Slashdot Mirror
Chinese AI Beats 15 Doctors In Tumor Diagnosis Competition (thenextweb.com)
An artificial intelligence system called BioMind has managed to defeated a team comprised of 15 of China's top doctors by a margin of two to one. The Next Web reports the details: When diagnosing brain tumors, BioMind was correct 87 percent of the time, compared to 66 percent by the medical professionals. The AI also only took 15 minutes to diagnose the 225 cases, while doctors took 30. In regards to predicting brain hematoma expansion, BioMind was victorious again, as it was correct in 83 percent of cases, with humans managing only 63 percent. Researchers trained the AI by feeding it thousands upon thousands of images from Beijing Tiantan Hospital's archives. This has made it as good at diagnosing neurological diseases as senior doctors, as it has a 90 percent accuracy rate. Further reading available via Xinhua.
This is really amazing. It is like computers are good at image recognition. I see a lot of potential in this AI.
There had to be a human that was right 100% of the time to judge the competition and know that the AI had 87% accuracy compared to the 15 doctors 66%.
what the hell kind of olympics are they having in china? brain tumor diagnoses? that's something that shouldn't be rushed in the first place
Did the AI give any false positives?
Did the doctors correctly diagnose any cases that the AI did not?
Whilst this is great news, I hope doctors use it as a learning and aid tool instead of a full diagnosis suite without a review by the specialist.
"It shouldn't be rushed in the first place" because doctors are human and make more mistakes when getting tired. Why would you think that would apply to AIs? The only thing that matters of AI is accuracy, and if it can be more accurate than humans while being faster, then "rushing" doesn't matter. Think it through.
Those who do not learn from commit history are doomed to regress it.
This is encouraging news. i had read previously of Watson helping doctors in diagnosing a baffling case of Leukemia that seemed totally untreatable, Watson correctly indicated that the patient had 2 kinds of Leukemia at once, which the human doctors didn’t realise.
Machine learning will not soon replace human doctors, I think, but you can see it becoming a powerful support tool for doctors, hopefully finding ailments earlier as well. Fascinating.
But sadly, the editors have not managed to defeated poor editing. Who wrote this? Creimer?
Is AI just an umbrella term that can mean vastly different things?
I've been running SpamAssassin for 15 years. Is that AI, too? I trained it by giving it thousands of non-spam and spam messages.
I like the name, since all minds are biological except this one.
i didn't say shit about AI. "The AI also only took 15 minutes to diagnose the 225 cases, while doctors took 30." i said no doctor should be rushing through that many diagnoses in 30 minutes
Finally something cool posted
it's not a tumah!
"Tempers are wearing thin. Let's just hope some robot doesn't kill everybody." --Bender
The Chinese 'top doctors' are kindergarten toddlers.
15 of them, neigh, even if they trot out 1.5 billion of their kindergarten toddlers none of them can diagnose any kind of tumor.
AI is now more a tarpaulin than an umbrella, it covers everything from machine learning, pattern recognition, expert systems, neural networks, data mining and analysis, navigation or anything with speech recognition and natural language processing. basically AI is a pretty meaningless term nowadays, usually when you here the term AI it is being followed by a lot of corporate bullshit or is being used by someone that doesn't really understand what they are talking about.
Assuming that the doctors in the contest are the FIRST. How do you guys know they did not take however many images of X-rays of people already diagnosed, treated and confirmed as having cancer or not, and gave THOSE to the people who arranged this contest? I have a feeling that, like me, almost everyone else discussing this here, did NOT read the original story either.
Machine learning is a subset of AI. A* search is also AI, even though it's not machine learning. "AI" means "technology that was found searching for AI," just like "Tang" is "space technology" even though it practically has nothing to do with space.
If you want to differentiate, it's more reasonable to differentiate between "strong AI" and "weak AI." Strong AI is what most people mean by AI, whereas "weak AI" is "cool stuff that's not really intelligent."
"First they came for the slanderers and i said nothing."
my cat beat those same Chinese doctors on the same test [/sarc]
Let me know when that AI beats doctors from Johns Hopkins, or CTCA, or City of Hope, etc.
It's not that I, as a tech freak, doubt that AI can become better at such tasks, but rather that I have worked with Chinese guys with PhDs who in reality were little better than high school graduates in the west were back in the 80's (not now perhaps given the American educational system plunge). I simply doubt the use of Chinese doctors as a yardstick. Oh, and before some moron leaps to the idea that I'm a racist: Nope, China is an officially Communist one-party-rule system with government control over speech, which includes academia in particular. That situation breeds academic backwardness, and a hobbled intellectual climate which results in mediocrity and very dangerous gaps in any field involving objectivity, which is not politically allowed. if China ever embraces freedom and liberty, it will be one of the most amazing and energetic nations on Earth.
At least 10 of them didn't even go to medical school - they just got some papers printed up.
...need not apply.
China's top doctors are worse than computer in tumer diagnosis.
Celebrate if and only if you wish more bad diagnosis in China.
I should mod you up just for the Tang.
As soon they will be the last humans on earth with this skill. If due to some calamity the AI systems fail or become destroyed, then who will still have the knowledge and experience to do these things manually?
(or to program new AI systems and feed them new sample data for that matter)
When doctors stare at a sample of blood they were consistently unable to identify that the sample had sickle-cells.
Just saying "AI did it" doesn't actually make this anything more that a centuries-long list of "a medical test has been invented". It's good and all that but there seems to be some idea that "AI" means "voodoo".
"Encyclopedia" is to "Wikipedia" what "Library" is to "Some people at a bus stop"
A rigged competion. Doctors are to diagnoze too many cases in too short time, and no analysis of false positives. In short: bs
ANUSHKA SHARMA HAVING THE VERY BAD CONVERSATION ON THE ROADðY±
https://wanchukone.blogspot.com/2018/07/Anushka-Sharma-bad-conversation-on-the-road.html?m=1
what i really find worrying is that doctors were only able to detect 66% of tumors, so a lot of people get false diagnostics for something very deadly.
On a long enough timeline, the survival rate for everyone drops to zero.
AI: 87% success
Chinese doctors: 66% success
Flip of a coin: 50% success
Chinese doctors are a bit closer to the coin flipping rather than to AI performance. Maybe that, rather praising AI, we should instead blame the professional capacities of the doctors. Furthermore, consider that since diagnosis is mostly based on images, AI can (and probably does) perform some basic image processing that can greatly enhance the detection of subtle features. To my experience doctors usually work with raw images (and usually are trained to do so). Teaching them to use some basic image processing on Xray images could enhance their skills at little cost.
False negatives are really important here.
100 cancer cases. 75 cancer diagnosis. 25 non cancer diagnosis. 75% success. 25% dead due to misdiagnosis.
100 non cancer cases. 75 non cancer diagnosis. 25 cancer diagnosis. 75% success. 0% dead due to misdiagnosis.
The way the AI deals with this can make or break it.
The self-governing monopoly trade union known as the AMA will see to it that this software never enters the public domain. The beach houses and Ferraris of featherbedded radiologists should never come under threat.
Prove anything by multiplying Huge Number times Tiny Number
Algorithms are faster at parsing data than people, just like calculators are faster at calculating. This isn't 'AI', and hard data is not always the only consideration. We have had algorithms forever (some of you seem to be under the impression this was created in your lifetime. It wasn't) processing power is magnitudes stronger now, that's all. They are useful tools, but they are not 'intelligence' and they should not be utilized or trusted implicitly beyond the calculator example. I doubt you could 'beat' that calculator at math, it doesn't make it 'smart'.
"50 Years of Successful Predictive Modeling Should Be Enough:(2002)
https://www.journals.uchicago.edu/doi/10.1086/341846
Unfortunately, for licensed occupations that is not so. Instead we rely on random individuals who, once trained, are within a generation cast away, to perform diagnoses.
Maybe the AI can be trained to do some grammar checks as well?
the program prescribed lion tails and rhino horns to treat the cancer.