3 ms·
Do you think there's potential for language models to play a role here? I know that AI can get tossed around as a buzzword, but hasn't it proved quite successfu
by ahepp 3y ago
Do you think there's potential for language models to play a role here? I know that AI can get tossed around as a buzzword, but hasn't it proved quite successful in fields like computer vision?
I'm not deeply familiar with the state of that art, but it seems like recovering the metadata from a PDF generated by LaTeX would be no more impressive than many other things we're currently seeing language models achieve?
- staunton 3y agoI'm absolutely positive a few million dollars could get you a system that can "read aloud" pdf math papers in no time. I guess people will wait for it to become cheaper though.
- hutzlibu 3y agoYou can also have that cheaper already. But having it stable and reliable - will take some time and possibly more money, depending on your definition of reliable.
- throwaway287391 3y agoYou wouldn't need to use computer vision on a picture of the PDF. arXiv has the tex source for most of the papers. An LLM trained on code could do a pretty good job of translating tex to readable html with a bit of effort.
- miki123211 3y agoMathpix is trying to achieve something like this, and they do consider the visually impaired market AFAIK, but it's pretty expensive and I have no experience with it personally, so I can't say how good it is.