tax dollars should be equitable to all. But by all means, take your credit card to Pepsi Presents: For Profit Medical Centre and get a full work up.
What makes the cut of a "social service"? Wouldn't it be anything people really need? Basic food staples, for example?
China tried socializing food production, and it worked terribly. Production tripled when they re-privatized it and let farmers grow for themselves. People do a better job when they get to keep the rewards.
US health care isn't capitalist in this sense of rewarding a good job. It's the worst of both worlds: a system whose regulations are superbly adapted to optimize profit for the administrative class at the expense of both doctors and patients.
Some Americans think that the US Healthcare way is “the right way but Slightly Off-Track(tm)” and will not be suaded.
You asked a direct question, though, to what extent do social services exist.
Social services exist to ensure that we all have a decent foundation on which to conduct our business of living.
For some that will be as you say, providing basic food and housing. In fact in Sweden food is given to children for free in school; however in the UK it is only poor students that get it.
In Sweden water is free, in the UK it’s charged but it’s a utility that cannot be turned off.
Everyone draws the line somewhere else, but the basis is meant to be that we have a solid foundation.
Unexpected medical expenses shouldn’t decimate a household economy for a decade, it doesn’t matter how unprepared the household is. But this is my personal feeling.
In addition: tying health insurance to employment and having at-will working conditions strikes me as ensuring compliance/docility in the work place, which I don’t believe in.
1) it's not a shallow dismissal if neither position is interested in having a conversation.
2) it's not unkind to point out the truth, it's not possible to persuade people that the American healthcare system has a fundamental architectural problem, they believe wholeheartedly that it would work if it wasn't for the <insert consequential factor here>
it's not an intellectually honest conversation, it's shallow, everyone has dug in. it's boring.
You've replied to my post, but I honestly cannot tell who you are talking to, or whether your assertions about whoever you're talking about are in any way connected to anything I've said.
It could be that this style of engagement is why you're bored. It does seem boring to talk to people without actually talking to them.
You're talking about "keeping the fruits of labour", in the context of socialised healthcare, however it seems the majority of the world is doing perfectly fine with socialised healthcare. In fact it seems that less tax money is spent on healthcare in countries that just nationalise it.
That's the relation, that's why it's boring, because you'll never truly argue the point or consider another perspective. You can't be persuaded, and neither can I, reasonably.
So, what's the point.
Go live in your squalid hole and I'll go live in mine and lets not pretend that either of us aren't brainwashed in some way, because there's no possible way that we're going to conclude this discussion that I keep seeing repeated on every public forum ad infinitum.
Why bother intruding into a discussion to announce how exhausted you are by it? It means nothing to us, and it's more effort for you. You could just ignore it and be less exhausted.
I'm basically telling you both to shut the fuck up because we're rehashing the same shit as forever.
Unless either of you are willing to engage in intellectually honest, curious conversation then you should honestly just keep your gob shut.
Flame wars help nothing, rehashing this helps nothing. We're walking in circles, save your energy, I'm wasting mine in an effort to save yours in future.
I made a couple points and it seems like you think that, based on those points, you know much more about what I believe than I even know myself, and hold a great deal of contempt for me and whatever it is I apparently believe.
I would love to have an open and inquisitive conversation. I am open-minded about how we structure our system. There are many considerations, and costs and benefits to any method. I do not think socialized or public structures are bad or inferior to private ones.
I do think it helps for people to be rewarded for their labor, but it isn't the only consideration, and it can be worked out in a range of systems.
Even though quite cheap in most places, water is not necessarily free in Sweden - at least not according to the bill I receive every month...
(I believe if you knock on someone's door and ask for water you can't say no to that person - if within reason - and there are usually lots of public/free places to get water to drink)
I share your disdain for a central authority in making these judgement calls, but I have even less confidence in the majority of people who think they can solve everything with AI. Signed, a data scientist with a PhD in biomedical engineering.
Misunderstanding here. The question is at what age is the base rate in the population sufficiently high that a test with a certain sensitivity and specificity useful net risks. Multiple studies have shown that these screening programmes don’t have a huge impact on mortality - excluding lead time bias etc. and noting the c. 10 in 100k extra cases of cancer caused by the screening. Here’s a review for prostate screening that’s even more damning about its usefulness - https://med-fom-urlgsci.sites.olt.ubc.ca/files/2007/06/P-ca-...
MRI scans take longer, are often more uncomfortable for the patient (being in an enclosed, noisy machine), require much larger capital investment than an X-ray (and therefore supply is more limited), and have a higher cost per exam than an X-ray.
It almost certainly would be possible to use MRI for screening, but the impact would be a reduction in availability and a higher cost.
> A lot of people, including myself, don’t believe that central health authorities have the right to make that call.
That’s absolutely fine. You’re free to spend your money on preventative healthcare that’s been determined to fail a cost benefit test. Organizations that have to save as many lives as possible or prolong healthspan as long as possible within a limited budget must make decisions somehow.
The NHS has finite resources. They have to decide if implementing a screening program is worthwhile or not, versus spending the money elsewhere. You can still go to your doctor and get a mammogram (or even a more useful test) if you have other reasons why this might be justified in your situation. They aren’t ‘banning’ mammograms for young women.
Any sort of treatment is invasive. Almost all form of medical treatment has side effects and risks.
Finding out you have "cancer" is traumatic and extremely emotional, though breast cancer is one of the most survivable (in part because, well, everyone loves boobs. Prostate cancer, on the other hand...)
Putting these tools in the hands of medical professionals is one thing. Putting them in the hands of the general public is beyond irresponsible.
People physically assaulted doctors and nurses for not being given ivermectin; imagine how insufferable people will get when some website examined their mammogram and said they have cancer.
This is absurd logic. If the next step for a test like is a procedure with a lot of risk, change the next step.
We need to be able to work in a world with frequent, imperfect, low cost diagnostic tools. Cancer is almost completely survivable if caught early enough. So working to figure out early detection is effectively the "cure" for cancer we have been looking for.
As I recall the actual logic was after decades of early & frequent screening, allegedly the data showed a greatly increased rate of mass discovery and treatment, but no improvement in mortality at the population level. This suggested despite a lot of activity, we may not have accomplished anything.
but that means that instead of treatment the correct next step after discovery is monitoring of those masses, no?
also it's quite possible that (just as with COVID tests) we would benefit from more testing even if that test is not that reliable. (so there's an argument for developing fast non-ionizing radiation imaging machines, eg. a fast stand-in MRI)
I'm always surprised with the "just close your eyes" attitude of medical policies. I mean, this way we essentially choose when to get women misdiagnosed? Isn't the solution to get a better idea of how common it is for non-cancerous masses to appear and adjust the risk predictions? Or to actually improve the diagnostic methods?
The NHS is massive bureaucracy. What it does or doesn’t do is peering inside the belly of a Byzantine whale.
The article is a misnomer calling him an ‘amateur’. Its a click bait title. He’s shown himself to a world leading researcher in the application of AI to cancer screening.
Plenty of managers in the NHS can’t even do simple math.
Thanks for the different perspective. What did you mean by "unnecessary treatment" though? If you have cancer, doesn't it need to be treated? Doesn't cancer anywhere always cause harm to the body?
Breast cancer for example is diagnosed by increasing levels of invasiveness. First a mammogram, then possibly a 3D mammogram, then an ultrasound, then a biopsy. There are possibilities for false positives all along this path and increasing levels of possible complications when performing procedures. If a false positive gets to a biopsy and you get an infection from it, you would not have ever gotten that infection if they didn't start testing you so young. False positives are very common with breast cancer screening.
Not to mention the fact that getting a biopsy can cause the cancer to spread all over the body where it might never have grown beyond its original position had it been left untouched.
For many detectable tumors, the best answer is "wait and see", not "immediately remove". There are many reasons for this- surgery itself is risky, the tumor itself might not ever become harmful.
1) I just downloaded the "The Mammographic Image Analysis Society database of digital mammograms" [0] and ran it against the tool [1] image by image. Results below, code here [2]:
true_pos 36
true_neg 207
false_pos 63
false_neg 16
total 322
2) How is it true when the site [1] says "We will not store your data on our server. Please don't worry about any privacy issues." when you can find all analyzed mammograms under the "static" directory?
Thank you for your efforts for validation and I appreciate that. There is a script running in the background to auto clean the files in static folder every day.
I think caching people's sensitive medical history in a way that is publicly accessible is worth calling out; especially as it could be trivially fixed to maintain privacy.
This is very cool work. I'm a radiologist, I also work on developing ML/AI based systems for cancer detection and characterization. Literally just took a break for a few minutes from creating some labels and saw this as the top HN post!
I think in some ways making the model available online can be good, but in other ways could be harmful too. Very complicated topic.
I've always felt the "could be harmful" was a rationalization by radiologists worried about their job security since it's easily mitigated with a warning and multiple tests.
And especially because in the future, most radiology work will be done by software. It's just a matter of whether it's 10 years or 100 years from now.
> I've always felt the "could be harmful" was a rationalization by radiologists worried about their job security
Surely concern for the well-being of the patient figures in there somewhere...
Or imagine this: A liver lesion is incidentally discovered on your abdominal CT performed for unrelated reasons. Its radiographic characteristics are equivocal. Additional imaging studies fail to completely exclude the possibility of a liver malignancy. You undergo a biopsy. But the biopsy is complicated by hemorrhage. Surgery is required. You develop a post-operative nosocomial infection. etc. etc.
To the extent that risks along this chain of unfortunate events is known, yes, warnings could put some of the quantified decision making in the patient's hands. Well, except for the rampant innumeracy in the general population...
This topic (breast screening) is a good example due to the sheer scale. If you increase the work-up rate by even a smallish amount, you are statistically pretty much guaranteed to kill people who did not have the disease. How this balances about gain (i.e. save other lives) is not obvious. Figuring out the "right" way to do this is real work.
Also a good example because there's actually been recent push back on mammography because the risks may outweigh the rewards (especially in younger age brackets). More testing does not automatically equal better outcomes. Here's one summary from the Nordic Cochrane Center (for those that don't know Cochrane is one of the gold standard organizations for evaluating the quality of biomedical research via systematic reviews https://en.wikipedia.org/wiki/Cochrane_(organisation) ):
Extremely important, and is the most important question to ask in any screening scenario.
It's not the case the more screening is always better. There are plenty of screening regimes that you've never heard about because the trial data didn't support it.
That’s not true. I would love to have someone check my work. It’s just got so many false positives. Would you trust the radiologist if a report shows a suspicious lesion marked by the software that the radiologist dismisses? CAD in mammography has been there since the 90s. rads don’t put up a fight.
Do you want a whole FSD/Autopilot level of blame absolution if software missed a cancer/critical finding? I sure as hell won’t be apart of that.
Yeah, I've read all the arguments about harm, and I just don't buy it. I'd get a CT scan every year if I could convince someone to sign off on it. There is absolutely no substitute for knowing.
> Yeah, I've read all the arguments about harm, and I just don't buy it.
"I have no medical background whatsoever but I do not buy undisputed medical science."
> I'd get a CT scan every year if I could convince someone to sign off on it
20mSv of unnecessary ionizing radiation exposure a year, what could possibly go wrong? That whole-body CT scan is equivalent to getting at least EIGHTY chest x-rays, and ten times what a uranium miner receives in a year, and well within the range where an increased risk of cancer is noticeable in epidemiological data.
Case and point why this tool should not be generally available.
You would choose annual body CT even if outcomes-based evidence showed no benefit?
Even among those at highest risk for lung carcinoma, studies of annual screening chest CT do not uniformly show improved disease-specific survival as compared to an unscreened but otherwise matched cohort.
What I found mind blowing when I first read these studies (particularly the second link) is that the positive-predictive value of the CT screen was only 3.8% for lung nodules that were 4mm or larger in size. Basically just means we find all kinds of lung abnormalities all the time that "could be cancer" but aren't.
Keep in mind, that the screening criteria was people over age 65 who had a 30+ pack-year smoking history (1 pack-year = smoking 1 pack of cigarettes per day in one year), and had to have smoked within the last 15 years.
I can also tell you from doing CT lung biopsies that it's not a procedure to take lightly. Lot of important structures in the chest.
Absolutely. Those studies show no improved survival across cohorts. Not for individuals. My only chance of living is to know. That's worth it for me.
I mean, plus I have extreme health anxiety, and I've had doctors say "nah you're too young, we don't need to test for that" and later it turned out I did have kidney failure at 35.
Yeah, that is the only concern. I guess whole-body MR would be fine too. I'd be willing to lay in the tube for 20 hours a year to know that I'm at least structurally normal still.
How could it be harmful? Is it just because of errors, i.e. false negative means person won't talk to doctor and won't be caught early; false positive means needless worry, maybe tarnish more general medical industry? Or something else?
Yeah, those are part of the issue. For example on the webpage, there is no discussion about the false negative/positive rate of the system, and no comparison or link to an academic paper that compares the performance of the system to a trained breast radiologist.
Does a patient take a negative exam as "I'm good!" and forego actually seeing a doctor, or having the exam read by a radiologist?
A positive result is also a challenging situation. I've talked to patients before after diagnosing them, and as you could expect most people are shocked, scared or they want to know what they should do next.
In breast radiology specifically (at least in the US) there is a well defined reporting lexicon and classification system. From what I can tell, the system does not use the lexicon or BI-RADS classification. https://www.acr.org/-/media/ACR/Files/RADS/BI-RADS/BIRADS-Re...
Diagnostic and screening images also undergo rigorous QA for the entire system. Does the model have some kind of QA built in?
There is a just a small disclaimer at the top that says "we will not store your data on the server. please do not worry about privacy issues". Not exactly a formal or legal binding agreement.
These are just a few things that come to mind off the top of my head. Like I said before, I do think that what coolwulf works on and is trying to do is good, and in the long run can help doctors better characterize findings and help patients. But like anything in health care, there are a lot of edge cases and side effects that you have to think about. The stakes are also very high.
I think care with messaging positive/negative rate is the right reasoning for a productized version of this positioned as an alternative to human radiologists. The difference here is that we know the user has a scan of their body and we can be pretty sure they couldn’t have gotten the scan without doctors already being involved.
So we don’t have to worry as much about the “Does a patient take a negative exam as ‘I'm good!’”, because in the normal flow of medicine that gets this scan the doctor will check the scan for them either way. The website probably could have better explanations of this, but most likely negatives will already be double checked independently by doctors, and positives will most likely be handled correctly by bringing it to their current current attention to double check the scan or run more tests.
> Does a patient take a negative exam as "I'm good!" and forego actually seeing a doctor, or having the exam read by a radiologist?
I wonder how the false positive/false negative rates for this tool compare to that of a trained human. At some point we may reach parity, in which case what's the harm of trusting the automated result?
I don't think that a human reader in the loop will be going away for a long time. Sure, they can be at parity, but does a combined read have better results? This is an active research question.
I think what will happen is that over time, human interpreters and AI systems will "co-evolve" in a sense, where people will pick up on where models are wrong and also learn how to use models to understand their own blind spots. These are also active research topics that are in their infancy.
The consequences are false positives and negatives in cancer detection are well known (easily googled), and are actually extremely important when looking at cancer treatment in large populations. You nailed the most important ones.
I'm not worried. If things are that different, the rest of the economy will be too. The job will also evolve over time, like any other profession that changes with new technology.
It's our job (as radiologists and engineers) to shape that in a way that benefits patients.
This is very true. Data availability (and moreso, label availability) is the biggest barrier to improvement here I suspect [ thanks for labeling!]. Access being another. Using a public site to bootstrap that could do very interesting things. On the other hand, public access to a poorly RA/QA'd algorithm could also cause more trouble than help, easily.
Label availability is a big problem. There are some academic collaborations for some diseases which is making this better, but datasets in general are minuscule compared to what is available for more general computer vision applications.
I am hopeful though. Few-shot learning and self supervision are very active questions right now and there are a lot of papers in the medical AI field that are getting published on these topics.
I'm personally interested in liver cancer, which does not have large, well-curated and shared database of cases.
Sharing data gets tricky, especially in the US. Labels I'm working with and creating (at least at this point) are for my own research, which won't be shared publicly any time soon.
> n the medical AI field that are getting published on these topics.
Most of what I've seen isn't very promising. The energy in these research areas are because it would be so much cheaper than the "right" way, far more than because of the likelihood of success. And also perversely, because it's hard for the academic researchers to get enough data to do other studies :) NB: I'm not saying there is nothing useful coming out of the learning literature in last few years, just that it a) isn't a silver bullet and b) is often being misapplied in these areas anyway.
Label quality and availability isn't the only big problem though. Many data sets exhibit problematic sampling bias, as well as being order(s) of magnitude too small, because of the way they are gathered and how access is granted.
Problem is though for most of these diseases there just aren't the number of samples available, period, to do it the "right way". HCC for example, has around 50K new cases/year in the US. Even if every single case went into a repository with perfect labels, would still take a long time to collect that info. Not to mention you need either a radiologist (4 year of medical school + 6 years of post-school training) or a very skilled and experienced technician to label the data.
Not to mention imaging protocols are not standardized, and the imaging technology is also evolving so scans we do today may not be "correct" or standard in 5-10 years.
Definitely diseases have different challenges. Breast cancer screening being a notable outlier as far as data availabilty. For some diseases ML is probably always going to be problematic although may help in diagnostics mostly by helping get rid of other possibilities.
I suspect we have similar overall views of the problem, but I'm pretty strongly in camp that recent advances in ML/AI are mostly really driven by data & label availability, not algorithmic advances - this colors where I think the wins to be had in medical ML can happen most easily. Either way though the non-technical barriers seem clearly higher than the technical ones still.
Really enjoyed your contributions to this thread, thanks.
I’m a first year radiology registrar (PGY3) in Australia looking to find others doing interesting work in this domain, if you think I could help with your efforts feel free to DM
True! I certainly wouldn't discourage anyone from trying.
On the other hand, I think it would be a huge mistake to trust that fancy learning approaches will solve everything so we shouldn't try and improve access and labling. Getting better there is still by far the most high probability of successful impact, imo.
What peer review or regulatory approval process has this been through? Seems pretty irresponsible -- there are many notorious pitfalls encountered with ML for medical imaging. You shouldn't play with people's lives.
I don't understand why this comment is downvoted. Automated screening of radiological images by means of neural net is an extensively researched topic. Ten years ago there had been predictions that such automated screening will displace the radiologists, but that clearly did not happen.
For instance, this article is silent on false positive/false negative rates of the software. There is no comparison with other research. It reads like a corporate press release promoting a product.
The quotations around "amateur" should be moved to "fought". Nonetheless, it's encouraging that this level of research can be executed at home, however the strict burden of proof required should still be maintained.
Well, he was director of R&D for a medical imaging company and worked directly with academia. So I think its appropriate to contextualise the "amateur" because his work looked much more amateur than it really was.
