Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Huh? How do you “solve” experimentally determined protein structures?

The AI models were basically trained on those.



You train on a subset and verify on another subset, validating your model.


And the results are? I'd love to see that analysis because I'd bet the accuracy rate isn't that great.


maybe by evaluating it on a test set not used for training?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: