Oddbean new post about login | logout I think AI will be a great match with MEDICAL IMAGING. While A FEW very experienced doctors can "read" an Xray or cat-scan or whatever, millions of "known images" can be assimilated by AI and produce QUICK ASSESSMENTS of images. This will change things. Instant results, no delay in treatment. Did I mention greater accuracy? There aren't enough experts to manually review all images (and there will be MORE). https://threadreaderapp.com/thread/1709585434689569066.html
@d46cf433 the problem is the LLM dataset. The input data must be nearly 100% accurate or the AI will learn incorrectly. There are millions of images out the, a huge dataset, but almost all have small mistakes. Also, how do we input the known good results? Radiology reports, like all dictated reports, have transcription errors that humans can parse but that would likely break the model. This means every image used for learning must be manually reviewed and typo-corrected.