This is an incredibly important point. Medical research must be taken seriously and I see many problems with the processes being applied here.
(for those who care- I'm a published ml biologist who works for a pharma that develops human health products. Having worked in this area for some time, I often see people who have no real idea of how the medical establishment works, or how diagnostics are marketed/sold/regulated. Overconfidence by naive individuals can have massive negative outcomes.
Why does everyone assume this guy has zero business attempting this? If you read his credentials, he should be every bit as qualified as you to attempt this kind of work while understanding the pitfalls.
According to his CV he's been active in the field for quite some time. The default assumption that he's an idiot and going to kill people just seems too cynical here.
Grandparent - you specifically mention having noted methodology problems, would you mind sharing where in the methodology you think he's gone wrong?
I'm not going to detail challenges in medical ML - the literature can do that. But just to mention one that other people haven't that goes beyond just precision-recall: algorithms can be biased based on variance in physiology (e.g. more accurate for men), and understanding how an algorithm is biased is very important for the person interpeting the information, who should be a trained doctor.
And it doesn't matter what his credentials are, that's appeal to authority. If he thinks this should be used and trusted by people for decision making, then he should submit it to independent peer review and regulatory approval.
This is like taking your temperature at home, are you making a diagnostic yourself? Not quite. But you can know some symptoms and take action (going to the doctor) maybe with less anxiety
Thermometers are well understood, simple devices, and there are other complementary checks (e.g. does my forehead feel hot) if they fail.
This project might lead to people thinking they're in the clear and not seek appropriate medical treatment, or be overtreated due to an error. You should always talk to a qualified doctor if you're concerned about your health, and not use projects like these for decision making.
I don't see how a passive scan like this can be harmful . Ofcourse if it does show a positive one should confirm further with a biopsy or other standard confirmation diagnostic.
As for false negatives...
If you feel something is wrong you should get it checked thoroughly anyway .
you literally just said the scan has no value. THe point of a scan like this is to have absurdly low false negative and positive rates so that it's actionable. Unactionable medical diagnostics are worthless and just cost (money, fuel, time).
In an ideal world, that would be the case. However, people aren't 100% rational agents motivated by logic.
My aunt was diagnosed with Multiple Sclerosis which led her to lose her eyesight from her left eye, because she refused to get a thorough checkup by a professional, and even today whenever I tell her that I visited the doctor for an issue she has... not very good words to say (something something "you are a chicken, you're hypochodriac etc). And I'm saying this without entertaining the probability of her visiting a professional who just happened to be in a bad day, which could potentialy lead to a wrong diagnosis.
I've been hunting down my own diagnosis for symptoms everyone seems to tell me that aren't serious (nail discoloration and a 24/7 headache that feels like my arteries are pulsing, which lasts for YEARS, cold fingertips during the winter, and more).
I get what you're trying to say and I agree with the general message. However, more checkpoints to catch a potential failure are good. For example, if someone were to make a take-at-home device which scans nailfold capillaries (no reason for something like that to exist) I'd get that in a heartbeat. I'm being actively ignored by every medical professional that I have visited, and if I'm not ignored they give the minimum amount of attention, kind of like "well, it's not like you're dying so who cares?"
If you don't feel good yes you should see a doctor regardless.
Let's say you get a scan every year . If the scan is able to detect something earlier than a radiologist is able to identify i think it's worth paying attention to.
I agree with you, but what's most important is the impression that the average person who uses it will have. And I don't think most people would think of this as like "taking your temperature at home". I think most people who might upload their x-ray scans would take this a lot more seriously.
A false positive could create a lot of anxiety and emotional distress, and the patient might need to go to 2, 3, or 4 other doctors to get second opinions before they feel comfortable that they really don't have cancer.
A false negative could be even worse. A patient might think "oh, the official-looking online thing said I don't have cancer, so I don't need to wait for or consider a human radiologist's results", and not believe they need treatment.
I think it's very important that people understand that -- until more research is done -- this is still not a substitute for having a human look at your x-rays. If we could be reasonably sure that everyone (or at least a very large majority) understood this when using this tool, then I think people would have far fewer objections. But I don't think that's the case.
Having said that, I think it's safe to assume that this tool has saved lives, so it's almost certainly been a net positive for people.
A thermometer mostly tells you if you have an infection, and how close you are to your body temperature becoming a medically urgent or life-threatening situation.
In addition, these kinds of things will still miss lobular ("normal" cancers are ductal) breast cancers as they don't form lumps.
15% of the women with breast cancer are waiting for a non-invasize diagnostic imaging system that can see their cancer. The only thing that can see these is an MRI with gadolinium. And that gadolinium contrast causes issues in about 1 in 1000 women, so it can't be used as a general screen.
If your decision making process is a negative result tells you nothing and a positive result warrants immediate follow up, what’s the risk here? I’m assuming doctors recommending that women get checked for breast cancer is the primary breast cancer is tested and diagnosed which presumably wouldn’t change because someone make a website.
In this case I feel better about it because there is a natural limitation in that most people doing this will only have the scan because they are getting tested through a real clinical process. So effectively they are getting "standard of care" treatment implicitly, and all this does is accelerate their response to true positives. The worst case scenario is a false positive gives them a lot of anxiety / costs them money through trying to accelerate their real diagnosis only to find it isn't real.
ETH will soon move from PoW to PoS (let's not debate the timeline or if it is a good idea). This will put about 32 million GPUs worth of compute and millions of CPUs searching for something else to do (or just flood the market with used equipment).
I have been searching, for years, for alternative workloads for these GPUs beyond just PoW mining and password cracking. Many of them are on systems with tiny cpus, little memory, little disk, little networking so the options are heavily limited. AI/ML/Rendering/Gaming actually make bad use cases.
If anyone has thoughts on this, I'd appreciate hearing them. Let it all die is certainly an option, but it also seems just as wasteful as keeping it going. Maybe we can find a better use case, like somehow curing cancer...
> ETH will soon move from PoW to PoS (let's not debate the timeline or if it is a good idea). This will put about 32 million GPUs worth of compute and millions of CPUs searching for something else to do (or just flood the market with used equipment).
Crypto markets crashing together could do this, but ETH's switch isn't going to do much for old cards.
Checking https://whattomine.com/ shows that ETH mining isn't even in the top 5 most profitable things to mine with a 1080Ti right now. The miners looking to squeeze every bit of profitability out of old hardware switched away from ETH a long time ago.
1) The cards have already paid for themselves. They are 100% ROI positive and even at the current low amounts very profitable. Regardless of what W2M says, ETH is still the top most profitable coin. Large miners don't sell immediately, they wait for the market to go up or the option against their ETH holdings.
What? ETH isn’t the most profitable thing for these cards to mine. Why are you so confident that ETH switching to PoS will open the floodgates when ETH is already not the driver of their usage?
The ETH proof-of-stake switch isn’t going to be the driving force they causes GPUs to be abandoned in bulk.
> ETH isn’t the most profitable thing for these cards to mine.
You can mine shitcoins on a small scale for maybe a bit more profit. The issue is always unloading a lot of shitcoin in a very small volume market. Even ETC has a 6.5 day deposit time on exchanges like Kraken. Nobody is going to want to deal with that.
> The ETH proof-of-stake switch isn’t going to be the driving force they causes GPUs to be abandoned in bulk.
The GPUs will move to other shitcoins, which will cause the prices of those shitcoins to go down as people will mine and dump them, which will cause a lot of GPU miners to shut off.
Usually, with POW, the issuance is the same. So they will mint the same amount, no way that will crash the price. What _will_ make the mining not profitable is the difficulty increase. It boggles my mind people don't seem to get this.
The merge will absolutely happen, there are 12.6m ETH in the staking contract [1] that people want out of there and the only way to get it is to get to the next steps.
I hear people say this, but I am absolutely certain their assumptions are wrong.
1) The sum total of rewards is fixed for POW
2) Introducing extra hashing power will increase the difficulties of these mining ops up to the profitability equilibrium point.
After the overall "free" hashing power increases to a point, GPU's will start flooding the market at dumping prices.
There are plenty of good uses, projects like BOINC have been using GPUs for good for over a decade. The problem is the incentive system disappears, it's a lot easier to get people to run 32 million GPUs when it makes them money instead of costs them money.
This seems like an incentivized BOINC built on top of an inflationary shitcoin. There is no utility in the coin itself and that is reflected in its price history.
What we need is something that has utility... like run BOINC, earn tokens that can be used in the real world for something other than just dumping on the market.
It's not worth implementing for MacOS because it woulnd't increase the overall folding rate very much given the relative weakness of the Mac GPUs compared to the existing f@h fleet.
folding@home has been doing this for 20+ years. They already did all the smart research and tech development. Just use that until somebody comes up with a workable DrugDiscoveryAtHome or CureCancerAtHome.
This is a very dismissive answer which seems odd coming from someone with a lot of karma. Running GPUs at scale isn't easy or cheap.
There is no incentive to run this other than good feelings. Unfortunately, that isn't enough in the business world to spend millions on cap/opx.
What I'm looking for is incentivized options. Even better if they come from a web3 situation where a business can operate without having actual customers.
It's ironic to be declaring the GP as dismissive while you're flippantly dismissing the work of everyone involved with F@H. You're conveniently ignoring the incentive of expanding our understanding of biology, which has very real applications and results[1] that benefit the entire world for the rest of human existence, instead of benefiting a single participant in the short term.
You literally said "no incentive [...] other than good feelings." Regardless, even adding the words "direct financial" is still incorrect. For one of many examples, insurance companies investing in F@H infrastructure would be beneficiaries, albeit via savings and not revenue.
You're moving the goal posts, on top of changing the definition of "direct" to fit your false narrative. First you claim there was no incentive at all, then there was no "direct" incentive, and now direct apparently means "immediate" instead of linear.
If a company invests in research specifically because the fruits of that research will reduce expenses, that is direct financial benefit.
Expenses are the "loss" portion of "profit and loss." That is 50% of the concept of accounting, making it as direct as it gets. Please stop this prolonged line of trolling.
Yes Digital Ocean. They are not a ponzi/pyramid because
1. They sell a useful product
2. They do not charge people to be affiliates.
3. They do not sell a token/security that you buy with the intent of making future gains, funded only by other people buying the token expecting more people will do the same.
If anyone uses the link ... I haven't checked ... I get some DO credit, not even money, which I will use for hobby projects. If I get $5 in credit in a year, I would be pretty suprised.
It is structurally more like a ponzi than a pyramid.
By large crypto has no MLM style referral system where the product is the business opportunity to sell the product.
However it does have a feature where crypto gains are only funded by newer people coming in and buying at a higher price. But no cashflows are produced other than selling the investment tokens. This is almost identical to a ponzi except it has decentralisation that makes it probably legally not a ponzi.
Imagine I did an IPO of a company I incorporated last night but hid behind a panama shell. That’s crypto.
> However it does have a feature where crypto gains are only funded by newer people coming in and buying at a higher price.
"only" is inaccurate. This was true early on before there were markets and contracts. Now we have DeFi which allows for gains on trades by anyone who wants to provide liquidity. Similar to any sort of commodity market where people buy and sell what they own (or even don't own with margin and options).
We conclusively demonstrated that kinetic models of folding are critical to do better drug discovery against GPCRs and other target classes.
Or did you mean something more fundamental, like "the biophysics of protein folding is primarily determined by entropic-driven hydrophobic collapse, not enthalpic contributions from hydrogen bonding?"
>We conclusively demonstrated that kinetic models of folding are critical to do better drug discovery against GPCRs and other target classes.
How was this demonstration done? Or are you referring to building a better high throughput screening mechanism? I'm by no means a protein expert, but I am trying to learn more about them.
>Or did you mean something more fundamental, like "the biophysics of protein folding is primarily determined by entropic-driven hydrophobic collapse, not enthalpic contributions from hydrogen bonding?"
I don't want to minimize anything or anyones work. But conformation of experimental observations via molecular dynamics simulation isn't what I was thinking of - but maybe I am misunderstanding your claim. Its my own ignorance.
Oh you know what an alternative use is? Oaths. Works with old ASICs as well...well I think. So you take a document, like this comment, you append a nonce (you'll see) and you hash it until you get a lot of zeroes in the front. Same as bitcoin, but you're not hashing the bitcoin protocol. Then, you know the document has been sworn, as a cryptographic oath, to that extent. Nonce: 38943
> oath = "Oh you know what an alternative use is? Oaths. Works with old ASICs as well...well I think. So you take a document, like this comment, you append a nonce (you'll see) and you hash it until you get a lot of zeroes in the front. Same as bitcoin, but you're not hashing the bitcoin protocol. Then, you know the document has been sworn, as a cryptographic oath, to that extent. Nonce: 38943"
Note the leading 0's. That's no timestamp, that's an oath.
In reply to the sibling comment: no. I like timestamp.com, and in fact I could have never found out about it other than by talking about the oath concept, but this is not just including it in the blockchain. It's proving its value to the author to bother doing the work of getting a good nonce for it. Literally putting my money where my mouth is. And swearing an oath to that extent, I could cryptographically swear it more, with more work, or use a smaller less impressive nonce if I'm not as sure.
And incentives? There is an incentive for me. At the same time it is effectively burning money, swearing by burning money. Took like seven seconds of compute, too. I had to wait human time for that. It's collateral, it's an oath. And it's an impediment to forgery, and in addition, an impediment to eg news sites telling different people different things. With oaths they have to tell everybody the same thing.
Because folding is too hard to verify. For a Blockchain to work it needs to be very easy to prove that the miner has actually done the work. With a folding problem it's very hard to prove the answer a miner gives is actually a solution to the problem rather than a quick guess. It's a shame!
Just make it centralized (like most cryptocurrency effectively is anyway). You would still need some checks and balances sure.
Make a token. Altruists buy the token to fund the “miners”. They may also make a profit but they buy knowing most likely they wont and it is for a good cause.
If anyone ever needs an example of how crypto is doing harm to the world: "Due to crypto mining, GPUs were in severe shortage and very over-priced on eBay"
The over-valuation of crypto is a two-fold negative impact on society: a massive brain-drain sucking talented engineers who would otherwise be solving real problems, and the opportunity cost of GPUs burning electricity to run unintentional ponzi schemes instead of training deep learning models.
You could argue the other way around. Crypto is creating market incentives to develop more powerful GPUs (for less $$$), which can make other applications possible. It's tricky to control market forces...
Is it actually doing that though? The main mine-able crypto is Ethereum and I don't think Nvidia is specifically optimizing their consumer gaming GPUs for maximizing hashrates. The only innovation in this regard is their "mining" gpus which are the same thing as their normal GPUs just without physical video outputs.
I'm not complaining about the market forces causing GPUs to be expensive, I'm complaining about the non-regulation of marketing these as securities to laymen, which is exactly what is happening right now every time you see an NFT or ICO ad on Facebook or Instagram.
Wolf on Wall Street is a movie about this happening before in history, and the current craze around crypto is exactly the same thing and the reason why the qualified investor rules were made.
If you stop dumb money being funneled into crypto, and stop the moral acceptance of chasing dumb money, you can free people to pursue more noble causes.
I'm not anti crypto, but anyone can see that a good chunk, if not a majority, of current interest in crypto is due to chasing dumb money and greed. If I recall correctly Vitalik Buterin tweeted about this excess attention before. Bad attention could kill crypto before it can truly even get its feet off the ground.
It doesn't matter if the GPUs are optimized or not for crypto - I'd argue it's even better if it's not. The point is that Nvidia and AMD reach better economies of scale (which are critical for such R&D intensive businesses as cutting edge GPUs). That benefits all, if it goes in somewhat the right direction.
* The free AI breast cancer detection website took coolwulf about three months of spare time, sometime he had to sleep in his office to get things done, before the site finally went live in 2018
* The website also gained a lot of attention from the industry, during which many domestic and foreign medical institutions, such as Fudan University Hospital, expressed their gratitude to him by email and were willing to provide financial and technical support
* Afterwards, he and Weiguo Lu, now a tenured professor at University of Texas Southwest Medical Center, founded two software companies targeting the radiotherapy and started working on product development for cancer radiotherapy and artificial intelligence technologies
* But in 2022, he returned with an even more important "brain cancer project"
* coolwulf (Hao Jiang) (right) He told us that his parents are not medical professionals, and his interest in programming was fostered from a young age
* A reliable AI for tumor detection can enable a large number of patients who cannot seek adequate medical diagnosis in time to know the condition earlier or provide a secondary opinion
* He said that he’s not sure actually how many people have used it because the data is not saved on the server due to patient privacy concerns
> A reliable AI for tumor detection can enable a large number of patients who cannot seek adequate medical diagnosis in time to know the condition earlier or provide a secondary opinion
It's an interesting subject, with a long history; I think many of the biggest challenges are not technical.
The first commercially available AI/ML approach to breast cancer screening was available (US) in the late 90s. There have been many iterations and some improvements since, none of which really knock it out of the park but most clinical radiologists see the value. Perhaps the more interesting question then is why are people getting value out of uploading their own scans, i.e. why does their standard care path not already include this?
The reason I made this project 100% free and available to the general public is to help patients, especially in the remote area who has limited access to experienced radiologists for diagnosis, to at least get a second opinion on their mammogram. And I think this has certain value and this is why I'm doing this project.
Question I have for you is that one of the biggest problems with cancer diagnoses is false positives: "Yes, there is something on your scan, we're not sure what it is, so we'll biopsy it." Biopsy is not a 0-risk procedure, and it can cause a lot of worry and pain, so it's not something to be taken lightly. Also, there are many cases of "OK, it's probably cancer, but the cure may be worse than the disease." This is the classic problem with detecting prostate cancer at an advanced age - it's very likely/probable something else will kill you before the cancer does.
How does your software deal with this issue? I'd be worried if, as you put it, people in a remote area with limited access to experienced radiologists, were given access to this and it came back with "Fairly decent chance of cancer" - what do they do then?
Great question. Definitely sensitivity / specificity balance is a crucial topic in AI-assisted diagnosis. I have to admit this model and website for mammography was done in 2018 and might not be the leading solution out there. At the moment, if I want to improve the results of my earlier work, I will add additional stage of radiomics model to do false positive reduction, and in the mean time lower the threshold in the first model to increase sensitivity. By doing this, combining the feature extractions of deep learning and the past 20 years of knowledge in medical imaging using Radiomics might give better performance in terms of sensitivity / specificity.
You're operating in a really serious area of medicine and I encourage you to takes the comments about false positives (and false negatives) more seriously. It's not really just a matter of making adjustments to a model; it has to be pervasive in the entire process of making reproducible modeels that are used for making decisiojns about humans.
The website is cleared marked as a breast health awareness tool and not for diagnosis. We are not doing any decision making for humans. However I would like to point out there are quite some FDA approved mammo AI products in the market at the moment.
Even if people use it for diagnosis, so what? Take the info and if it’s positive to confirm with doctor, if it’s negative but you have reservations go see doctor.
I think this comment is perhaps a bit uncharitable.
Even if you have done everything you are supposed to do in the process, at the end of the day you are looking at ROC curves or equivalent and trying to understand sensitivity vs. specificity trade offs, and often you have some sort of (indirect) parameters than can move to different points of that tradeoff.
This is quite critical in deployment, if you are screening you usually want something different than in diagnosis; as alluded to elsewhere if you raise the work-up rate too much you definitely risk killing more people from biopsy complications than you help with higher sensitivity (it's more complicated than that in practice)
deploying ML medical diagnostics is like everything else in ML: the ML part is 1% of a much larger "thing", which involves business, legal, and many other concerns far beyond the data analytics.
Nothing abotu what I'm saying is uncharitable- in a sense, it's charitable because I'm helping warn a person who is going down a dangerous path to consult more with experts in this field.
The charitable response would have been to assume they do have some understanding the broader context, and perhaps raise specific concerns or points of interest.
What you did was assume that they were ignorant in potentially dangerous ways, and assert they should do something different.
I don't think a reading of the comments/responses (at least at your time of posting) really supported that assumption, especially considering the limitations of the medium. Hence my reply to you, while also detailing the trade offs a tiny bit.
I read the entire article and I didn't see anything in there that would convince me this author is anything other than an amateur programmer (I can't parse the section about bruker- is the "amateur programmer" also a director at Bruker who ddevelops medical devices full-time"?).
Please be assured that I put a fair amount of thought into this - for example, I used to do due diligence for a VC firm evaluating proposals like this all the time and we had to reject most of them because the founders didn't understand the basic rules of deploying medical technology in highly regulated environments like the US.
Based on my interactions with the author in the various parts this post, I continue to conclude this individual is lacking core knowledge and wisdom required to execute a project like this successfully at scale.
The article was pretty fluffy, but it was about them not by them. If article was accurate about the role at Siemens they have for certain been exposed to RA/QA work and know what a DHF is, etc.
Anyway and least at they time you posted (since then there were more interaction) I didn't find the same information nearly enough to dismiss their competence out of hand.
I went back and dwetermined that the article was wrong. He wasn't a "director" at Bruker, he was a "detector imaging scientist". There's nothing about Siemens.
So this isn't an amateur programmer, it's a person who got a phd in nuclear engineering and radiological sciences, was a scientist at bruker, has some experience with health systems, and then became a serial entrepeneur with a small company that has some funding. BTW, people who have the job title "Director" are normally fairly senior (old), as well.
You will save even more lives if you over-communicate about the need to use your awareness tools (with better false positive reduction) to drive more effective diagnosis.
Inconsistent care is a really good point. I wasn't trying to be negative - hope it didn't come across that way. I was hoping to point out that systemic issues in health care management, at least in a lot of countries, seems to be more of a problem that tech for things like this.
Out of curiosity, how are you handling the data access and labelling issues here? I suspect that's the key issue that has limited the performance of the commercial offerings (hardly limited to this problem or this space).
OTOH in terms of real impact, properly leveraging a more modestly successful algorithm will probably help more people than getting a few more %. With the (strong) caveat that in a space like this you really have to look at work-up rate and balance risks.
There is history of breast cancer in my family and anything that can be done to improve outcomes has to be highly commended. I did however have the same question and this
> to at least get a second opinion on their mammogram.
for me makes a lot of sense, even in developed countries where you get a result but want extra assurances.
It would be interesting to know (assuming you have the data, even anecdotally) if the second opinions using this overturned professional ones and from those how many were corrected an original false negative mistake.
Not speaking for coolwulf obviously but I can perhaps shed some light.
Screening breast mammo has an occurrence rate problme. Something like less than 10 in 1000 studies will require further review; this means in practice as a radiologist you look at a lot of negative films before seeing a TP. It also means a typical read is done fast. Seconds-to-small minutes.
This results in a couple of things. Reader variability based on experience/throughput, and false negatives. There were some double reader studies that caught something like 15% (going from memory here) of FN - but nobody can affort to have two radiologists read everything.
So the profession is already conceptually used to the idea of using an algorithm as a "second read" and reconsidering. Typically this won't "overturn" anything here but rather say 'hey have another look', but the decision to proceed or not is still the clinicians. Having a positive from the algorithm makes them review carefully, but you have to watch the FP rate here or nobody would get anything else done.
I have heard of health systems using algorithms as a first pass too (i.e. radiologist only see films that have had a postive in a tuned-to-be-senstive version), but that has it's own set of issues.
'Amateur' oughtn't be scare-quoted because it's not a slur, many of the finest programmers were amateurs for many years before they were old enough to be given a job in the profession.
I had an art teacher affectionately remind me the etymology for amateur
> borrowed from French, going back to Middle French, "one who loves, lover," borrowed from Latin amātor "lover, enthusiastic admirer, devotee," from amāre "to have affection for, love, be in love, make love to" (of uncertain origin) + -tōr-, -tor, agent suffix
https://www.merriam-webster.com/dictionary/amateur#etymology...
changes the feeling of it all when you get that context, someone who loves a subject pretty much--no qualifications skill wise or regarding depth but they love it and should presumably take things seriously to some degree as any lover would.
Using 'amateur' (quoted or not) is click bait. It's an embellishment to the rest of the headline. For that matter even though it's true the graphics cards are as well. Only thing that could have made it more click bait would be to also put in AI in the headline.
True, rather like age (or any discrimination category actually) is often used to make something sound more of an achievement, even though it's actually just about right place right time, experience (which you could happen to have at almost any age), etc.
If u used to be a paid software programmer and got a different job, but continued doing programming side projects without pay, are u an amateur or not?
Yes, but you'd also be a former professional. Amateur also isn't meant to be a negative term, it just means you do it as a hobby rather than as a profession.
As an AI/ML researcher who publishes in this area regularly, I will be using this as a case study for AI ethics classes. That this is allowed to go on is shocking.
> In 2018, a programmer named “coolwulf” started a thread about a website he had made. Users just need to upload their X-ray images, then they can let AI to carry out their own fast diagnosis of breast cancer disease.
Literally the worst fears that we have as a community is that people will recklessly apply ML to things like cancer screening on open websites and cause countless deaths, bankruptcies, needless procedures, etc. How many people went to this website, uploaded images, were told were ok and didn't follow up? How many were told they have cancer and insisted on procedures they didn't need?
The website is totally unaccountable. Totally unregulated. Totally without any of the most basic ethical standards in medicine. Without even the most basic human rights for patients. This is frankly disgusting.
In the US this would have been shut down by the FDA immediately.
We should not be celebrating this unethical "science" that doesn't meet even the most basic of scientific standards or ethical standards.
Unfortunately, it's just a matter of time before someone gets severe complications for whatever outcome your program gets wrong. You have a false negative? Big problem, you just gave someone worse treatment options (if lucky) or led to someone being diagnosed too late and potentially getting metastasized breast cancer. Those bone metastases are quite painful, you know.
False positives are even worse, because they are far far far more likely to happen in practice. Imagine your program telling someone it has a malignant mass (VERY bad wording, only the pathologist can say something is malignant). I speak from experience that this WILL lead to the patient going to the doctor, and the doctor, seeing these very strong words, might want to take a biopsy to confirm the malignancy. These are painful procedures that can and will go wrong, eventually (complication rate is ~1%, which is not a whole lot on its own, but is unacceptable on a known healthy population). If lucky, the biopsy can be done stereotactical, but if unlucky it'll have to be done MRI-guided. You just cost society a few thousand dollars/euros/yuan. And that's if everything goes right, worst case it'll be hospital admission due to complications, like an infection.
Your blog post says you are trying to fight cancer, which is a noble cause. If the tool is not for diagnostical purposes, it's not doing a whole lot in fighting cancer as it is just a play thing then. At the moment, it's more like hindering cancer by taking resources from people who need them and giving them to people who don't need them.
Source: researcher in AI for breast cancer screening/diagnosis.
Of course, this is going to harm some people. Is it possible the number of people it helps is significantly larger than the number of people it harms? If it were so, would you not consider that a reasonable tradeoff?
A slight digression: I find in certain countries (cough cough US), everyone is free to give advice/opinion on anything, except for medical/legal matters, which are considered sacred and so no one is free to give any advice/opinion on those. I saw a person scolded by HR for literally saying "make sure you stay well hydrated". This is madness! Medical professionals should not have a monopoly on advice concerning hydration.
The proper response to being scolded by HR for saying that is to ask what the person's username is, so you can speak to their manager.
Then, when you speak to the manager, ask them not to have their report telling people not to do things that HR doesn't have authority over (coworkers discussing health topics like that are not within the scope of HR).
Is there any possible way that a test that is not entirely reliable can be used for screening? Like, couldn't they have a radiologist look at the photo to confirm the result?
Yeah you could do that and it's happening in practice. Many countries implement double reading where two radiologists look at it. In the USA, single reading + computer-aided diagnosis is quite common using commercial, FDA-approved software.
> just a matter of time before someone gets severe complications for whatever outcome your program
If you are going to indirectly call someone out for publishing code which you deem severely detrimental to human health, you better back your stuff up.
> You have a false negative? Big problem
How does one get a false negative from multi-class probabilities? There aren't any...
> you just gave someone worse treatment options (if lucky) or led to someone being diagnosed too late and potentially getting metastasized breast cancer. Those bone metastases are quite painful, you know.
An image is already checked by a doctor, before given to the patient. A "False negative" then, is simply the doctor saying they are OK, and the ML confirming this, while both missing cancer. If doctor said, not OK, then no patient will take an OK from a website and be diagnosed too late. The last sentence is needlessly spiteful. I hope I misunderstood the intent of it.
> False positives are even worse, because they are far far far more likely to happen in practice. [...] I speak from experience that this WILL lead to the patient going to the doctor, and the doctor, seeing these very strong words, might want to take a biopsy to confirm the malignancy.
If doctor says OK, but my automated second opinion says not OK, then I go back to the doctor, and the doctor will either sooth my worries, explaining why the patch is benign. This is the desired outcome for my health. Or the doctor will second-guess their diagnosis, and perform, with my permission, more thorough tests. If tests come back OK, this will sooth my worries and the doctor's worries, and all will be fine. If tests come back not OK, then the application helped save my life.
There is zero reason to believe this app will bamboozle doctors into spending 1000s of dollars over nothing.
> If the tool is not for diagnostical purposes, it's not doing a whole lot in fighting cancer as it is just a play thing then.
This does not follow. There is a significant contribution with this app for fighting and raising awareness of cancer. Low-expertise hospitals use it as a diagnostics aid. People reported getting a second opinion from this app, returning to the hospital, and finding cancer early.
> At the moment, it's more like hindering cancer by taking resources from people who need them and giving them to people who don't need them.
Yes, it really seems you are deeming this app both severely dangerous and hindering cancer research. Without any solid (ML) reasoning and plenty of non-sequiturs. I'd suggest that even if you continue to hold such views, that you apologize to a fellow researcher.
Source: citizen scientist who open-sourced data analysis tools with disclaimers, now widely used for cancer screening, precision medicine, and improving chemo-therapy.
The words “found malicious mass”, disclaimer or not, could be considered a diagnosis. You can probably say “anomaly”, or maybe even “classified as malicious mass”. With you having connections in the US, it is probably worth talking to an expert if you want to keep this online.
Thank you for all your efforts and significant work. The intent and selfless dedication is moving. As grandma would say, talk does not cook rice. As a former student RN caring for patients at UCSF Neurosurgery & Neurology I witnessed a lot of research innovation on the "hardware" side when the "god" of last hope did grand rounds [1] One could only imagine something like this in the early 90s, software as another adjunct layer to the overall care approach. How can clinicians diagnose and treat if the patient has not been admitted into the point of service? Data driven approaches like yours would have saved a lot of $$$/RIP to increase awareness leading to early detection and care access. Pioneers shine a light for others to see, may detractors never have to face an unintended LP. [2]
You know, if people who teach ML classes at universities point out to you that what you're doing is going to be taught in classes as an example of "what not to do and how not to do it", you should probably immediately stop what you are doing and get opinions from experts in the field.
I will repeat what others are say: this is irresponible due to naivete and could be harmful to people. Please consult with experts on how to proceed.
and this is why healthcare is the #1 source of bankruptcy in the US.
Some people believe that every single person must have Mercedes-Benz type of care in the US.
They cannot fathom that some of the plebs (do they even exist for them?) may want to make their own independent healthcare choices, and are willing to accept the risk ... (or can only afford!) a Suzuki.
What is unethical about this citizen science project? What is ethical about keeping it only for yourself, and not sharing it with the world?
You are saying you have the expertise to build a similar product, but releasing it would mean the worst fear of your community?
> people will recklessly apply ML
What are the indications that this is a reckless application of ML?
> How many people went to this website, uploaded images, were told were ok and didn't follow up?
Common sense dictates exactly zero. Their follow up was taking their images and getting an automated second opinion. Either a doctor already deemed them OK, or a doctor deemed them not OK, in which case, they would not rely on a second opinion, to think they are suddenly OK.
> How many were told they have cancer and insisted on procedures they didn't need?
Again, exactly zero. The app returns probabilities not binary diagnostics. No hospital would do anti-cancer procedures on a patient without cancer, even when they insist, because some website, friend, or religious leader told them so.
> The website is totally unaccountable.
Good. Or make the good-faith open-source project accountable and liable? That would simply mean shutting it down. No more diagnostics help for low-expertise hospitals: not good at all.
> Totally without any of the most basic ethical standards in medicine.
List a basic ethical standard in medicine which this project runs afoul of.
> basic human rights for patients
What right is that? The right not to upload your images to a site of your choosing? I thought human rights include self-determination, and keeping possession of your imaging to do however you see fit.
> In the US this would have been shut down by the FDA immediately.
But is that a good, ethical thing? Or simply that red tape and authority in US does not allow for such projects?
> We should not be celebrating this unethical "science" that doesn't meet even the most basic of scientific standards or ethical standards.
You should not talk about ethics or science, when you did not do even a proper evaluation of the work of a fellow scientist.
> I can't believe this is getting upvoted here.
Awaiting your work on cancer research and ML. Post it here. If devoid of ethical issues, and strongly scientific, it will also be upvoted and celebrated. Or is your major contribution going to be a snipe at someone who actually contributed?
I share your sentiment, people are focusing on successes too much but not scrutinise what potential outcomes false negatives in software like this can have.
Admittedly I just skimmed the article but I feel like the title should be more - "This "amateur" medical...". His primary expertise is more physics/CS/programmer related than it is human biology/medicine.
At the time of this model / work developed (late 2017 / early 2018), very few public mammogram datasets were available. So the model was trained on DDSM/MIAS and tested on InBreast dataset. The 90% accuracy was calculated with results on InBreast dataset.
> As for why he doesn't commercialize the website and collect some money, this question was also asked by us.
You don't get into open source development for the big bucks. It's sort of like asking someone who starts a soup kitchen why they don't open a restaurant.
I think it's worth noting how open source development spurs technical innovation in a way commercial development does not. If a company develops a proprietary technology, almost nobody will get access to the source or datasets, and so it can't then be reused and built upon to advance the state of the art. If it's open source, everyone can build on it and improve things. Sort of like how science needs to be public in order to develop more science.
But just as we need large investments to push science further, we should also have large investments in open source. We need to make access to these 50-GPU clusters easier for open source research, so a lone developer doesn't have to built it himself. Some meager funding for the actual development would be nice too... How many OSS researchers could we fund at $35K per?
The 90% accuracy figure in the article unfortunately exposes the author as very ignorant.
On a German breast cancer screening population, 90% accuracy is abysmal as ~99.2% of cases are negative.
Just predicting „no cancer“ would achieve 99.2% accuracy.
Accuracy is a very bad metric for such highly asymmetric problems.
To provide some context:
The German screening system is able to identify cancer in ~6/1000 of patients screened, and missing it for the remaining 2/1000.
IIRC this is achieved by re-inviting in ~3% of cases to further examination, where ultrasound / needle biopsy / etc. can be done on a case by case basis.
Reading this article only makes me realise how crypto industry has crippled the progress independent researchers like this would badly need gpus for AI.
How many really useful, cool and meaningful projects are stuck because such authors can't find or afford gpus - as they are being used to calculate meaningless hashes instead :/
> In short, it is to let the AI help you "look at the film", and the accuracy rate is almost comparable to professional doctors, and it is completely free.
>> Furthermore, the accuracy of tumor identification has reached 90%.
How is this accuracy calculated? Further in the article it is noted that there is no patient data saved by the project:
>> He said that he’s not sure actually how many people have used it because the data is not saved on the server due to patient privacy concerns. But during that time, he received a lot of thank-you emails from patients, many of them from China.
Considering user privacy is laudable in my opinion, but I'm still curious to know how accuracy is known.
I’m currently working with Digistain (S21) and we’re using AI to predict breast (and eventually other) cancer recurrence.
The tests are performed using infrared spectroscopy to measure protein synthesis and then fed into AI in order to make proper measurements and predictions.
We’ve shown we’re able to predict better than any other known method and are beginning our partnership and rollout to many hospitals around the world.
I don't consider him an amateur programmer if he found that many GFX cards. Not amateur not "amateur" either. It's not the kind of thing that really gets called amateur.
I'm a consummate algorithmist and have difficulty getting that amount hardware...for now at least. And he got them all of the same kind, that's pro. coolwulf is in fact cool.
I don't understand why this is useful. The starting point is a mammogram. Are there a lot of people who are able to get a mammogram performed, but not able to get it analyzed?
I applaud the author. These tools seems like a great addition to health care providers. I'm just less sure about when you would use it directly as a patient.
1) Radiologists' experience on reading is not the same. The reading from a 5+ years experience radiologist sometime could be quite different from the one with a one or two years experience radiologist. As a matter of fact, due to the complicated issues with mammogram reading, some inexperienced radiologist don't get mammogram reading tasks when they started working in the clinics.
2) 2nd opinion sometime could be quite useful and important for the patients. By providing a utility for patients to have certain awareness on their mammogram, apart from the readings from their radiologists alone, could be useful, especially for people who are from remote area which lacks of experienced radiologists.
3) Even in big cities which has more medical resources (like big cities in China), due to the amount of patient, usually each radiologist only has about 30s to read a mammogram image (We have been told by multiple doctors and they could confirm that.). Mis-reading is very common due to the work-load. So a utility like this could help on finding or at lease warning doctors on possible missed.
There are other reasoning behind this but above are things on top of my head right now.
great story kudos to the author etc.
however from my knowledge reaching the 90 (although an amazing task) is not enough. I have been in contact with different researchers in similar domains and they were struggling for 96% and above. The difficulty in which you acquire those 1% after 90% is exponential and most hospitals/research facilities etc set a lower limit. 90% is by far not acceptable.
Furthermore as others have mentioned, security etc play always a bit part.
