Ankur Parikh

Really excited about our new work (led by @ptshaw2) on ProtEx: A Retrieval Augmented Approach to Protein Function Prediction. Gives SOTA results on several tasks (especially for rare classes and sequences far from training set) to help better characterize the protein universe!

We made FLIP2, a protein fitness benchmark spanning seven new datasets, including enzymes, protein-protein interactions, and light-sensitive proteins, as well as splits that measure generalization relevant to real-world protein engineering campaigns.

@JBlunt1018 hope you and your family continue to stay safe

My family and I are stuck in Puerto Vallarta and was heading back home to the United States, the airport in Puerto Vallarta is also closed with the cartel violence, pray for us and our family.

Super proud of what this team is doing! And I can’t wait to share more soon.

🚨🔔Foundational graph search task as testbed: for some distribution, transformers can learn to search (100% acc). We interpreted their algo!! But as graph size ↑, transformers struggle. Scaling up # params does not help; CoT does not help. 1.5 years of learning in 10 pages!

Excited to share a new paper: “ALTA: Compiler-Based Analysis of Transformers” (w/ @James_Cohan, @jacobeisenstein, @kentonctlee, @JonathanBerant, @toutanova) arxiv.org/abs/2410.18077

Most protein function predictors only make predictions for labels seen in training. We used LLM embeddings of text describing protein function to train ProtNote, which can generalize to new functional labels described in free text. Code is available for everybody to try!

Excited to announce our new work! 🧬 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) biorxiv.org/content/10.110…

Self-rewarding LMs at #icml2024 ! Thru iterative DPO (w/ a small amount of seed data), LLM instruction following ↑ (AlpacaEval 2.0, human, MT-bench) & reward modeling ↑ (corr w human rankings). @jingxu_ml will be presenting in Vienna (Tues 7/23 11:30am); please stop by! (1/2)

Check out our new @GoogleAI paper: we curate a mixture of 5M human judgments to train general-purpose foundational autoraters. Strong LLM-as-judge scores on RewardBench (87.8%), and highest perf among baselines on LLMAggreFact + 6 other benchmarks! 📰 arxiv.org/abs/2407.10817 👇

CARE: a benchmark for the Classification And Retrieval of Enzymes and CREEP: Contrastive Reaction-EnzymE Pretraining as a baseline method for retrieving EC class from reaction. @jsunn_y Ariane Mora @Shengchao_Liu Bruce Wittmann @AnimaAnandkumar @francesarnold @yisongyue

Protein function == image caption?! ProtNLM from @GoogleAI helps scientists by annotating an impressive ✨49M✨ sequences! @andreea_gane @mlbileschi_pub Blog: 310.ai/2024/06/12/pro… Try: 310.ai/copilot/booking Tutorial: youtu.be/7a09Gnqc1Lg?fe…

Excited to announce the release of TORAX, a tokamak transport simulator from our @GoogleDeepMind Fusion team! #fusionenergy - Open-source: github.com/google-deepmin… - Uses JAX: fast, differentiable - Easy coupling of ML-surrogates Hot off the press → arxiv.org/abs/2406.06718

Very excited about our release of Pfam-N v3 on @InterProDB! A collaboration between @GoogleDeepMind and @emblebi. Check out the blogpost: 👇 xfam.wordpress.com/2024/05/31/pfa… 🧵1/5

Excited to share new work from @GoogleDeepMind / @GoogleResearch on improving LLM evals using ML predictions together with a simple but effective stratified sampling approach that strategically divides the underlying data for better performance. Paper: arxiv.org/abs/2406.04291

ProtEx: A Retrieval-Augmented Approach for Protein Function Prediction | @GoogleDeepMind - ProtEx, combines homology-based similarity search with pre-trained model - Pretrain, Multi-sequence pretraining, span denoising and sequence similarity scoring. Finetune conditions on query and retrieved exemplar sequences to predict functional labels Preprint: biorxiv.org/content/10.110…

Also Accepted as Poster for ICML 24'

Excited to share new work from @GoogleDeepMind / @GoogleResearch on “Robust Preference Optimization through Reward Model Distillation”. arxiv.org/abs/2405.19316

Happy to share our recent work on antibody library design, affinity prediction, and optimisation - 'Baselining the Buzz. Trastuzumab-HER2 affinity, and beyond!' Preprint (inc. SI) - doi.org/10.1101/2024.0… Code - github.com/oxpig/Tz_her2_… Data - doi.org/10.5281/zenodo… 🧵

"Engineering of highly active and diverse nuclease enzymes by combining machine learning and ultra-high-throughput screening" from @countablyfinite @LucyColwell37 and team looks amazing. They compare directed evolution, hit recombination, and ML-guided design of NucB.