Neil Thomas
@countablyfinite
Followers
745
Following
1K
Statuses
389
Building AI for biological design @evoscaleai š§¬ Formerly: @Theteamatx š PhD @UCBerkeley š
Act V, Scene II
Joined November 2015
RT @etowah0: Can we learn protein biology from a language model? In new work led by @liambai21 and me, we explore how sparse autoencodersā¦
0
90
0
RT @alexrives: We're thrilled to present ESM3 in @ScienceMagazine. ESM3 is a generative language model that reasons over the three fundamenā¦
0
252
0
I'll be in Vancouver for NeurIPS / MLSB from Dec 13-16! If you're interested in protein language models, semi-supervised learning, or otherwise interested in meeting up, reach out! :)
0
0
16
RT @alexrives: Introducing ESM Cambrian. Unsupervised learning can invert biology at scale to reveal the hidden structure of the natural wā¦
0
175
0
RT @ElanaPearl: š§¬What are protein language models (PLMs) actually learning about biology? Our paper introduces InterPLM - a framework thatā¦
0
130
0
RT @yun_s_song: Large protein language models can learn complex epistatic interactions, but how much does that help with predicting variantā¦
0
6
0
RT @liambai21: Remember Golden Gate Claude? @etowah0 and I have been working on applying the same mechanistic interpretability techniquesā¦
0
197
0
RT @avapamini: excited to share & release ProtNote! š§¬š° ProtNote enables protein-function annotation on new, unseen functions by fusing LLMā¦
0
38
0
RT @milind_jagota: Happy to share this work on a new source of negative examples for antibody modeling, which has been in the oven for a whā¦
0
2
0
RT @yun_s_song: Antibodies are highly diverse, but most possible sequences are unstable or polyreactive. In this work led by @milind_jagotaā¦
0
43
0
RT @KevinKaichuang: Join the BioML group at MSR New England as a summer 2025 PhD intern! In addition to working with me, @avapamini @alexiā¦
0
18
0
RT @SluskyLab: These laureates are VERY deserving. But an important addendumāno model could be trained without 40 years of Helen Berman andā¦
0
202
0
Incredible recognition for the people who have revolutionized the field of protein design!
BREAKING NEWS The Royal Swedish Academy of Sciences has decided to award the 2024 #NobelPrize in Chemistry with one half to David Baker āfor computational protein designā and the other half jointly to Demis Hassabis and John M. Jumper āfor protein structure prediction.ā
0
0
4
Very nice work from @CadeGordonML, @amyxlu and @pabbeel ! Should save everyone O(L) forward passes when using protein MLMs to predict fitness.
Excited to announce our new work! š§¬ Some highlights are: - sequences likelihoods predict zero-shot fitness capabilities - a new method to calculate pLM likelihood in O(1) instead of O(L) forward passes - providing a causal between training data and outputs - suggesting a new finetuning method to improve pLM capabilities (1/9)
2
1
23
RT @Lux_Capital: 1/ APPS CLOSE IN 2 DAYS! Want your ideas in front of the best minds? Weāre thrilled to bring together a powerhouse of judgā¦
0
2
0
Iām excited to be on the judging panel for @Lux_Capital, @EvoscaleAI & @envedabioās Bio x ML hackathon! Looking forward to seeing what folks build with our models and discover in the datasets! Apply here by Sep 28: - open globally.
1/ ONE MORE WEEK! Weāve extended the hackathon deadline to Sat Sep 28 @ 11:59 pm PT. With increased demand from all our exciting new resources and prizes, we want to give teams more time to submit a proposal.Ā Preview our resources & apply by Sep 28: (No team or idea yet? Join our Discord)
1
8
47
RT @EliWeinstein6: Iām thrilled to announce: at @jura_bio weāve constructed a new kind of generative protein model, which allows us to manuā¦
0
78
0
RT @sacdallago: Phd students: submit your projects for a chance to get an NVIDIA fellowship! Time until Sept 11! And yes: applications inā¦
0
32
0
RT @Lux_Capital: Lux Capital is running a Bio x ML Hackathon with Lux fam cos @EvoscaleAI and @envedabio from Oct 10-20! Build on top of lā¦
0
17
0
RT @sokrypton: @kapil_devkota_ @anthonygitter @rohitsingh8080 Here is a notebook if you wanna experiment: I endedā¦
0
4
0