Having a doctor miss-read/miss the tumor etc can have legal actions etc. but having a machine do that is also not within the legal scope, there are so many aspects that need to be taken into consideration.
Great effort especially for the person, but science wise not much.
Very cool, but the desktop app for chest CT doesn't work - you can open an image but nothing get diagnosed.
How is the algorithm differ from existing image classification algorithms?
This guy is the real role model of our field (computer science). Not a Steve Job, not an Elon Musk. Hao Jiang (Coolwulf) is who we should all aspirate to be.
Fsck money already, we have the power to save live beings.
This is a regulated industry. If you don't jump through the necessary hoops, the US government will shut you down and there are many laws on the books they can use for this. There are many laws beyond HIPAA (which exists to make it easier to share data, not harder) which apply.
We may be talking about different things - I was referring to writing the software and having people upload their own data to it. You guys must be referring to more formal usage in the healthcare industry.
OP is probably talking about the legality of American hospitals using this software in an official capacity like some Chinese hospitals seem to be doing.
I'm completely unfamiliar, but it wouldn't surprise me if for diagnosing? software like this to be used in an official medical capacity in America it would need to go through some sort of particular vetting process because if it isn't it might leave hospitals who use it open to lawsuits.
that would be a potential YC idea. A company that enables smaller groups to do clinical research w/o needing an army of people to wade through the regulatory red tape. That also isn't in and of itself a giant predatory CRO type organization.
that's literally what pharma is now- companies that exist to help smaller groups get their research through the clinical and approval process. It would be hard to buidl that level of expertise in a smaller company.
I am David Sheryl, From Rosenberg Taxes. I want to say a Grateful thank you to Wizard Web Recovery after a difficult experience with fake hackers online. I have a friend who recommend me to Wizard Web Recovery, I contacted Wizard web Recovery and told him about My medical school grade and credit score, told me not to worry and gave me his word, I don’t know him, never met him before but makes me feel comfortable with his words and experience, He keeps to his word and upgrades my credit score, I had to share this as am sure that most people here, are looking for the right hacker and have not seen any yet. try and contact Wizard Web Recovery on any kind of service you may need from him as a hacker. it is a testimony, you have the right to give yours too, by helping others to know there is still a high rate of good hackers that are highly experienced, he responds on email and WhatsApp Number: +1 (917) 725-3296) Email: wizardwebrecovery@programmer.net) or (wizardwebrecoveryprogrammer@gmail.com)
He gives a listening ear to hear you out. Greetings to you all
Recently a Chinese media interviewed me and I talked about a few side projects I have done in the past. I talked about the Neuralrad Mammo Screening project and Neuralrad multiple brain mets SRS platform. More awareness on radiation therapy to the general public will greatly help the community and we believe Stereotactic Radiosurgery (SRS) will eventually replace majority of the whole brain radiation therapy (WBRT) in the next five years.
Here is the link to the original article: https://www.toutiao.com/article/7094940100450107935/
Super effort. I understand your reluctance to accept funding but if you ever change your mind on that be sure to publish it here on HN. If giving you more tools means more progress in this domain without the usual red tape then I'm all for giving you as much of a push as possible.
As a general thought, it's super easy to start a Patreon account to accept donations.
Seems pretty effective. :)
Oh, it's you!
What a beacon of light and inspiration you are. Thanks for your work.
That said, I welcome you to publish your work so it can become even better after a formalized peer-review process.
For the multiple brain mets SRS project, we will be presenting at this year's AAPM annual meeting.
What's the best way to contact you? I've been fighting brain cancer for 7 years and work closely with a group of neuro-oncologists, researchers, non-profits, and investors in the space.
I'd love to chat.
Feel free to send me emails: coolwulf@gmail.com
this is incredible. thanks.
1. do you have a patreon or something where people can support you?
2. do you know how to obtain the nucleotide sequences of c-MYC?
3. have you seen machine learning applications that can predict protein function? specifically, the goal is to predict if protein X could participate in cleaving of DNA similar to CRISPR Cas9? false positives are okay, provided the negatives are accurate (i.e., guarantees protein cannot cleave DNA like Cas9).
>do you know how to obtain the nucleotide sequences of c-MYC?
Check:
https://www.ensembl.org/Homo_sapiens/Transcript/Summary?db=c...
This is the main transcript, the remaining 9 are in the table
thanks so much! would it be possible to contact you for other questions? or could you recommend a reddit sub for questions like this? researchgate doesn't seem responsive.
for instance, how to obtain the nucleotide sequences for TP53, or for the NHN/RuvC domains of CRISPR Cas9?
thanks again for your help.
Re p53:
same site (ENSEMBL): https://www.ensembl.org/Homo_sapiens/Gene/Summary?db=core;g=...
also you can check human genes at GeneCards: https://www.genecards.org/
re other questions: The crucial thing is: what are your goals? Bioinformatics and cancer biology in particular are rather vast fields. Questions and answers make sense if you can place what you just got in some framework. And getting there is non-trivial, see i.e. +800pp book: https://www.amazon.com/Lewins-GENES-XII-Jocelyn-Krebs/dp/128...
thanks so much. genecards looks interesting.
is it possible to find nucleotide sequences from amplified c-MYC from cancerous tissue, either burkitt's lymphoma or TNBC?
ensembl is valuable, but it would be more efficient to hire someone to decode the different options. what type of researchers understand the different sequence options on an ensembl page: cancer biologist, geneticist, bioinformatics scientist?
thanks again for your help!
the goal is conducting some first-principles analysis of cancer, nothing fancy and very likely a waste of time.
if at this point in the game no one has advised you to eat organic, consider it
WBRT is pretty brutal. Am I right in thinking you are focusing on multiple site treatment/palliative treatment of metastatic presentations? High site count also or sticking to say < 5?
Exactly, I'm working on a workflow platform for multiple brain metastasis stereotactic radiosurgery. This will greatly benefit patients with more than 5 BMs.
Cool stuff! I've done some work in adjacent areas - there are huge challenges (not only technical) but great to feel you are making an impact.
ELI5, please.
Metastatic disease is when a cancer spreads to multiple locations. This can make it difficult/impossible/impractical to treat effectively, especially surgically.
Whole brain radiotherapy works by killing everything a little bit in the hopes that the tumors die first (e.g. like chemo). There are good reasonswhy this tends to mostly-sort-of be true, but getting the balance right is hard and too much dose will definitely cause other problems.
SRS is a way of targeting radiation directly to locations to kill cells, with less effective dose (hence damage) to other parts of the brain.
It's all pretty harsh stuff, and you can die from the necrotic tissue caused by it, also.
Often with this kind of disease you know you aren't going to cure someone, but you can get rid of symptoms and make people more comfortable (palliative care).
Where did you learn to program on distributed Nvidia GPUs? The article implied you were self taught and learning to do this is quite challenging for various reasons.
Not least, Nvidia's documentation is not the best resource to learn from. This seems like quite a lot of work to understand ML and write custom CUDA code to get this to work. Do you have any insight about how you taught yourself these things and what tools you use?
Not the OP but I taught myself in a couple of weeks picking apart some of the sample CUDA code and reading some of the (excellent) pdfs on the architecture of the Nvidia range. At the time the GTX285 was hot stuff, the same code runs unchanged on a 1080ti and I would expect it to continue to work on even more modern incarnations. CUDA is pretty good as a platform to build on if you understand the basic idea behind the engine, and some low level experience will come in very handy. And ML on CUDA vs ML in C or some other language is typically a matter of shuttling the data and the results back and forth between main memory and the card as well as implementing the most time consuming portions of the algorithm you are using in a custom kernel, you can usually get 50% or so of the theoretical maximum speed with relatively little effort. Getting to full speed is going to be a lot harder, but then you could of course also add another card (or another three) get get an instant boost.
Usually you would - nowadays at least - use someone else's optimized kernel + ML library but if you wanted to roll your own that's doable.
I'm not your OP but I learnt all these things at univeristy during my BCompSci. Understanding ML algorithms came down to a lot of math / statistics units. I learnt about parallel computing during a dedicated unit called "Distributed and Parallel Computing"
Did coolwulf write his own CUDA code, or use machine learning libraries that already come with CUDA implementations, like TensorFlow or PySpark?
Google Translate link - https://www-toutiao-com.translate.goog/article/7094940100450...
Thanks -- and here's the site as well -- http://mammo.neuralrad.com:5300/upload
Imagine for a moment if FB and Google took all their developers and resources and applied it to solving cancer, how long do you think it would take to make a dent in the problem space? Or is there a hard limit stopping us from progressive like Moores law for scientific understanding.
> Imagine for a moment if FB and Google took all their developers and resources...
It's a nice thought exercise, but of course not realistic. FAANGs wouldn't be able to attract or retain the talent if they were purely in scientific medical research (where do you think those fancy offices and fat paychecks comes from?).
Having said that, it doesn't have to be either/or. I strongly believe that companies can be successful in balancing between thriving commercial businesses and using its profits to advance research that can benefit the society at large. Case in point: Google's DeepMind has for years published groundbreaking work in various areas including breast cancer research [1], protein folding [2], energy savings and climate change [3], and more. Also Calico [4], or its investments in health startups via GV, and of course all the well known services to users.
What I'd like to see is more companies following a path where a portion of its profits is used to fund research in areas that can cause societal change for the better (which, in its turn, may eventually became new sources of revenue down the road - e.g. protein folding, self-driving cars).
(disclaimer: Googler here, but no relationship with any of the projects mentioned)
[1] https://www.deepmind.com/publications/international-evaluati...
[2] https://www.technologyreview.com/2021/07/22/1029973/deepmind...
[3] https://www.deepmind.com/blog/real-world-challenges-for-agi
[4] https://www.calicolabs.com/
Yea… but instead, ads are like 3% more effective each year. So, you know, silver linings.
Google does have a company
Calico (California Life Corporation)
(I’m hoping to work there after my PhD)
Sure if you are willing to ignore all evidence and instead go with a hopeful expectation of “doing good”.
Sir, We've all been hearing that AI will revolutionize medical diagnostics for a long time, seen startups come & go in this space but seeing your work impact lives on the ground has convinced me that the secret missing ingredient was a self-less human who can build an accessible(non-commercial) solution for the masses.
A decade ago I did a voluntary work on a simple app for oral cancer detection(questionnaire, data-entry) with an oncologist, Who used it to for survey in the tribal regions of India and he used to say how lack of early detection is the number one reason for so many deaths(many die without knowing that they had cancer).
He's now settled in Germany, But I would still pass this story to him & perhaps his colleagues in India could make use of it.
Thank you.
Thank you for all you've done for people, it's amazing and inspiring!
Your story was inspirational. It's really cool to run this project to help others without expecting any payment.
Hi. Some great projects. What's more commendable is your dedication towards your projects and seeing them through to end- to the point that they are actually useful. This is what I truly admire.
I have a question for you. What is the tech stack that you use?
And if it is not too much: What resources did you use to learn Deep Learning?
> Recently a Chinese media interviewed me and I talked about a few side projects I have done in the past.
I apologize if this has been asked and answered before, but do you speak Mandarin, or was the interview in English?
Asking out of curiosity if it’s the former, and if so, how difficult was it to learn whilst also working on this and other things? And are there any resources or tips you might share that you found helpful?
According to the article, Coolwulf went to Nanjing University for undergrad, so pretty safe to assume he would do an interview with Chinese media in Mandarin. And since he likely grew up in China, it was probably very easy for him to learn Mandarin indeed!
Great work. I'm inspired seeing others use their skills to work on something important. So many of the smartest, most well educated engineers of my generation have put their talents to use doing things that are of little value to society. Thank you for demonstrating to us what else is possible.
I am not a religious man at all but God Bless you. You are an amazing human being and a source of inspiration.
Unless it's more expensive than existing treatments the medical industry will close the circles around you excluding you.
That's why not one startup has hacked healthcare in America, not one. No breakaway successes making pharma cheaper. Like those incubators in Bangladesh, for premature babies not startups that is, those did OK. Some pill startups yes, but again that's an expensivification of medicine. If you can make medicine more expensive, they welcome you in!
Jim Clark tried this, he was on a roll after Silicon Graphics and Netscape. Huge roll about as strong as Elon Musk as a serial entrepreneur. Then he targeted healthcare and couldn't do shit, just couldn't get anything to happen. He literally talked about getting "rid of all the assholes" by which he meant insurance and doctors and hospitals and middlemen and pharma and all the other "assholes" of that nature in his own words, but leave "only one asshole in the middle--us [paraphrased]." It's in a book. That book also talks about guys going on airplanes and chasing goats off cliffs, saying "Some people do this."
Well the real structure of medicine isn't designed around the human body, it's designed around cornering the market. Market dominance. So of course it has this immune system against cost reduction and efficiencies--efficiencies especially--and you do know it lobbies, don't you? And can bribe the FDA like the Sacklers did? Or lobby the FDA, and then bribe underneath so when people see favoritism they think it's the over-the-counter placebo causing a placebo effect without suspecting an additional more potent under-the-table dosage of money. In case the administration has built up a tolerance to the over-the-counter stuff.
>That's why not one startup has hacked healthcare in America, not one.
This post makes a lot of points, but in general I think they boil down to the above statement : the belief that large, complex systems are just run by stupid and/or malicious people and that a sufficiently clever 'hack' will fix all the problems. I think that is an attitude that is common on HN, but wrong.
Most big problems are not technology problems, they are People problems with a capital P. Technology problems can be fixed with 'one simple hack they don't want you to know about!!!' People problems are complex and messy and cause and effect can be intermingled vertically and horizontally with other seemingly unrelated factors as well as temporally with things that don't even exist yet or used to exist but don't anymore!
The way we fix these messy, complex People problems is by respecting that they are real problems, that the people acting on those systems are (mostly) reasonable people just doing what reasonable people do, and slogging through solutions a day at a time with the oldest technology around - political power. These problems resolve if you can get enough people to agree they need solved.
I think you basically agreed with the GP.
What about the NHS in the UK? They should be more aligned to wanting cheaper cancer diagnosis and also anything to help people.
If America has dysfunctional healthcare there is still the rest of the world. Which might be good for Americans eventually as the tech will come across one way or another.
> If America has dysfunctional healthcare there is still the rest of the world. Which might be good for Americans eventually as the tech will come across one way or another.
Which raises the question - why does the vast majority of healthcare tech development come from the US? (I included "development" to get around the out-sourcing of testing to China.)
It’s hard to understand how much more available capital is in the states. Getting 100 million together is drastically easier there than anywhere else.
Similar to silicon valley’s tech scene, you get migration of people who want to work at a level where that sort of capital is required.
I was reacting to "Which might be good for Americans eventually as the tech will come across one way or another."
It is probably much much better for the wealthy and well insured, but on average not as good for all people. With that skew the high end is going to be better. That is my guess.
I believe it's too small a market.
Modern treatments are hugely expensive to develop, and they also tend to be very specific (ie. only 1 in 5,000 people might get the exact right kind of brain tumour for your treatment to be an option).
With only about 1 million all-cause deaths per year in the UK, that means your treatment for a specific terminal brain tumour might only have 200 patients per year.
The 'we just saved 4 hours of clinician time by using fancy AI' just isn't worth it if it only saves 800 hours of clinician time per year, yet costs millions to develop.
The fix is to roll this out somewhere with more patients (eg. China) and where trials are cheaper (ie. China).
Use rugged AI like coolwulf. That's what's up. He hauled ass, be like coolwulf.
No don't be like coolwulf, be coolwulf. Be him, do what he did, make the algorithms from your brain protect the algorithms that make up your brain. You know they're symbiotic right?
Transistor and neuron fighting together against destruction. Symbiosis.
EDIT: coolwulf might as well be John von Neumann, the last line of defense against brain cancer and brain surgery, those neurons reaching beyond the skull to make a change in the World for some part of that World to then reach inside that skull and protect it. The chip as the final garrison of the biological human brain. Ghost in the Shell. Ghost retreating from the Shell and regrouping, no ruling out whether it can triumph when it returns to reconquer its damaged Shell.
Shell game.
Money moves too many times, like the pebbles under the shells. So that's a deflection so people don't see who's serving whom. Every time someone thinks they can remember what's under a shell before it's shuffled, they move it really fast, and like with feints. And of course this doesn't work for real, hands aren't fast enough, so you gotta layer it with like something to injure their eye with a flash while they're looking, say a laser to the pupil, or maybe...
You could make a little trapdoor for the pebble provided the guy will not ask which shell the pebble was actually in. So deflect all questions.
But the end outcome is servitude comes from all over the world to an American banker, an American doctor, and an American executive. There's a few others. But the long tail of the American middle class? That was only necessary because those bankers, doctors and executives would never themselves (or with their sons) man the nuclear ballistic submarines, dogfight jets, aircraft carriers, CIA undercover agents, and especially infantrymen--never never never infantrymen. I've only heard of a single lonely guy becoming an enlisted man in a high-bullets-flying environment, when he could have gone anywhere--from Philips Exeter in fact. Anecdotal.
Yeah undercover agents get wrapped up and tortured, but infantrymen are bullet sponges. Riflemen, holy shit. WW2 riflemen, no let's not talk about that. At least spies there's a whole song and dance to the whole thing, part two of 1984, and speaking as someone who's stood up to torture, there's not the immediacy of the infantry where the brain surgery is much messier. Well on second thought it's not that different from getting rocks thrown at you, like I have this year...
Listen. So this is what's up. There's these guys, they eat all this work and don't do anything. Doctors don't do anything. They only use 5% of what they learned, the rest being a masochistic filter the elders imposed on the youngers to keep them from bringing them down to human reality with competition. So morally a doctor should only study for 1 year, not 20 (if not less, because of retention). That's five percent! Don't argue with that part, argue with something else!
And then plus surgeons, spesh brain surgeons which is what every medical student "wants to be when he grows up" as elder doctors put it, so the thing there is younger surgeons have much better manual dexterity than older surgeons, you can't do it past your fifties generally. What does that mean? Learn it faster! But no, they have to say "I studied for 20 years to tell you my bullshit" like come on, that's masochism. You spent 1 year learning worthwhile things and 19 years in the medical school "slammer" learning to be a good bitch, that's the whole game in medicine, hazing. Because otherwise people would say hey just pump out more doctors then the doctors have to say "nonono, you can't because it's soooooo expensive to make a doctor and America loses money [pretends to lose money] on training doctors and you have to be the absolute best of the best of the best to learn how to multiply 3-digit numbers".
It's so so stupid, especially because of this: it's all multiple choice. All of it. It's all multiple choice. No essays or explanations, just filling in black circles. I have contempt for using that form of testing alone because I'm so good at it with no effort. Like not even doing the reading. Show up "naked", a good enough mathematician should be able to learn from the test itself, from NOTHING, NOWHERE, and NOBODY. Destroy the integrity of these measurements. Get the test to leak the answers. Like I could do it in Chinese [I don't read Kanji, I can only make out Oracle Bones from 2600 BC] and be measurably better than a monkey. I don't generally brag about it because I have no respect for it, like not having respect for torture. But if that's the underlying reason some doctor claims to be better than me? If such a doctor says he's better than me, then let us agree we are not equal.
