UCE: Cross-Species Protein Language Modeling
jure · x · 2026-07-14
Jure Leskovec's team released a paper and code for UCE (Universal Cell Embeddings). UCE bridges molecular and cellular scales in biology, using the ESM model to represent gene information based on the proteins they encode.
The model can generalize to unseen novel species across hundreds of millions of years of evolution, breaking through the limitation of treating genes merely as columns in traditional expression matrices.
Related event: UCE Single-Cell Foundation Model Published in Nature(5 posts)→
More from Research
- DeepMind paper: LLMs can't sample randomly on their own — pass random numbers from code instead — rohanpaul_ai · 2026-10-27
- Token Collapse: Study Finds Graph Soft Prompting Tokens in LLMs Fail to Encode Structure — Sapienza · 2026-10-27
- Mathathon 2.0 launches with Terence Tao's SAIR, $1.5M in compute for teams — _sathvikr · 2026-10-27
- The 'sweet lesson' is real: two kinds of ML problems need opposite strategies — mike64_t · 2026-10-27
- PixGR: Pixel-space generative decoding for fast feed-forward 3D reconstruction — kwangmoo_yi · 2026-10-27
- UniAction: unified action conditioning lets one world model handle mixed action types — ZeYanjie · 2026-10-27