“The Microsoft AI team shares research that demonstrates how AI can sequentially investigate and solve medicine’s most complex diagnostic challenges—cases that expert physicians struggle to answer.”
Total_Brick_2416 on
Things like this make me really think we are on the verge of enormous societal change.
I think people that still just believe that AI is “only a pattern recognition tool that predicts that next character” are going to be in for a rude awakening in the next few years.
DueAnnual3967 on
But it is a glorified chatbot, how are such things possible
blamestross on
> Since SDBench is built from complex, pedagogically curated NEJM CPC cases, the case distribution does not match that of a real-world deployment scenario, and indeed there are no cases where the patients are in fact healthy or have benign syndromes. Thus, we do not know whether MAI-DxO’s performance gains on hard cases generalize to common, everyday clinical conditions, and could not measure false positive rates
So we call this overfitting to the test. Seems like a “hard to diagnose case” is a useful prior criteria to guessing the answer.
It’s like they trained an ai to be “House MD” and not a real doctor.
4 Comments
“The Microsoft AI team shares research that demonstrates how AI can sequentially investigate and solve medicine’s most complex diagnostic challenges—cases that expert physicians struggle to answer.”
Things like this make me really think we are on the verge of enormous societal change.
I think people that still just believe that AI is “only a pattern recognition tool that predicts that next character” are going to be in for a rude awakening in the next few years.
But it is a glorified chatbot, how are such things possible
> Since SDBench is built from complex, pedagogically curated NEJM CPC cases, the case distribution does not match that of a real-world deployment scenario, and indeed there are no cases where the patients are in fact healthy or have benign syndromes. Thus, we do not know whether MAI-DxO’s performance gains on hard cases generalize to common, everyday clinical conditions, and could not measure false positive rates
So we call this overfitting to the test. Seems like a “hard to diagnose case” is a useful prior criteria to guessing the answer.
It’s like they trained an ai to be “House MD” and not a real doctor.