While the US health care system as a whole is mess, I don't think it's intractable if you can attack it from the right directions. GoodRX did with pharmaceuticals, and Mark Cuban seems to be doing a good job with CostPlus Drugs. (Valisure is also pretty interesting, doing in-house validation of generics).
Marty Makary wrote a book a couple years back, The Price We Pay on the wicked knot of a problem that is US health care. For those mildly interested, there was a Peter Attia Podcast inteview a while back that covers the gist of it: https://peterattiamd.com/martymakary/
For those that want to get some color from some of the biggest problems from a clinician/practioner's perpective, I found some of these podcast episodes to be pretty great/eye opening: https://zdoggmd.com/podcasts/
Super impressed. Amateurs can and do play their parts in science --- there are numerous discoveries made in astronomy, mathematics, and definitely in computer applied sciences that sprang from the minds of amateurs.
As a “professional” programmer, I’m humbled by your accomplishments. I really must find ways to contribute more to the world. It seems there’s a lot of opportunities in AI to do it.
You're clearly well accomplished in multiple areas. How do approach learning something new?
Thank you for doing this, you deserve a superhero badge! It's very inspirational.
You are a hero. God bless you and thank you for efforts to help others.
You are an inspiration. Thank you for restoring my faith in this field.
Sir, you are super cool.
Fantastic work dude. On behalf of anyone who might one day benefit thanks and congrats.
Nice job!
javascript only link. Any compatible link with noscript/basic (x)html browsers?
And I would like the content sent to me in Morse code via telegraph.
https://telegra.ph/This-amateur-programmer-once-used-50-1080...
There, it may content errors, since it's a machine translated copy what had to be telegraphed by the operator. Thank him, he's a brilliant guy.
In the UK the NHS don't do screening for breast cancer for under 50s because it's believed that it would do more harm than good by leading to unnecessary treatment for cancers that would never have actually caused any harm, and even where no treatment is carried out it causes great distress. Though there are arguments that the age cut off is too high, and should be set at 40.
The above is with regard to a well-funded and regulated screening program that presumably has much better precision/recall than this website. I wonder what the cut off age is for this website before the diagnoses cause more harm than good? 60? 70?
This is getting lots of upvotes because it's confirmation bias for the large segment of HN readers who believe that problems would easily be solved by a small number of brilliant technologists, if only it weren't for governments and big organisations with all their rules and regulations.
A lot of people, including myself, don’t believe that central health authorities have the right to make that call.
Moreover, I personally don’t have confidence in their ability to make those kinds of decisions, and I believe the abysmal performance of the NHS supports my view.
I think central authorities absolutely must make that call. Who else is going to decide how to dole out a scarce social resource? Americanizing healthcare is obviously not a good choice given how much worse it does overall by basically every measure (unless you're rich and don't give a toss about other people).
I certainly agree that central authorities can be better. But that's kind of a truism.
What alternative options do you have in mind? Admittedly I'm short on alternative ideas.
I understand why people don't want money to completely decide who gets health care. But I don't think it should be a complete non-factor either. If I've worked hard and saved diligently for an emergency, and an emergency arrives, I feel I should be able to tell a doctor "will you treat me? I have money".
But if you insist upon perfect equality of health care access, I guess it can't be. Some central authority has to decide if anyone is allowed to treat you, regardless of what you've saved for the eventuality.
Medical tourism will always be a thing for the rich. But I personally do not see an argument for driving a social service, even if only slightly, with capitalism.
Perhaps that’s part of my developing thought on the matter: the public system paid by tax dollars should be equitable to all. But by all means, take your credit card to Pepsi Presents: For Profit Medical Centre and get a full work up.
What makes the cut of a "social service"? Wouldn't it be anything people really need? Basic food staples, for example?
China tried socializing food production, and it worked terribly. Production tripled when they re-privatized it and let farmers grow for themselves. People do a better job when they get to keep the rewards.
US health care isn't capitalist in this sense of rewarding a good job. It's the worst of both worlds: a system whose regulations are superbly adapted to optimize profit for the administrative class at the expense of both doctors and patients.
This is not an interesting conversation honestly.
Some Americans think that the US Healthcare way is “the right way but Slightly Off-Track(tm)” and will not be suaded.
You asked a direct question, though, to what extent do social services exist.
Social services exist to ensure that we all have a decent foundation on which to conduct our business of living.
For some that will be as you say, providing basic food and housing. In fact in Sweden food is given to children for free in school; however in the UK it is only poor students that get it.
In Sweden water is free, in the UK it’s charged but it’s a utility that cannot be turned off.
Everyone draws the line somewhere else, but the basis is meant to be that we have a solid foundation.
Unexpected medical expenses shouldn’t decimate a household economy for a decade, it doesn’t matter how unprepared the household is. But this is my personal feeling.
In addition: tying health insurance to employment and having at-will working conditions strikes me as ensuring compliance/docility in the work place, which I don’t believe in.
> This is not an interesting conversation honestly.
"Please don't post shallow dismissals" - HN Guidelines https://news.ycombinator.com/newsguidelines.html
> Some Americans think that the US Healthcare way is “the right way but Slightly Off-Track(tm)” and will not be suaded.
"Be kind. Don't be snarky... Please don't sneer, including at the rest of the community." - HN Guidelines
1) it's not a shallow dismissal if neither position is interested in having a conversation.
2) it's not unkind to point out the truth, it's not possible to persuade people that the American healthcare system has a fundamental architectural problem, they believe wholeheartedly that it would work if it wasn't for the <insert consequential factor here>
it's not an intellectually honest conversation, it's shallow, everyone has dug in. it's boring.
You've replied to my post, but I honestly cannot tell who you are talking to, or whether your assertions about whoever you're talking about are in any way connected to anything I've said.
It could be that this style of engagement is why you're bored. It does seem boring to talk to people without actually talking to them.
You're talking about "keeping the fruits of labour", in the context of socialised healthcare, however it seems the majority of the world is doing perfectly fine with socialised healthcare. In fact it seems that less tax money is spent on healthcare in countries that just nationalise it.
That's the relation, that's why it's boring, because you'll never truly argue the point or consider another perspective. You can't be persuaded, and neither can I, reasonably.
So, what's the point.
Go live in your squalid hole and I'll go live in mine and lets not pretend that either of us aren't brainwashed in some way, because there's no possible way that we're going to conclude this discussion that I keep seeing repeated on every public forum ad infinitum.
it's fucking exhausting.
Why bother intruding into a discussion to announce how exhausted you are by it? It means nothing to us, and it's more effort for you. You could just ignore it and be less exhausted.
I'm basically telling you both to shut the fuck up because we're rehashing the same shit as forever.
Unless either of you are willing to engage in intellectually honest, curious conversation then you should honestly just keep your gob shut.
Flame wars help nothing, rehashing this helps nothing. We're walking in circles, save your energy, I'm wasting mine in an effort to save yours in future.
You're welcome.
I made a couple points and it seems like you think that, based on those points, you know much more about what I believe than I even know myself, and hold a great deal of contempt for me and whatever it is I apparently believe.
I would love to have an open and inquisitive conversation. I am open-minded about how we structure our system. There are many considerations, and costs and benefits to any method. I do not think socialized or public structures are bad or inferior to private ones.
I do think it helps for people to be rewarded for their labor, but it isn't the only consideration, and it can be worked out in a range of systems.
Even though quite cheap in most places, water is not necessarily free in Sweden - at least not according to the bill I receive every month...
(I believe if you knock on someone's door and ask for water you can't say no to that person - if within reason - and there are usually lots of public/free places to get water to drink)
That sounds more like a red herring, food is not a service, it's a product.
Administration bloat is a problem everywhere, however I don't think that is the reason for or against national Healthcare.
there is private healthcare available in the uk for those who want to pay. it isnt nhs or nothing.
I'm not saying otherwise. I was replying to a specific comment that I did not interpret to be referring to a specific real health system.
I share your disdain for a central authority in making these judgement calls, but I have even less confidence in the majority of people who think they can solve everything with AI. Signed, a data scientist with a PhD in biomedical engineering.
Misunderstanding here. The question is at what age is the base rate in the population sufficiently high that a test with a certain sensitivity and specificity useful net risks. Multiple studies have shown that these screening programmes don’t have a huge impact on mortality - excluding lead time bias etc. and noting the c. 10 in 100k extra cases of cancer caused by the screening. Here’s a review for prostate screening that’s even more damning about its usefulness - https://med-fom-urlgsci.sites.olt.ubc.ca/files/2007/06/P-ca-...
would it be possible to use MRI for screening?
MRI scans take longer, are often more uncomfortable for the patient (being in an enclosed, noisy machine), require much larger capital investment than an X-ray (and therefore supply is more limited), and have a higher cost per exam than an X-ray.
It almost certainly would be possible to use MRI for screening, but the impact would be a reduction in availability and a higher cost.
> A lot of people, including myself, don’t believe that central health authorities have the right to make that call.
That’s absolutely fine. You’re free to spend your money on preventative healthcare that’s been determined to fail a cost benefit test. Organizations that have to save as many lives as possible or prolong healthspan as long as possible within a limited budget must make decisions somehow.
The NHS has finite resources. They have to decide if implementing a screening program is worthwhile or not, versus spending the money elsewhere. You can still go to your doctor and get a mammogram (or even a more useful test) if you have other reasons why this might be justified in your situation. They aren’t ‘banning’ mammograms for young women.
It's not just a matter of finite resources.
Any sort of treatment is invasive. Almost all form of medical treatment has side effects and risks.
Finding out you have "cancer" is traumatic and extremely emotional, though breast cancer is one of the most survivable (in part because, well, everyone loves boobs. Prostate cancer, on the other hand...)
Putting these tools in the hands of medical professionals is one thing. Putting them in the hands of the general public is beyond irresponsible.
People physically assaulted doctors and nurses for not being given ivermectin; imagine how insufferable people will get when some website examined their mammogram and said they have cancer.
> ...breast cancer is one of the most survivable (in part because, well, everyone loves boobs. Prostate cancer, on the other hand...)
Extremely weird take. Not least because the survival rates for breast cancer and prostate cancer are very similar.
Yeah, just look at their disastrous handling of Covid for example.
the real problem is that we still don't have combined vaccines for all the new variants (+ influenza)
and it's completely ridiculous why there isn't a public/government project to do this.
This is absurd logic. If the next step for a test like is a procedure with a lot of risk, change the next step.
We need to be able to work in a world with frequent, imperfect, low cost diagnostic tools. Cancer is almost completely survivable if caught early enough. So working to figure out early detection is effectively the "cure" for cancer we have been looking for.
As I recall the actual logic was after decades of early & frequent screening, allegedly the data showed a greatly increased rate of mass discovery and treatment, but no improvement in mortality at the population level. This suggested despite a lot of activity, we may not have accomplished anything.
but that means that instead of treatment the correct next step after discovery is monitoring of those masses, no?
also it's quite possible that (just as with COVID tests) we would benefit from more testing even if that test is not that reliable. (so there's an argument for developing fast non-ionizing radiation imaging machines, eg. a fast stand-in MRI)
I'm always surprised with the "just close your eyes" attitude of medical policies. I mean, this way we essentially choose when to get women misdiagnosed? Isn't the solution to get a better idea of how common it is for non-cancerous masses to appear and adjust the risk predictions? Or to actually improve the diagnostic methods?
NHS recommendations: https://www.cancerresearchuk.org/about-cancer/breast-cancer/...
Anecdotally, I know (directly and indirectly) many, many women who had breast cancer before 50 years of age.
The NHS is massive bureaucracy. What it does or doesn’t do is peering inside the belly of a Byzantine whale.
The article is a misnomer calling him an ‘amateur’. Its a click bait title. He’s shown himself to a world leading researcher in the application of AI to cancer screening.
Plenty of managers in the NHS can’t even do simple math.
Your solution to anxiety derived from lack of medical knowledge is more bureaucracy between a patient and their healthcare provider?
I think the suggestion was more along the lines of “screening programmes have their place, but their variables should be optimised”
Thanks for the different perspective. What did you mean by "unnecessary treatment" though? If you have cancer, doesn't it need to be treated? Doesn't cancer anywhere always cause harm to the body?
Breast cancer for example is diagnosed by increasing levels of invasiveness. First a mammogram, then possibly a 3D mammogram, then an ultrasound, then a biopsy. There are possibilities for false positives all along this path and increasing levels of possible complications when performing procedures. If a false positive gets to a biopsy and you get an infection from it, you would not have ever gotten that infection if they didn't start testing you so young. False positives are very common with breast cancer screening.
Wait, how is an ultrasound more invasive than a mammogram, an X-ray based technique?
I guess they are physically more invasive - you have to undress and the gel has to be applied on the body part to get the ultrasound scan.
Not to mention the fact that getting a biopsy can cause the cancer to spread all over the body where it might never have grown beyond its original position had it been left untouched.
Highly contested, and there is no evidence. Link to this published research: https://gut.bmj.com/content/64/7/1105.abstract
For many detectable tumors, the best answer is "wait and see", not "immediately remove". There are many reasons for this- surgery itself is risky, the tumor itself might not ever become harmful.
See for example https://jamanetwork.com/journals/jamaoncology/fullarticle/27... for some more context.
1) I just downloaded the "The Mammographic Image Analysis Society database of digital mammograms" [0] and ran it against the tool [1] image by image. Results below, code here [2]:
2) How is it true when the site [1] says "We will not store your data on our server. Please don't worry about any privacy issues." when you can find all analyzed mammograms under the "static" directory?http://mammo.neuralrad.com:5300/static/mamo.jpg
http://mammo.neuralrad.com:5300/static/mammo.jpg
(trying file names at random)
[0] https://www.repository.cam.ac.uk/handle/1810/250394
[1] http://mammo.neuralrad.com:5300/upload
[2] https://github.com/gregsadetsky/mias-check
Thank you for your efforts for validation and I appreciate that. There is a script running in the background to auto clean the files in static folder every day.
You just admitted you do store images.
Also, you're serving up on http. Don't do that.
They cache images.
I think caching people's sensitive medical history in a way that is publicly accessible is worth calling out; especially as it could be trivially fixed to maintain privacy.
True, it definitely shouldn't be accessible to others.
caching is temporary storage, and as shown elsewhere, they're serving up the site on http with other user's data visible through guessing.
This is very cool work. I'm a radiologist, I also work on developing ML/AI based systems for cancer detection and characterization. Literally just took a break for a few minutes from creating some labels and saw this as the top HN post!
I think in some ways making the model available online can be good, but in other ways could be harmful too. Very complicated topic.
恭喜coolwulf， 祝你继续成功。
I've always felt the "could be harmful" was a rationalization by radiologists worried about their job security since it's easily mitigated with a warning and multiple tests.
And especially because in the future, most radiology work will be done by software. It's just a matter of whether it's 10 years or 100 years from now.
> I've always felt the "could be harmful" was a rationalization by radiologists worried about their job security
Surely concern for the well-being of the patient figures in there somewhere...
Or imagine this: A liver lesion is incidentally discovered on your abdominal CT performed for unrelated reasons. Its radiographic characteristics are equivocal. Additional imaging studies fail to completely exclude the possibility of a liver malignancy. You undergo a biopsy. But the biopsy is complicated by hemorrhage. Surgery is required. You develop a post-operative nosocomial infection. etc. etc.
To the extent that risks along this chain of unfortunate events is known, yes, warnings could put some of the quantified decision making in the patient's hands. Well, except for the rampant innumeracy in the general population...
Perhaps a better example is prostate biopsies. They have a significant risk of producing incontinence or impotence.
This is a real scenario that happens regularly.
> "could be harmful" was a rationalization
This topic (breast screening) is a good example due to the sheer scale. If you increase the work-up rate by even a smallish amount, you are statistically pretty much guaranteed to kill people who did not have the disease. How this balances about gain (i.e. save other lives) is not obvious. Figuring out the "right" way to do this is real work.
Also a good example because there's actually been recent push back on mammography because the risks may outweigh the rewards (especially in younger age brackets). More testing does not automatically equal better outcomes. Here's one summary from the Nordic Cochrane Center (for those that don't know Cochrane is one of the gold standard organizations for evaluating the quality of biomedical research via systematic reviews https://en.wikipedia.org/wiki/Cochrane_(organisation) ):
Gøtzsche PC. Mammography screening is harmful and should be abandoned. J R Soc Med. 2015;108(9):341-345. doi:10.1177/0141076815602452 https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4582264/
Extremely important, and is the most important question to ask in any screening scenario.
It's not the case the more screening is always better. There are plenty of screening regimes that you've never heard about because the trial data didn't support it.
That’s not true. I would love to have someone check my work. It’s just got so many false positives. Would you trust the radiologist if a report shows a suspicious lesion marked by the software that the radiologist dismisses? CAD in mammography has been there since the 90s. rads don’t put up a fight. Do you want a whole FSD/Autopilot level of blame absolution if software missed a cancer/critical finding? I sure as hell won’t be apart of that.
Yeah, I've read all the arguments about harm, and I just don't buy it. I'd get a CT scan every year if I could convince someone to sign off on it. There is absolutely no substitute for knowing.
> Yeah, I've read all the arguments about harm, and I just don't buy it.
"I have no medical background whatsoever but I do not buy undisputed medical science."
> I'd get a CT scan every year if I could convince someone to sign off on it
20mSv of unnecessary ionizing radiation exposure a year, what could possibly go wrong? That whole-body CT scan is equivalent to getting at least EIGHTY chest x-rays, and ten times what a uranium miner receives in a year, and well within the range where an increased risk of cancer is noticeable in epidemiological data.
Case and point why this tool should not be generally available.
> I'd get a CT scan every year
You would choose annual body CT even if outcomes-based evidence showed no benefit?
Even among those at highest risk for lung carcinoma, studies of annual screening chest CT do not uniformly show improved disease-specific survival as compared to an unscreened but otherwise matched cohort.
https://www.nejm.org/doi/full/10.1056/nejmoa1102873
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3762603/
What I found mind blowing when I first read these studies (particularly the second link) is that the positive-predictive value of the CT screen was only 3.8% for lung nodules that were 4mm or larger in size. Basically just means we find all kinds of lung abnormalities all the time that "could be cancer" but aren't.
Keep in mind, that the screening criteria was people over age 65 who had a 30+ pack-year smoking history (1 pack-year = smoking 1 pack of cigarettes per day in one year), and had to have smoked within the last 15 years.
I can also tell you from doing CT lung biopsies that it's not a procedure to take lightly. Lot of important structures in the chest.
