Paul Smyth
@smyth7.bsky.social
Director of AI/ML Engineering at GSK.ai. Ex-physicist. Mostly interested in deep learning for biological sequences. All opinions my own.
Seems like an easy heuristic to apply when working out where to avoid!
December 3, 2024 at 9:02 AM
Seems like an easy heuristic to apply when working out where to avoid!
Interesting, but there are some points to note. Embedding compression is not usually how pLMs are used for variant effect prediction- see e.g ESM1v. EMS1v model was optimised for DMS prediction, so no surprise it does well. Scaling should consider data and compute, as in some earlier work this year.
November 25, 2024 at 11:18 AM
Interesting, but there are some points to note. Embedding compression is not usually how pLMs are used for variant effect prediction- see e.g ESM1v. EMS1v model was optimised for DMS prediction, so no surprise it does well. Scaling should consider data and compute, as in some earlier work this year.