🚨ICML Paper Alert🚨 What if finding the right protein homologs wasn't a slow search, but a learned part of the model itself? We introduce 𝐏𝐫𝐨𝐭𝐫𝐢𝐞𝐯𝐞𝐫, an end-to-end framework that learns to retrieve the most useful homologs for self-supervised reconstruction! (1/12)

🧬 Announcing PoET-2: A breakthrough protein language model that achieves trillion-parameter performance with just 182M parameters, transforming our ability to understand proteins.

Excited to share PoET-2, our next breakthrough in protein language modeling. It represents a fundamental shift in how AI learns from evolutionary sequences. 🧵 1/13

Turns out PoET embeddings are powerful for protein representation learning, especially with small mutagenesis datasets. With PoET, we match the performance of ProteinNPT with 15x less data! 1/4

✒️PoET is our generative protein language model for controllable sequence generation, zero-shot protein mutagenesis & clinical variant effect prediction. ⚡️Unparalleled accuracy, speed, & generalizability 🧬Learn more in our latest blog post:

In a new blogpost for @openprotein, @TimT51 and I dive into some new results for PoET. Turns out PoET reaches perplexities on heldout UniRef50 sequences that would require 500 billion parameters with a typical masked protein language model! 🤯 https://t.co/XFwbNCmulP

👑New model topping the ProteinGym leaderboard👑 Congrats to the SaProt team! @duguyuan @LTEnjoy Exciting to see the progress happening at the intersection of protein sequence & structure modeling.

It's remarkable how much functional and structural information is captured by protein language models, despite being trained only on sequences. @curcuas and I dive into this in a new blog post for @openprotein. More to come! https://t.co/YMwb71kfOr

Excited to announce PoET, our (@timt1630, @tbepler1) retrieval-augmented generative protein language model that achieves state-of-the-art unsupervised variant function prediction performance on #ProteinGym. #MachineLearning #ProteinML 1/9 https://t.co/KVX6nP1uUC