Absolutely. Those studies show no improved survival across cohorts. Not for individuals. My only chance of living is to know. That's worth it for me.
I mean, plus I have extreme health anxiety, and I've had doctors say "nah you're too young, we don't need to test for that" and later it turned out I did have kidney failure at 35.
You’d dramatically raise your risk of cancer from x-ray exposure for almost zero clinical benefit.
Yeah, that is the only concern. I guess whole-body MR would be fine too. I'd be willing to lay in the tube for 20 hours a year to know that I'm at least structurally normal still.
> Yeah, that is the only concern.
It really isn't. Your chances of negative consequence for unnecessary follow up procedures would rapidly become significant.
Until you've had an unnecessary biopsy that kills you :)
How could it be harmful? Is it just because of errors, i.e. false negative means person won't talk to doctor and won't be caught early; false positive means needless worry, maybe tarnish more general medical industry? Or something else?
Yeah, those are part of the issue. For example on the webpage, there is no discussion about the false negative/positive rate of the system, and no comparison or link to an academic paper that compares the performance of the system to a trained breast radiologist.
Does a patient take a negative exam as "I'm good!" and forego actually seeing a doctor, or having the exam read by a radiologist?
A positive result is also a challenging situation. I've talked to patients before after diagnosing them, and as you could expect most people are shocked, scared or they want to know what they should do next.
In breast radiology specifically (at least in the US) there is a well defined reporting lexicon and classification system. From what I can tell, the system does not use the lexicon or BI-RADS classification. https://www.acr.org/-/media/ACR/Files/RADS/BI-RADS/BIRADS-Re...
Breast cancer can be very subtle, and can be hard to figure out, especially if a patient has had surgery or other benign abnormalities. https://radiopaedia.org/cases/development-of-dcis?case_id=de...
Diagnostic and screening images also undergo rigorous QA for the entire system. Does the model have some kind of QA built in?
There is a just a small disclaimer at the top that says "we will not store your data on the server. please do not worry about privacy issues". Not exactly a formal or legal binding agreement.
In fact, you can't even connect over https. http://mammo.neuralrad.com:5300/upload
These are just a few things that come to mind off the top of my head. Like I said before, I do think that what coolwulf works on and is trying to do is good, and in the long run can help doctors better characterize findings and help patients. But like anything in health care, there are a lot of edge cases and side effects that you have to think about. The stakes are also very high.
I think care with messaging positive/negative rate is the right reasoning for a productized version of this positioned as an alternative to human radiologists. The difference here is that we know the user has a scan of their body and we can be pretty sure they couldn’t have gotten the scan without doctors already being involved.
So we don’t have to worry as much about the “Does a patient take a negative exam as ‘I'm good!’”, because in the normal flow of medicine that gets this scan the doctor will check the scan for them either way. The website probably could have better explanations of this, but most likely negatives will already be double checked independently by doctors, and positives will most likely be handled correctly by bringing it to their current current attention to double check the scan or run more tests.
To be fair, how is a patient going to get a mammogram or xray without having seen a doctor?
> Does a patient take a negative exam as "I'm good!" and forego actually seeing a doctor, or having the exam read by a radiologist?
I wonder how the false positive/false negative rates for this tool compare to that of a trained human. At some point we may reach parity, in which case what's the harm of trusting the automated result?
Secondary follow up.
I don't think that a human reader in the loop will be going away for a long time. Sure, they can be at parity, but does a combined read have better results? This is an active research question.
I think what will happen is that over time, human interpreters and AI systems will "co-evolve" in a sense, where people will pick up on where models are wrong and also learn how to use models to understand their own blind spots. These are also active research topics that are in their infancy.
https://www.rsna.org/news/2022/march/AI-Potential-in-Breast-...
EDIT: To be clear, I believe that this is a different group of authors from coolwulf. Pubmed link:
https://pubmed.ncbi.nlm.nih.gov/35348377/
The consequences are false positives and negatives in cancer detection are well known (easily googled), and are actually extremely important when looking at cancer treatment in large populations. You nailed the most important ones.
Well this radiologist could lose their job to an algorithm. That's pretty harmful to the industry.
I'm not worried. If things are that different, the rest of the economy will be too. The job will also evolve over time, like any other profession that changes with new technology.
It's our job (as radiologists and engineers) to shape that in a way that benefits patients.
If AI automates radiology, it will automate the rest of the world first.
>Very complicated topic.
This is very true. Data availability (and moreso, label availability) is the biggest barrier to improvement here I suspect [ thanks for labeling!]. Access being another. Using a public site to bootstrap that could do very interesting things. On the other hand, public access to a poorly RA/QA'd algorithm could also cause more trouble than help, easily.
Label availability is a big problem. There are some academic collaborations for some diseases which is making this better, but datasets in general are minuscule compared to what is available for more general computer vision applications.
I am hopeful though. Few-shot learning and self supervision are very active questions right now and there are a lot of papers in the medical AI field that are getting published on these topics.
I'm personally interested in liver cancer, which does not have large, well-curated and shared database of cases.
Sharing data gets tricky, especially in the US. Labels I'm working with and creating (at least at this point) are for my own research, which won't be shared publicly any time soon.
> n the medical AI field that are getting published on these topics.
Most of what I've seen isn't very promising. The energy in these research areas are because it would be so much cheaper than the "right" way, far more than because of the likelihood of success. And also perversely, because it's hard for the academic researchers to get enough data to do other studies :) NB: I'm not saying there is nothing useful coming out of the learning literature in last few years, just that it a) isn't a silver bullet and b) is often being misapplied in these areas anyway.
Label quality and availability isn't the only big problem though. Many data sets exhibit problematic sampling bias, as well as being order(s) of magnitude too small, because of the way they are gathered and how access is granted.
Problem is though for most of these diseases there just aren't the number of samples available, period, to do it the "right way". HCC for example, has around 50K new cases/year in the US. Even if every single case went into a repository with perfect labels, would still take a long time to collect that info. Not to mention you need either a radiologist (4 year of medical school + 6 years of post-school training) or a very skilled and experienced technician to label the data.
Not to mention imaging protocols are not standardized, and the imaging technology is also evolving so scans we do today may not be "correct" or standard in 5-10 years.
Definitely diseases have different challenges. Breast cancer screening being a notable outlier as far as data availabilty. For some diseases ML is probably always going to be problematic although may help in diagnostics mostly by helping get rid of other possibilities.
I suspect we have similar overall views of the problem, but I'm pretty strongly in camp that recent advances in ML/AI are mostly really driven by data & label availability, not algorithmic advances - this colors where I think the wins to be had in medical ML can happen most easily. Either way though the non-technical barriers seem clearly higher than the technical ones still.
Really enjoyed your contributions to this thread, thanks.
I’m a first year radiology registrar (PGY3) in Australia looking to find others doing interesting work in this domain, if you think I could help with your efforts feel free to DM
Second follow up.
Results are not impressive until they are :)
It's certainly not a solved problem, and it's easy to have a pessimistic view now but I'm generally bullish on where things will be 10 years from now.
> Results are not impressive until they are :)
True! I certainly wouldn't discourage anyone from trying.
On the other hand, I think it would be a huge mistake to trust that fancy learning approaches will solve everything so we shouldn't try and improve access and labling. Getting better there is still by far the most high probability of successful impact, imo.
What peer review or regulatory approval process has this been through? Seems pretty irresponsible -- there are many notorious pitfalls encountered with ML for medical imaging. You shouldn't play with people's lives.
I don't understand why this comment is downvoted. Automated screening of radiological images by means of neural net is an extensively researched topic. Ten years ago there had been predictions that such automated screening will displace the radiologists, but that clearly did not happen.
For instance, this article is silent on false positive/false negative rates of the software. There is no comparison with other research. It reads like a corporate press release promoting a product.
The quotations around "amateur" should be moved to "fought". Nonetheless, it's encouraging that this level of research can be executed at home, however the strict burden of proof required should still be maintained.
Well, he was director of R&D for a medical imaging company and worked directly with academia. So I think its appropriate to contextualise the "amateur" because his work looked much more amateur than it really was.
This is an incredibly important point. Medical research must be taken seriously and I see many problems with the processes being applied here.
(for those who care- I'm a published ml biologist who works for a pharma that develops human health products. Having worked in this area for some time, I often see people who have no real idea of how the medical establishment works, or how diagnostics are marketed/sold/regulated. Overconfidence by naive individuals can have massive negative outcomes.
Why does everyone assume this guy has zero business attempting this? If you read his credentials, he should be every bit as qualified as you to attempt this kind of work while understanding the pitfalls.
According to his CV he's been active in the field for quite some time. The default assumption that he's an idiot and going to kill people just seems too cynical here.
Grandparent - you specifically mention having noted methodology problems, would you mind sharing where in the methodology you think he's gone wrong?
I'm not going to detail challenges in medical ML - the literature can do that. But just to mention one that other people haven't that goes beyond just precision-recall: algorithms can be biased based on variance in physiology (e.g. more accurate for men), and understanding how an algorithm is biased is very important for the person interpeting the information, who should be a trained doctor.
And it doesn't matter what his credentials are, that's appeal to authority. If he thinks this should be used and trusted by people for decision making, then he should submit it to independent peer review and regulatory approval.
This is like taking your temperature at home, are you making a diagnostic yourself? Not quite. But you can know some symptoms and take action (going to the doctor) maybe with less anxiety
edit: grammar
Thermometers are well understood, simple devices, and there are other complementary checks (e.g. does my forehead feel hot) if they fail.
This project might lead to people thinking they're in the clear and not seek appropriate medical treatment, or be overtreated due to an error. You should always talk to a qualified doctor if you're concerned about your health, and not use projects like these for decision making.
I don't see how a passive scan like this can be harmful . Ofcourse if it does show a positive one should confirm further with a biopsy or other standard confirmation diagnostic. As for false negatives... If you feel something is wrong you should get it checked thoroughly anyway .
you literally just said the scan has no value. THe point of a scan like this is to have absurdly low false negative and positive rates so that it's actionable. Unactionable medical diagnostics are worthless and just cost (money, fuel, time).
There are more costs, like others have described it might lead to unnecessary biopty that has real risks.
So you should ignore it and see a doctor regardless.
In an ideal world, that would be the case. However, people aren't 100% rational agents motivated by logic.
My aunt was diagnosed with Multiple Sclerosis which led her to lose her eyesight from her left eye, because she refused to get a thorough checkup by a professional, and even today whenever I tell her that I visited the doctor for an issue she has... not very good words to say (something something "you are a chicken, you're hypochodriac etc). And I'm saying this without entertaining the probability of her visiting a professional who just happened to be in a bad day, which could potentialy lead to a wrong diagnosis.
I've been hunting down my own diagnosis for symptoms everyone seems to tell me that aren't serious (nail discoloration and a 24/7 headache that feels like my arteries are pulsing, which lasts for YEARS, cold fingertips during the winter, and more).
I get what you're trying to say and I agree with the general message. However, more checkpoints to catch a potential failure are good. For example, if someone were to make a take-at-home device which scans nailfold capillaries (no reason for something like that to exist) I'd get that in a heartbeat. I'm being actively ignored by every medical professional that I have visited, and if I'm not ignored they give the minimum amount of attention, kind of like "well, it's not like you're dying so who cares?"
Fair note: I'm from Europe.
If you don't feel good yes you should see a doctor regardless. Let's say you get a scan every year . If the scan is able to detect something earlier than a radiologist is able to identify i think it's worth paying attention to.
I agree with you, but what's most important is the impression that the average person who uses it will have. And I don't think most people would think of this as like "taking your temperature at home". I think most people who might upload their x-ray scans would take this a lot more seriously.
A false positive could create a lot of anxiety and emotional distress, and the patient might need to go to 2, 3, or 4 other doctors to get second opinions before they feel comfortable that they really don't have cancer.
A false negative could be even worse. A patient might think "oh, the official-looking online thing said I don't have cancer, so I don't need to wait for or consider a human radiologist's results", and not believe they need treatment.
I think it's very important that people understand that -- until more research is done -- this is still not a substitute for having a human look at your x-rays. If we could be reasonably sure that everyone (or at least a very large majority) understood this when using this tool, then I think people would have far fewer objections. But I don't think that's the case.
Having said that, I think it's safe to assume that this tool has saved lives, so it's almost certainly been a net positive for people.
A thermometer mostly tells you if you have an infection, and how close you are to your body temperature becoming a medically urgent or life-threatening situation.
Not even remotely the same thing.
What's the chance of taking the wrong temperature though?
Pretty high, actually.
In addition, these kinds of things will still miss lobular ("normal" cancers are ductal) breast cancers as they don't form lumps.
15% of the women with breast cancer are waiting for a non-invasize diagnostic imaging system that can see their cancer. The only thing that can see these is an MRI with gadolinium. And that gadolinium contrast causes issues in about 1 in 1000 women, so it can't be used as a general screen.
If your decision making process is a negative result tells you nothing and a positive result warrants immediate follow up, what’s the risk here? I’m assuming doctors recommending that women get checked for breast cancer is the primary breast cancer is tested and diagnosed which presumably wouldn’t change because someone make a website.
In this case I feel better about it because there is a natural limitation in that most people doing this will only have the scan because they are getting tested through a real clinical process. So effectively they are getting "standard of care" treatment implicitly, and all this does is accelerate their response to true positives. The worst case scenario is a false positive gives them a lot of anxiety / costs them money through trying to accelerate their real diagnosis only to find it isn't real.
Right. Let's let the pharma and medical establishment continue to "innovate" on our behalf.
Ahh yes, why would we want to give poor people a potential route to improve their health? it would definitely be more ethical to let them die.
ETH will soon move from PoW to PoS (let's not debate the timeline or if it is a good idea). This will put about 32 million GPUs worth of compute and millions of CPUs searching for something else to do (or just flood the market with used equipment).
I have been searching, for years, for alternative workloads for these GPUs beyond just PoW mining and password cracking. Many of them are on systems with tiny cpus, little memory, little disk, little networking so the options are heavily limited. AI/ML/Rendering/Gaming actually make bad use cases.
If anyone has thoughts on this, I'd appreciate hearing them. Let it all die is certainly an option, but it also seems just as wasteful as keeping it going. Maybe we can find a better use case, like somehow curing cancer...
> ETH will soon move from PoW to PoS (let's not debate the timeline or if it is a good idea). This will put about 32 million GPUs worth of compute and millions of CPUs searching for something else to do (or just flood the market with used equipment).
Crypto markets crashing together could do this, but ETH's switch isn't going to do much for old cards.
Checking https://whattomine.com/ shows that ETH mining isn't even in the top 5 most profitable things to mine with a 1080Ti right now. The miners looking to squeeze every bit of profitability out of old hardware switched away from ETH a long time ago.
Wrong. Why?
1) The cards have already paid for themselves. They are 100% ROI positive and even at the current low amounts very profitable. Regardless of what W2M says, ETH is still the top most profitable coin. Large miners don't sell immediately, they wait for the market to go up or the option against their ETH holdings.
2) ETH doesn't require latest hardware because the algo is memory hard, which means that the bottle neck is in the memory controller, not in the speed of the GPU chip itself. https://www.vijaypradeep.com/blog/2017-04-28-ethereums-memor...
3) The actual consumable is electricity price, which really hasn't changed much in the last few years for large miners who have contracts.
> Wrong. Why?
What? ETH isn’t the most profitable thing for these cards to mine. Why are you so confident that ETH switching to PoS will open the floodgates when ETH is already not the driver of their usage?
The ETH proof-of-stake switch isn’t going to be the driving force they causes GPUs to be abandoned in bulk.
> ETH isn’t the most profitable thing for these cards to mine.
You can mine shitcoins on a small scale for maybe a bit more profit. The issue is always unloading a lot of shitcoin in a very small volume market. Even ETC has a 6.5 day deposit time on exchanges like Kraken. Nobody is going to want to deal with that.
> The ETH proof-of-stake switch isn’t going to be the driving force they causes GPUs to be abandoned in bulk.
The GPUs will move to other shitcoins, which will cause the prices of those shitcoins to go down as people will mine and dump them, which will cause a lot of GPU miners to shut off.
Usually, with POW, the issuance is the same. So they will mint the same amount, no way that will crash the price. What _will_ make the mining not profitable is the difficulty increase. It boggles my mind people don't seem to get this.
More hashrate will earn more of the pie.
> the ETH proof-of-stake switch isn’t going to be the driving force they causes GPUs to be abandoned in bulk.
The proof-of-stake switch is also pretty unlikely to ever actually happen.
The merge will absolutely happen, there are 12.6m ETH in the staking contract [1] that people want out of there and the only way to get it is to get to the next steps.
[1] https://etherscan.io/address/0x00000000219ab540356cBB839Cbe0...
Sorry but I'll believe it when it happens.
When is irrelevant in my eyes, it honestly doesn't matter. It will happen though.
I hear people say this, but I am absolutely certain their assumptions are wrong.
1) The sum total of rewards is fixed for POW 2) Introducing extra hashing power will increase the difficulties of these mining ops up to the profitability equilibrium point.
After the overall "free" hashing power increases to a point, GPU's will start flooding the market at dumping prices.
It will be incredibly rad!
Proof of Stake has been 6-18months away for 5 years now.
As far as I'm concerned it'll release along with Star Citizen
There are plenty of good uses, projects like BOINC have been using GPUs for good for over a decade. The problem is the incentive system disappears, it's a lot easier to get people to run 32 million GPUs when it makes them money instead of costs them money.
Well, exactly. It has to be incentivized.
Please please check gridcoin.world
This seems like an incentivized BOINC built on top of an inflationary shitcoin. There is no utility in the coin itself and that is reflected in its price history.
What we need is something that has utility... like run BOINC, earn tokens that can be used in the real world for something other than just dumping on the market.
Folding@home would love to take a swing at a sliver of that compute
It's peculiar to me that Folding@home never managed to get GPUs working for MacOS, given the platform's popularity: https://foldingforum.org/viewtopic.php?f=83&t=32895
As a consequence, it seems not worth installing on MacOS: https://stats.foldingathome.org/os
It's not worth implementing for MacOS because it woulnd't increase the overall folding rate very much given the relative weakness of the Mac GPUs compared to the existing f@h fleet.
folding@home has been doing this for 20+ years. They already did all the smart research and tech development. Just use that until somebody comes up with a workable DrugDiscoveryAtHome or CureCancerAtHome.
This is a very dismissive answer which seems odd coming from someone with a lot of karma. Running GPUs at scale isn't easy or cheap.
There is no incentive to run this other than good feelings. Unfortunately, that isn't enough in the business world to spend millions on cap/opx.
What I'm looking for is incentivized options. Even better if they come from a web3 situation where a business can operate without having actual customers.
"Mining", but with not such "wasteful" work.
It's ironic to be declaring the GP as dismissive while you're flippantly dismissing the work of everyone involved with F@H. You're conveniently ignoring the incentive of expanding our understanding of biology, which has very real applications and results[1] that benefit the entire world for the rest of human existence, instead of benefiting a single participant in the short term.
