I happen to know that they infact are. One of the actual uses of AI.
Millions of images from specimens collected over decades have been fed into these nueral networks.
Essentially, when used for anything other than chatbots AI should do one specific job extremely well. This is because it is trained in the same manner as any human. You give it images of specimens and the diagnosis (bit more complicated than that, but it’s the important part).
I have worked adjacent to it. Impressive is an understatement. I wouldn’t benefit financially of course as I wouldn’t be on here talking about it.
I don’t exactly understand the implementation I just know big research institutions, Google, nVidia, and many smaller companies are working it. The amount of data is nothing short of enormous, and even better still, it appears to raise the standard of care across insured and uninsured populations.
Not only is the ROI there, it’s already assisting in radiology and medical imaging as well. Behind all the bullshit being sold there is actual good stuff being done.
Imagine being in the bumfuck middle of nowhere Montana, and you have to get something checked out. Anywhere in the US a doctor will be able to tell if something is abnormal, but you need specialists to determine how abnormal a specimen is, and if tissue needs to be examined closer or specialized treatment is necessary. Instead of having to send it to John Hopkins or Mayo Clinic your closest hospital had one of these machines - just like an MRI - and they could get a diagnosis quickly? You can get the specialized treatment or testing faster at a lower cost.
peoplebeproblems@midwest.social 3 weeks ago
I happen to know that they infact are. One of the actual uses of AI.
Millions of images from specimens collected over decades have been fed into these nueral networks.
Essentially, when used for anything other than chatbots AI should do one specific job extremely well. This is because it is trained in the same manner as any human. You give it images of specimens and the diagnosis (bit more complicated than that, but it’s the important part).
bumblefumble@mander.xyz 3 weeks ago
I know some people from uni that made a startup doing exactly this type of stuff, they seem to be very successful. It’s impressive stuff, really.
peoplebeproblems@midwest.social 3 weeks ago
I have worked adjacent to it. Impressive is an understatement. I wouldn’t benefit financially of course as I wouldn’t be on here talking about it.
I don’t exactly understand the implementation I just know big research institutions, Google, nVidia, and many smaller companies are working it. The amount of data is nothing short of enormous, and even better still, it appears to raise the standard of care across insured and uninsured populations.
Not only is the ROI there, it’s already assisting in radiology and medical imaging as well. Behind all the bullshit being sold there is actual good stuff being done.
Imagine being in the bumfuck middle of nowhere Montana, and you have to get something checked out. Anywhere in the US a doctor will be able to tell if something is abnormal, but you need specialists to determine how abnormal a specimen is, and if tissue needs to be examined closer or specialized treatment is necessary. Instead of having to send it to John Hopkins or Mayo Clinic your closest hospital had one of these machines - just like an MRI - and they could get a diagnosis quickly? You can get the specialized treatment or testing faster at a lower cost.