[1] https://www.hpcwire.com/2020/10/14/how-foldinghome-identifie...
What? I'm not dismissing their work.
I'm just saying there is no direct financial incentive for 32 million GPUs to move over to it. If there was, they'd be on that instead of ETH.
You literally said "no incentive [...] other than good feelings." Regardless, even adding the words "direct financial" is still incorrect. For one of many examples, insurance companies investing in F@H infrastructure would be beneficiaries, albeit via savings and not revenue.
I literally clarified my statement since you came to an incorrect conclusion.
"direct financial" is correct when you're giving examples that are not direct.
You're moving the goal posts, on top of changing the definition of "direct" to fit your false narrative. First you claim there was no incentive at all, then there was no "direct" incentive, and now direct apparently means "immediate" instead of linear.
If a company invests in research specifically because the fruits of that research will reduce expenses, that is direct financial benefit.
> Reducing expenses is not direct financial benefit.
You've created a seven word sentence which contradicts itself in every conceivable way. Expenses are financial. Reducing them is beneficial.
>Expenses are financial. Reducing them is beneficial.
You're missing the word 'direct'.
Expenses are the "loss" portion of "profit and loss." That is 50% of the concept of accounting, making it as direct as it gets. Please stop this prolonged line of trolling.
The dark side is you need a ponzi scheme to fund it. When that ponzi collapses the work stops. Good feelings are better IMO!
It is odd that you think of crypto as a ponzi while promoting a referral network in your HN profile.
Yes Digital Ocean. They are not a ponzi/pyramid because
1. They sell a useful product
2. They do not charge people to be affiliates.
3. They do not sell a token/security that you buy with the intent of making future gains, funded only by other people buying the token expecting more people will do the same.
If anyone uses the link ... I haven't checked ... I get some DO credit, not even money, which I will use for hobby projects. If I get $5 in credit in a year, I would be pretty suprised.
I'm not saying it is a pyramid/ponzi. It is an affiliate marketing scheme, and those are never slimy. /s
So any job in sales is slimey.
yeah, everybody knows crypto is a pyramid scheme, ponzi schemes are specifically about a type of coupon fraud.
It is structurally more like a ponzi than a pyramid.
By large crypto has no MLM style referral system where the product is the business opportunity to sell the product.
However it does have a feature where crypto gains are only funded by newer people coming in and buying at a higher price. But no cashflows are produced other than selling the investment tokens. This is almost identical to a ponzi except it has decentralisation that makes it probably legally not a ponzi.
Imagine I did an IPO of a company I incorporated last night but hid behind a panama shell. That’s crypto.
> However it does have a feature where crypto gains are only funded by newer people coming in and buying at a higher price.
"only" is inaccurate. This was true early on before there were markets and contracts. Now we have DeFi which allows for gains on trades by anyone who wants to provide liquidity. Similar to any sort of commodity market where people buy and sell what they own (or even don't own with margin and options).
> This is a very dismissive answer which seems odd coming from someone with a lot of karma
Are you kidding? Shallow dismissals masquerading as insight are easily the most popular genre of comment on Hacker News.
only good feelings? f@H has made fundamental contributions to our understanding of biophysics of folding proteins!
I am aware of domain applications from F@H, but not core science work. What are those fundamental contributions?
We conclusively demonstrated that kinetic models of folding are critical to do better drug discovery against GPCRs and other target classes.
Or did you mean something more fundamental, like "the biophysics of protein folding is primarily determined by entropic-driven hydrophobic collapse, not enthalpic contributions from hydrogen bonding?"
>We conclusively demonstrated that kinetic models of folding are critical to do better drug discovery against GPCRs and other target classes.
How was this demonstration done? Or are you referring to building a better high throughput screening mechanism? I'm by no means a protein expert, but I am trying to learn more about them.
>Or did you mean something more fundamental, like "the biophysics of protein folding is primarily determined by entropic-driven hydrophobic collapse, not enthalpic contributions from hydrogen bonding?"
I don't want to minimize anything or anyones work. But conformation of experimental observations via molecular dynamics simulation isn't what I was thinking of - but maybe I am misunderstanding your claim. Its my own ignorance.
Sorry, it was just a way to say "not financially incentivized".
Oh you know what an alternative use is? Oaths. Works with old ASICs as well...well I think. So you take a document, like this comment, you append a nonce (you'll see) and you hash it until you get a lot of zeroes in the front. Same as bitcoin, but you're not hashing the bitcoin protocol. Then, you know the document has been sworn, as a cryptographic oath, to that extent. Nonce: 38943
So then go to https://geraintluff.github.io/sha256/
> oath = "Oh you know what an alternative use is? Oaths. Works with old ASICs as well...well I think. So you take a document, like this comment, you append a nonce (you'll see) and you hash it until you get a lot of zeroes in the front. Same as bitcoin, but you're not hashing the bitcoin protocol. Then, you know the document has been sworn, as a cryptographic oath, to that extent. Nonce: 38943"
> sha256(oath) 00009ea9ab415b7f60cd43571c159d1bf1e01de4bae6a706ec9053ceb94d385c
Note the leading 0's. That's no timestamp, that's an oath.
In reply to the sibling comment: no. I like timestamp.com, and in fact I could have never found out about it other than by talking about the oath concept, but this is not just including it in the blockchain. It's proving its value to the author to bother doing the work of getting a good nonce for it. Literally putting my money where my mouth is. And swearing an oath to that extent, I could cryptographically swear it more, with more work, or use a smaller less impressive nonce if I'm not as sure.
And incentives? There is an incentive for me. At the same time it is effectively burning money, swearing by burning money. Took like seven seconds of compute, too. I had to wait human time for that. It's collateral, it's an oath. And it's an impediment to forgery, and in addition, an impediment to eg news sites telling different people different things. With oaths they have to tell everybody the same thing.
Could you kindly point me/others to some info about this oath concept?
I mostly get pointed to Oauth stuff when searching for "oath sha256 nonce".
There's nothing out there beyond what you've read in my two comments above yours.
Well I suppose I can still point you to it: https://news.ycombinator.com/item?id=31451260
You know what? I'll make a post about it and link it here.
This is effectively timestamp.com
There is also very little incentive structure.
Why isn’t there a folding coin? Productive mining and you reward the new protein folds or whatever
Because folding is too hard to verify. For a Blockchain to work it needs to be very easy to prove that the miner has actually done the work. With a folding problem it's very hard to prove the answer a miner gives is actually a solution to the problem rather than a quick guess. It's a shame!
Just make it centralized (like most cryptocurrency effectively is anyway). You would still need some checks and balances sure.
Make a token. Altruists buy the token to fund the “miners”. They may also make a profit but they buy knowing most likely they wont and it is for a good cause.
That is what gridcoin is. Unfortunately, it doesn't work.
There's Folding Coin (FLDC)
https://foldingcoin.net/
But it never took off.
Crypto is <valley-girl>literally</valley-girl> stopping us from finding the cure for cancer!
What about federated learning to deal with the little memory issue?
If anyone ever needs an example of how crypto is doing harm to the world: "Due to crypto mining, GPUs were in severe shortage and very over-priced on eBay"
The over-valuation of crypto is a two-fold negative impact on society: a massive brain-drain sucking talented engineers who would otherwise be solving real problems, and the opportunity cost of GPUs burning electricity to run unintentional ponzi schemes instead of training deep learning models.
You could argue the other way around. Crypto is creating market incentives to develop more powerful GPUs (for less $$$), which can make other applications possible. It's tricky to control market forces...
(not that I'm a fan of crypto)
Is it actually doing that though? The main mine-able crypto is Ethereum and I don't think Nvidia is specifically optimizing their consumer gaming GPUs for maximizing hashrates. The only innovation in this regard is their "mining" gpus which are the same thing as their normal GPUs just without physical video outputs.
I'm not complaining about the market forces causing GPUs to be expensive, I'm complaining about the non-regulation of marketing these as securities to laymen, which is exactly what is happening right now every time you see an NFT or ICO ad on Facebook or Instagram.
Wolf on Wall Street is a movie about this happening before in history, and the current craze around crypto is exactly the same thing and the reason why the qualified investor rules were made.
If you stop dumb money being funneled into crypto, and stop the moral acceptance of chasing dumb money, you can free people to pursue more noble causes.
I'm not anti crypto, but anyone can see that a good chunk, if not a majority, of current interest in crypto is due to chasing dumb money and greed. If I recall correctly Vitalik Buterin tweeted about this excess attention before. Bad attention could kill crypto before it can truly even get its feet off the ground.
It doesn't matter if the GPUs are optimized or not for crypto - I'd argue it's even better if it's not. The point is that Nvidia and AMD reach better economies of scale (which are critical for such R&D intensive businesses as cutting edge GPUs). That benefits all, if it goes in somewhat the right direction.
My key takeaways:
* The free AI breast cancer detection website took coolwulf about three months of spare time, sometime he had to sleep in his office to get things done, before the site finally went live in 2018
* The website also gained a lot of attention from the industry, during which many domestic and foreign medical institutions, such as Fudan University Hospital, expressed their gratitude to him by email and were willing to provide financial and technical support
* Afterwards, he and Weiguo Lu, now a tenured professor at University of Texas Southwest Medical Center, founded two software companies targeting the radiotherapy and started working on product development for cancer radiotherapy and artificial intelligence technologies
* But in 2022, he returned with an even more important "brain cancer project"
* coolwulf (Hao Jiang) (right) He told us that his parents are not medical professionals, and his interest in programming was fostered from a young age
* A reliable AI for tumor detection can enable a large number of patients who cannot seek adequate medical diagnosis in time to know the condition earlier or provide a secondary opinion
* He said that he’s not sure actually how many people have used it because the data is not saved on the server due to patient privacy concerns
Link to the technology: http://mammo.neuralrad.com:5300/
> A reliable AI for tumor detection can enable a large number of patients who cannot seek adequate medical diagnosis in time to know the condition earlier or provide a secondary opinion
Citation Required?
that's an unadorned http link. Really?
It's an interesting subject, with a long history; I think many of the biggest challenges are not technical.
The first commercially available AI/ML approach to breast cancer screening was available (US) in the late 90s. There have been many iterations and some improvements since, none of which really knock it out of the park but most clinical radiologists see the value. Perhaps the more interesting question then is why are people getting value out of uploading their own scans, i.e. why does their standard care path not already include this?
The reason I made this project 100% free and available to the general public is to help patients, especially in the remote area who has limited access to experienced radiologists for diagnosis, to at least get a second opinion on their mammogram. And I think this has certain value and this is why I'm doing this project.
Hao
Great project and really cool to see this.
Question I have for you is that one of the biggest problems with cancer diagnoses is false positives: "Yes, there is something on your scan, we're not sure what it is, so we'll biopsy it." Biopsy is not a 0-risk procedure, and it can cause a lot of worry and pain, so it's not something to be taken lightly. Also, there are many cases of "OK, it's probably cancer, but the cure may be worse than the disease." This is the classic problem with detecting prostate cancer at an advanced age - it's very likely/probable something else will kill you before the cancer does.
How does your software deal with this issue? I'd be worried if, as you put it, people in a remote area with limited access to experienced radiologists, were given access to this and it came back with "Fairly decent chance of cancer" - what do they do then?
Great question. Definitely sensitivity / specificity balance is a crucial topic in AI-assisted diagnosis. I have to admit this model and website for mammography was done in 2018 and might not be the leading solution out there. At the moment, if I want to improve the results of my earlier work, I will add additional stage of radiomics model to do false positive reduction, and in the mean time lower the threshold in the first model to increase sensitivity. By doing this, combining the feature extractions of deep learning and the past 20 years of knowledge in medical imaging using Radiomics might give better performance in terms of sensitivity / specificity.
Hao
You're operating in a really serious area of medicine and I encourage you to takes the comments about false positives (and false negatives) more seriously. It's not really just a matter of making adjustments to a model; it has to be pervasive in the entire process of making reproducible modeels that are used for making decisiojns about humans.
The website is cleared marked as a breast health awareness tool and not for diagnosis. We are not doing any decision making for humans. However I would like to point out there are quite some FDA approved mammo AI products in the market at the moment.
So, you think if you put up a website, people won't use it for diagnosis anyway?
yes, I know mammo AI products are on the market- those that were approved followed a collection of regulations that I think you are not.
Even if people use it for diagnosis, so what? Take the info and if it’s positive to confirm with doctor, if it’s negative but you have reservations go see doctor.
It doesn’t really change anything.
so you're saying the test is worthless in terms of actionability? Why woudl you want to take it then?
It’s not that it’s worthless, it’s simply another data point to give to a professional.
I think this comment is perhaps a bit uncharitable.
Even if you have done everything you are supposed to do in the process, at the end of the day you are looking at ROC curves or equivalent and trying to understand sensitivity vs. specificity trade offs, and often you have some sort of (indirect) parameters than can move to different points of that tradeoff.
This is quite critical in deployment, if you are screening you usually want something different than in diagnosis; as alluded to elsewhere if you raise the work-up rate too much you definitely risk killing more people from biopsy complications than you help with higher sensitivity (it's more complicated than that in practice)
deploying ML medical diagnostics is like everything else in ML: the ML part is 1% of a much larger "thing", which involves business, legal, and many other concerns far beyond the data analytics.
Nothing abotu what I'm saying is uncharitable- in a sense, it's charitable because I'm helping warn a person who is going down a dangerous path to consult more with experts in this field.
The charitable response would have been to assume they do have some understanding the broader context, and perhaps raise specific concerns or points of interest.
What you did was assume that they were ignorant in potentially dangerous ways, and assert they should do something different.
I don't think a reading of the comments/responses (at least at your time of posting) really supported that assumption, especially considering the limitations of the medium. Hence my reply to you, while also detailing the trade offs a tiny bit.
I read the entire article and I didn't see anything in there that would convince me this author is anything other than an amateur programmer (I can't parse the section about bruker- is the "amateur programmer" also a director at Bruker who ddevelops medical devices full-time"?).
Please be assured that I put a fair amount of thought into this - for example, I used to do due diligence for a VC firm evaluating proposals like this all the time and we had to reject most of them because the founders didn't understand the basic rules of deploying medical technology in highly regulated environments like the US.
Based on my interactions with the author in the various parts this post, I continue to conclude this individual is lacking core knowledge and wisdom required to execute a project like this successfully at scale.
The article was pretty fluffy, but it was about them not by them. If article was accurate about the role at Siemens they have for certain been exposed to RA/QA work and know what a DHF is, etc.
Anyway and least at they time you posted (since then there were more interaction) I didn't find the same information nearly enough to dismiss their competence out of hand.
I went back and dwetermined that the article was wrong. He wasn't a "director" at Bruker, he was a "detector imaging scientist". There's nothing about Siemens.
What I didn't totally grasp from the article is this is a company https://www.reddit.com/r/MedicalPhysics/comments/t5u2c9/intr...
So this isn't an amateur programmer, it's a person who got a phd in nuclear engineering and radiological sciences, was a scientist at bruker, has some experience with health systems, and then became a serial entrepeneur with a small company that has some funding. BTW, people who have the job title "Director" are normally fairly senior (old), as well.
We've reached diminishing returns here I think but from the article linked
Is where I got Siemens from.right, but that's not on his LinkedIn.
How can we help? Is there some way for us to contact you regarding this?
You will save even more lives if you over-communicate about the need to use your awareness tools (with better false positive reduction) to drive more effective diagnosis.
Inconsistent care is a really good point. I wasn't trying to be negative - hope it didn't come across that way. I was hoping to point out that systemic issues in health care management, at least in a lot of countries, seems to be more of a problem that tech for things like this.
Out of curiosity, how are you handling the data access and labelling issues here? I suspect that's the key issue that has limited the performance of the commercial offerings (hardly limited to this problem or this space).
OTOH in terms of real impact, properly leveraging a more modestly successful algorithm will probably help more people than getting a few more %. With the (strong) caveat that in a space like this you really have to look at work-up rate and balance risks.
There is history of breast cancer in my family and anything that can be done to improve outcomes has to be highly commended. I did however have the same question and this
> to at least get a second opinion on their mammogram.
for me makes a lot of sense, even in developed countries where you get a result but want extra assurances.
It would be interesting to know (assuming you have the data, even anecdotally) if the second opinions using this overturned professional ones and from those how many were corrected an original false negative mistake.
Not speaking for coolwulf obviously but I can perhaps shed some light.
Screening breast mammo has an occurrence rate problme. Something like less than 10 in 1000 studies will require further review; this means in practice as a radiologist you look at a lot of negative films before seeing a TP. It also means a typical read is done fast. Seconds-to-small minutes.
This results in a couple of things. Reader variability based on experience/throughput, and false negatives. There were some double reader studies that caught something like 15% (going from memory here) of FN - but nobody can affort to have two radiologists read everything.
So the profession is already conceptually used to the idea of using an algorithm as a "second read" and reconsidering. Typically this won't "overturn" anything here but rather say 'hey have another look', but the decision to proceed or not is still the clinicians. Having a positive from the algorithm makes them review carefully, but you have to watch the FP rate here or nobody would get anything else done.
I have heard of health systems using algorithms as a first pass too (i.e. radiologist only see films that have had a postive in a tuned-to-be-senstive version), but that has it's own set of issues.
'Amateur' oughtn't be scare-quoted because it's not a slur, many of the finest programmers were amateurs for many years before they were old enough to be given a job in the profession.
I had an art teacher affectionately remind me the etymology for amateur
> borrowed from French, going back to Middle French, "one who loves, lover," borrowed from Latin amātor "lover, enthusiastic admirer, devotee," from amāre "to have affection for, love, be in love, make love to" (of uncertain origin) + -tōr-, -tor, agent suffix https://www.merriam-webster.com/dictionary/amateur#etymology...
changes the feeling of it all when you get that context, someone who loves a subject pretty much--no qualifications skill wise or regarding depth but they love it and should presumably take things seriously to some degree as any lover would.
Using 'amateur' (quoted or not) is click bait. It's an embellishment to the rest of the headline. For that matter even though it's true the graphics cards are as well. Only thing that could have made it more click bait would be to also put in AI in the headline.
True, rather like age (or any discrimination category actually) is often used to make something sound more of an achievement, even though it's actually just about right place right time, experience (which you could happen to have at almost any age), etc.
If u used to be a paid software programmer and got a different job, but continued doing programming side projects without pay, are u an amateur or not?
Yes, but you'd also be a former professional. Amateur also isn't meant to be a negative term, it just means you do it as a hobby rather than as a profession.
It literally means "for the love".
As an AI/ML researcher who publishes in this area regularly, I will be using this as a case study for AI ethics classes. That this is allowed to go on is shocking.
> In 2018, a programmer named “coolwulf” started a thread about a website he had made. Users just need to upload their X-ray images, then they can let AI to carry out their own fast diagnosis of breast cancer disease.
Literally the worst fears that we have as a community is that people will recklessly apply ML to things like cancer screening on open websites and cause countless deaths, bankruptcies, needless procedures, etc. How many people went to this website, uploaded images, were told were ok and didn't follow up? How many were told they have cancer and insisted on procedures they didn't need?
The website is totally unaccountable. Totally unregulated. Totally without any of the most basic ethical standards in medicine. Without even the most basic human rights for patients. This is frankly disgusting.
In the US this would have been shut down by the FDA immediately.
We should not be celebrating this unethical "science" that doesn't meet even the most basic of scientific standards or ethical standards.
I can't believe this is getting upvoted here.
Thank you for your reply. On the site, it's cleared marked and noted this is not for diagnosis.
"This tool is only to provide you with the awareness of breast mammogram, not for diagnosis."
Unfortunately, it's just a matter of time before someone gets severe complications for whatever outcome your program gets wrong. You have a false negative? Big problem, you just gave someone worse treatment options (if lucky) or led to someone being diagnosed too late and potentially getting metastasized breast cancer. Those bone metastases are quite painful, you know.
False positives are even worse, because they are far far far more likely to happen in practice. Imagine your program telling someone it has a malignant mass (VERY bad wording, only the pathologist can say something is malignant). I speak from experience that this WILL lead to the patient going to the doctor, and the doctor, seeing these very strong words, might want to take a biopsy to confirm the malignancy. These are painful procedures that can and will go wrong, eventually (complication rate is ~1%, which is not a whole lot on its own, but is unacceptable on a known healthy population). If lucky, the biopsy can be done stereotactical, but if unlucky it'll have to be done MRI-guided. You just cost society a few thousand dollars/euros/yuan. And that's if everything goes right, worst case it'll be hospital admission due to complications, like an infection.
Your blog post says you are trying to fight cancer, which is a noble cause. If the tool is not for diagnostical purposes, it's not doing a whole lot in fighting cancer as it is just a play thing then. At the moment, it's more like hindering cancer by taking resources from people who need them and giving them to people who don't need them.
Source: researcher in AI for breast cancer screening/diagnosis.
Of course, this is going to harm some people. Is it possible the number of people it helps is significantly larger than the number of people it harms? If it were so, would you not consider that a reasonable tradeoff?
A slight digression: I find in certain countries (cough cough US), everyone is free to give advice/opinion on anything, except for medical/legal matters, which are considered sacred and so no one is free to give any advice/opinion on those. I saw a person scolded by HR for literally saying "make sure you stay well hydrated". This is madness! Medical professionals should not have a monopoly on advice concerning hydration.
The proper response to being scolded by HR for saying that is to ask what the person's username is, so you can speak to their manager.
Then, when you speak to the manager, ask them not to have their report telling people not to do things that HR doesn't have authority over (coworkers discussing health topics like that are not within the scope of HR).
Is there any possible way that a test that is not entirely reliable can be used for screening? Like, couldn't they have a radiologist look at the photo to confirm the result?
Yeah you could do that and it's happening in practice. Many countries implement double reading where two radiologists look at it. In the USA, single reading + computer-aided diagnosis is quite common using commercial, FDA-approved software.
> just a matter of time before someone gets severe complications for whatever outcome your program
If you are going to indirectly call someone out for publishing code which you deem severely detrimental to human health, you better back your stuff up.
> You have a false negative? Big problem
How does one get a false negative from multi-class probabilities? There aren't any...
> you just gave someone worse treatment options (if lucky) or led to someone being diagnosed too late and potentially getting metastasized breast cancer. Those bone metastases are quite painful, you know.
An image is already checked by a doctor, before given to the patient. A "False negative" then, is simply the doctor saying they are OK, and the ML confirming this, while both missing cancer. If doctor said, not OK, then no patient will take an OK from a website and be diagnosed too late. The last sentence is needlessly spiteful. I hope I misunderstood the intent of it.
> False positives are even worse, because they are far far far more likely to happen in practice. [...] I speak from experience that this WILL lead to the patient going to the doctor, and the doctor, seeing these very strong words, might want to take a biopsy to confirm the malignancy.
If doctor says OK, but my automated second opinion says not OK, then I go back to the doctor, and the doctor will either sooth my worries, explaining why the patch is benign. This is the desired outcome for my health. Or the doctor will second-guess their diagnosis, and perform, with my permission, more thorough tests. If tests come back OK, this will sooth my worries and the doctor's worries, and all will be fine. If tests come back not OK, then the application helped save my life.
There is zero reason to believe this app will bamboozle doctors into spending 1000s of dollars over nothing.
> If the tool is not for diagnostical purposes, it's not doing a whole lot in fighting cancer as it is just a play thing then.
This does not follow. There is a significant contribution with this app for fighting and raising awareness of cancer. Low-expertise hospitals use it as a diagnostics aid. People reported getting a second opinion from this app, returning to the hospital, and finding cancer early.
> At the moment, it's more like hindering cancer by taking resources from people who need them and giving them to people who don't need them.
Yes, it really seems you are deeming this app both severely dangerous and hindering cancer research. Without any solid (ML) reasoning and plenty of non-sequiturs. I'd suggest that even if you continue to hold such views, that you apologize to a fellow researcher.
Source: citizen scientist who open-sourced data analysis tools with disclaimers, now widely used for cancer screening, precision medicine, and improving chemo-therapy.
The words “found malicious mass”, disclaimer or not, could be considered a diagnosis. You can probably say “anomaly”, or maybe even “classified as malicious mass”. With you having connections in the US, it is probably worth talking to an expert if you want to keep this online.
Also, 50 GPUs seems like more than necessary!
It is, you can easily do this one one GPU, but it will a bit longer. Doubt it'll be prohibitively so.
Thank you for all your efforts and significant work. The intent and selfless dedication is moving. As grandma would say, talk does not cook rice. As a former student RN caring for patients at UCSF Neurosurgery & Neurology I witnessed a lot of research innovation on the "hardware" side when the "god" of last hope did grand rounds [1] One could only imagine something like this in the early 90s, software as another adjunct layer to the overall care approach. How can clinicians diagnose and treat if the patient has not been admitted into the point of service? Data driven approaches like yours would have saved a lot of $$$/RIP to increase awareness leading to early detection and care access. Pioneers shine a light for others to see, may detractors never have to face an unintended LP. [2]
[1] https://pituitaryworldnews.org/one-worlds-greatest-neurosurg... [2] https://en.wikipedia.org/wiki/Lumbar_puncture
A disclaimer that it isn’t for diagnosis isn’t generally useful when everything else says the tool does diagnosis.
You know, if people who teach ML classes at universities point out to you that what you're doing is going to be taught in classes as an example of "what not to do and how not to do it", you should probably immediately stop what you are doing and get opinions from experts in the field.
I will repeat what others are say: this is irresponible due to naivete and could be harmful to people. Please consult with experts on how to proceed.
Should the internet also be shut down because people get false conclusions from WebMD, Reddit, Twitter, Google search results, etc?
and this is why healthcare is the #1 source of bankruptcy in the US.
Some people believe that every single person must have Mercedes-Benz type of care in the US.
They cannot fathom that some of the plebs (do they even exist for them?) may want to make their own independent healthcare choices, and are willing to accept the risk ... (or can only afford!) a Suzuki.
and this is why healthcare is the #1 source of bankruptcy in the US.
May not be true: https://www.washingtonpost.com/blogs/post-partisan/wp/2018/0...
The regulatory burden on medical treatments is far too high, however, and we should have the right to try: https://fee.org/archive/topics/Right%20to%20Try
> a case study for AI ethics classes
What is unethical about this citizen science project? What is ethical about keeping it only for yourself, and not sharing it with the world?
You are saying you have the expertise to build a similar product, but releasing it would mean the worst fear of your community?
> people will recklessly apply ML
What are the indications that this is a reckless application of ML?
> How many people went to this website, uploaded images, were told were ok and didn't follow up?
Common sense dictates exactly zero. Their follow up was taking their images and getting an automated second opinion. Either a doctor already deemed them OK, or a doctor deemed them not OK, in which case, they would not rely on a second opinion, to think they are suddenly OK.
> How many were told they have cancer and insisted on procedures they didn't need?
Again, exactly zero. The app returns probabilities not binary diagnostics. No hospital would do anti-cancer procedures on a patient without cancer, even when they insist, because some website, friend, or religious leader told them so.
> The website is totally unaccountable.
Good. Or make the good-faith open-source project accountable and liable? That would simply mean shutting it down. No more diagnostics help for low-expertise hospitals: not good at all.
> Totally without any of the most basic ethical standards in medicine.
List a basic ethical standard in medicine which this project runs afoul of.
> basic human rights for patients
What right is that? The right not to upload your images to a site of your choosing? I thought human rights include self-determination, and keeping possession of your imaging to do however you see fit.
> In the US this would have been shut down by the FDA immediately.
But is that a good, ethical thing? Or simply that red tape and authority in US does not allow for such projects?
> We should not be celebrating this unethical "science" that doesn't meet even the most basic of scientific standards or ethical standards.
You should not talk about ethics or science, when you did not do even a proper evaluation of the work of a fellow scientist.
> I can't believe this is getting upvoted here.
Awaiting your work on cancer research and ML. Post it here. If devoid of ethical issues, and strongly scientific, it will also be upvoted and celebrated. Or is your major contribution going to be a snipe at someone who actually contributed?
I share your sentiment, people are focusing on successes too much but not scrutinise what potential outcomes false negatives in software like this can have.
This is unimaginably worse than what Theranos did!
really fascinated by this take!
is your issue with this project.. being public? not accurate enough? shouldn't be pursued at all because it'd never work?
what is the goal of regulation?
Exactly my thoughts! Thanks for speaking up.
Admittedly I just skimmed the article but I feel like the title should be more - "This "amateur" medical...". His primary expertise is more physics/CS/programmer related than it is human biology/medicine.
What does "90% accuracy" mean? Is this before or after applying Bayes' theorem?
At the time of this model / work developed (late 2017 / early 2018), very few public mammogram datasets were available. So the model was trained on DDSM/MIAS and tested on InBreast dataset. The 90% accuracy was calculated with results on InBreast dataset.
> As for why he doesn't commercialize the website and collect some money, this question was also asked by us.
You don't get into open source development for the big bucks. It's sort of like asking someone who starts a soup kitchen why they don't open a restaurant.
I think it's worth noting how open source development spurs technical innovation in a way commercial development does not. If a company develops a proprietary technology, almost nobody will get access to the source or datasets, and so it can't then be reused and built upon to advance the state of the art. If it's open source, everyone can build on it and improve things. Sort of like how science needs to be public in order to develop more science.
But just as we need large investments to push science further, we should also have large investments in open source. We need to make access to these 50-GPU clusters easier for open source research, so a lone developer doesn't have to built it himself. Some meager funding for the actual development would be nice too... How many OSS researchers could we fund at $35K per?
Is 90% correct rate considered good enough for this kind of use?
Seems like 1/10 wrong would be bad, how does that compare with a doctor doing it?
The 90% accuracy figure in the article unfortunately exposes the author as very ignorant.
On a German breast cancer screening population, 90% accuracy is abysmal as ~99.2% of cases are negative. Just predicting „no cancer“ would achieve 99.2% accuracy.
Accuracy is a very bad metric for such highly asymmetric problems.
To provide some context: The German screening system is able to identify cancer in ~6/1000 of patients screened, and missing it for the remaining 2/1000.
IIRC this is achieved by re-inviting in ~3% of cases to further examination, where ultrasound / needle biopsy / etc. can be done on a case by case basis.
According to the American Cancer Society
> About half of the women getting annual mammograms over a 10-year period will have a false-positive finding at some point.
Thanks.
Any idea what the false negative is?
I guess that’s the more important factor?
The american screening system has very high rates of recalls to further examination (around 2-4x higher than good european screening systems).
In Germany, roughly 20-25% of cancers are missed in screening (and often found 2 years later)
What a service to society. Hats off to you to using your skills like that!
Reading this article only makes me realise how crypto industry has crippled the progress independent researchers like this would badly need gpus for AI.
How many really useful, cool and meaningful projects are stuck because such authors can't find or afford gpus - as they are being used to calculate meaningless hashes instead :/
> In short, it is to let the AI help you "look at the film", and the accuracy rate is almost comparable to professional doctors, and it is completely free.
In the US, the issue is getting the Xray
>> Furthermore, the accuracy of tumor identification has reached 90%.
How is this accuracy calculated? Further in the article it is noted that there is no patient data saved by the project:
>> He said that he’s not sure actually how many people have used it because the data is not saved on the server due to patient privacy concerns. But during that time, he received a lot of thank-you emails from patients, many of them from China.
Considering user privacy is laudable in my opinion, but I'm still curious to know how accuracy is known.
Probably based on a test set from the original dataset.
The expression "has reached" makes it sound more like an extrinsinc evaluation process has taken place (i.e. not on a test set).
Is this machine translated? Some parts don't make much sense.
This is impressive. Wonderful work to OP.
I’m currently working with Digistain (S21) and we’re using AI to predict breast (and eventually other) cancer recurrence.
The tests are performed using infrared spectroscopy to measure protein synthesis and then fed into AI in order to make proper measurements and predictions.
We’ve shown we’re able to predict better than any other known method and are beginning our partnership and rollout to many hospitals around the world.
I don't consider him an amateur programmer if he found that many GFX cards. Not amateur not "amateur" either. It's not the kind of thing that really gets called amateur.
I'm a consummate algorithmist and have difficulty getting that amount hardware...for now at least. And he got them all of the same kind, that's pro. coolwulf is in fact cool.
I don't understand why this is useful. The starting point is a mammogram. Are there a lot of people who are able to get a mammogram performed, but not able to get it analyzed?
I applaud the author. These tools seems like a great addition to health care providers. I'm just less sure about when you would use it directly as a patient.
Let me try to explain:
1) Radiologists' experience on reading is not the same. The reading from a 5+ years experience radiologist sometime could be quite different from the one with a one or two years experience radiologist. As a matter of fact, due to the complicated issues with mammogram reading, some inexperienced radiologist don't get mammogram reading tasks when they started working in the clinics.
2) 2nd opinion sometime could be quite useful and important for the patients. By providing a utility for patients to have certain awareness on their mammogram, apart from the readings from their radiologists alone, could be useful, especially for people who are from remote area which lacks of experienced radiologists.
3) Even in big cities which has more medical resources (like big cities in China), due to the amount of patient, usually each radiologist only has about 30s to read a mammogram image (We have been told by multiple doctors and they could confirm that.). Mis-reading is very common due to the work-load. So a utility like this could help on finding or at lease warning doctors on possible missed.
There are other reasoning behind this but above are things on top of my head right now.
great story kudos to the author etc. however from my knowledge reaching the 90 (although an amazing task) is not enough. I have been in contact with different researchers in similar domains and they were struggling for 96% and above. The difficulty in which you acquire those 1% after 90% is exponential and most hospitals/research facilities etc set a lower limit. 90% is by far not acceptable.
Furthermore as others have mentioned, security etc play always a bit part.
Having a doctor miss-read/miss the tumor etc can have legal actions etc. but having a machine do that is also not within the legal scope, there are so many aspects that need to be taken into consideration.
Great effort especially for the person, but science wise not much.
Very cool, but the desktop app for chest CT doesn't work - you can open an image but nothing get diagnosed. How is the algorithm differ from existing image classification algorithms?
“Malicious mass”… “benign calcification”. Not appropriate to tell people they do or don’t have cancer based on a radiograph gpt guess.
Also, can it tell the difference between a radiograph of a watermelon and a cancer filled breast? Prove it.
Goes to show that you don’t need to be a long time programmer to have impact in most areas of society.
Why is Python so good? It democratizes by lowering the bar to coding.
For an evolutionary perspective of cancer and possible new ways to treat them, I recommend the new book "The cheating cell" by Athena Aktipis.
This guy is the real role model of our field (computer science). Not a Steve Job, not an Elon Musk. Hao Jiang (Coolwulf) is who we should all aspirate to be.
Fsck money already, we have the power to save live beings.
Imagine if all the GPU power that goes into crypto mining could be used on projects like this instead..
Amazing work. I'm glad this is going to help so many people.
Sadly, this would be illegal in the USA and get shut down pretty quickly.
More like a patent holder would usurp all the work someone else did and make a fortune off of it after taking 5 years to get through the red tape.
Na, this is all well-known work. This field is seriously big, lots of publications on it dating back to the early '90s.
What exactly is illegal about this? If you are thinking HIPAA laws, they don't apply when you are sharing your own medical information/images.
This is a regulated industry. If you don't jump through the necessary hoops, the US government will shut you down and there are many laws on the books they can use for this. There are many laws beyond HIPAA (which exists to make it easier to share data, not harder) which apply.
We may be talking about different things - I was referring to writing the software and having people upload their own data to it. You guys must be referring to more formal usage in the healthcare industry.
who operates the servers for the software? Each individual at home is going to have a machine that does inference on their own images?
unlessI'm wrong, he's in Michigan.
OP is probably talking about the legality of American hospitals using this software in an official capacity like some Chinese hospitals seem to be doing.
I'm completely unfamiliar, but it wouldn't surprise me if for diagnosing? software like this to be used in an official medical capacity in America it would need to go through some sort of particular vetting process because if it isn't it might leave hospitals who use it open to lawsuits.
that would be a potential YC idea. A company that enables smaller groups to do clinical research w/o needing an army of people to wade through the regulatory red tape. That also isn't in and of itself a giant predatory CRO type organization.
that's literally what pharma is now- companies that exist to help smaller groups get their research through the clinical and approval process. It would be hard to buidl that level of expertise in a smaller company.
The FDA may or may not attempt to classify it as a medical device and then shut it down. Otherwise legal if it includes a disclaimer.
way cool, no need to label yourself amateur, software, hardware, radiology, and real impact. Hats off to you.
did they win?
I am David Sheryl, From Rosenberg Taxes. I want to say a Grateful thank you to Wizard Web Recovery after a difficult experience with fake hackers online. I have a friend who recommend me to Wizard Web Recovery, I contacted Wizard web Recovery and told him about My medical school grade and credit score, told me not to worry and gave me his word, I don’t know him, never met him before but makes me feel comfortable with his words and experience, He keeps to his word and upgrades my credit score, I had to share this as am sure that most people here, are looking for the right hacker and have not seen any yet. try and contact Wizard Web Recovery on any kind of service you may need from him as a hacker. it is a testimony, you have the right to give yours too, by helping others to know there is still a high rate of good hackers that are highly experienced, he responds on email and WhatsApp Number: +1 (917) 725-3296) Email: wizardwebrecovery@programmer.net) or (wizardwebrecoveryprogrammer@gmail.com)
He gives a listening ear to hear you out. Greetings to you all