amyx.lu https://amyx.lu/ Recent content on amyx.lu amyx.lu https://amyx.lu/images/papermod-cover.png https://amyx.lu/images/papermod-cover.png Hugo -- 0.147.9 en-us 2025 Amy X. Lu Wed, 01 Oct 2025 00:00:00 +0000 [In silico #002](https://luma.com/txswapqo), IDEALondon https://amyx.lu/talks/2025-insilico/ Wed, 01 Oct 2025 00:00:00 +0000 https://amyx.lu/talks/2025-insilico/ Invited talk at IDEALondon. https://amyx.lu/news/2025-10-01/ Wed, 01 Oct 2025 00:00:00 +0000 https://amyx.lu/news/2025-10-01/ I've joined [Isomorphic Labs](https://www.isomorphiclabs.com/) as a Research Scientist 🥳! https://amyx.lu/news/2025-07-07/ Mon, 07 Jul 2025 00:00:00 +0000 https://amyx.lu/news/2025-07-07/ [PhD Dissertation Talk](https://events.berkeley.edu/eecs/event/298296-dissertation-talk-generative-models-for-real-world) at the BAIR seminar series 👩‍🎓. https://amyx.lu/news/2025-04-28/ Mon, 28 Apr 2025 00:00:00 +0000 https://amyx.lu/news/2025-04-28/ BAIR Seminar / [UC Berkeley EECS PhD Dissertation Talk](https://events.berkeley.edu/eecs/event/298296-dissertation-talk-generative-models-for-real-world) https://amyx.lu/talks/2025-dissertation/ Mon, 28 Apr 2025 00:00:00 +0000 https://amyx.lu/talks/2025-dissertation/ Excited to be in Singapore for [ICLR 2025](https://iclr.cc/) to present our work on protein language model likelihoods. https://amyx.lu/news/2025-04-23/ Wed, 23 Apr 2025 00:00:00 +0000 https://amyx.lu/news/2025-04-23/ New post on the [BAIR blog](https://bair.berkeley.edu/blog/2025/04/08/plaid/) on our work with CHEAP and PLAID. https://amyx.lu/news/2025-04-08/ Tue, 08 Apr 2025 00:00:00 +0000 https://amyx.lu/news/2025-04-08/ The [Exploration in AI Today (EXAIT)](https://exait-workshop.github.io/) workshop is accepted at ICML 2025. https://amyx.lu/news/2025-03-20/ Thu, 20 Mar 2025 00:00:00 +0000 https://amyx.lu/news/2025-03-20/ New post on [Nathan's substack](https://ncfrey.substack.com/p/hit-the-vintage-store-and-get-yourself) on CHEAP and PLAID. https://amyx.lu/news/2025-03-19/ Wed, 19 Mar 2025 00:00:00 +0000 https://amyx.lu/news/2025-03-19/ [Baker Lab](https://www.bakerlab.org/) Journal Club, UW Institute for Protein Design https://amyx.lu/talks/2025-bakerlab/ Mon, 03 Mar 2025 00:00:00 +0000 https://amyx.lu/talks/2025-bakerlab/ [Latent Labs](https://www.latentlabs.com/) Journal Club https://amyx.lu/talks/2025-latent-labs/ Sun, 02 Mar 2025 00:00:00 +0000 https://amyx.lu/talks/2025-latent-labs/ [NVIDIA Fundamental Generative AI Research](https://research.nvidia.com/labs/genair/) https://amyx.lu/talks/2025-nvidia/ Mon, 24 Feb 2025 00:00:00 +0000 https://amyx.lu/talks/2025-nvidia/ [Xaira Therapeutics](https://www.xaira.com/) https://amyx.lu/talks/2025-xaira/ Fri, 21 Feb 2025 00:00:00 +0000 https://amyx.lu/talks/2025-xaira/ Genome modeling and design across all domains of life with Evo 2 https://amyx.lu/publications/2025-evo2/ Fri, 21 Feb 2025 00:00:00 +0000 https://amyx.lu/publications/2025-evo2/ Evo 2 is a 40B parameter genomic foundation model capable of predicting functional impacts of genetic variations, autonomously learning biological features, and generating novel genomic sequences across all domains of life. [Ginkgo Bioworks](https://www.ginkgo.bio/) Journal Club https://amyx.lu/talks/2025-ginkgo/ Mon, 10 Feb 2025 00:00:00 +0000 https://amyx.lu/talks/2025-ginkgo/ [Isomorphic Labs](https://www.isomorphiclabs.com/) https://amyx.lu/talks/2025-isomorphic-labs/ Fri, 07 Feb 2025 00:00:00 +0000 https://amyx.lu/talks/2025-isomorphic-labs/ [Lila Sciences](https://www.lila.ai/) https://amyx.lu/talks/2025-lila-sciences/ Fri, 07 Feb 2025 00:00:00 +0000 https://amyx.lu/talks/2025-lila-sciences/ [Genesis Therapeutics](https://genesistherapeutics.ai/) https://amyx.lu/talks/2025-genesis-therapeutics/ Tue, 28 Jan 2025 00:00:00 +0000 https://amyx.lu/talks/2025-genesis-therapeutics/ [Profluent Bio](https://www.profluent.bio/) https://amyx.lu/talks/2025-profluent/ Tue, 28 Jan 2025 00:00:00 +0000 https://amyx.lu/talks/2025-profluent/ [Microsoft Research AI for Science](https://www.microsoft.com/en-us/research/lab/microsoft-research-ai-for-science/) https://amyx.lu/talks/2025-msr-ai-science/ Wed, 22 Jan 2025 00:00:00 +0000 https://amyx.lu/talks/2025-msr-ai-science/ [EvolutionaryScale](https://www.evolutionaryscale.ai/) https://amyx.lu/talks/2025-evolutionaryscale/ Tue, 21 Jan 2025 00:00:00 +0000 https://amyx.lu/talks/2025-evolutionaryscale/ Remote talk at [EvolutionaryScale](https://www.evolutionaryscale.ai/) on our work with CHEAP and PLAID. https://amyx.lu/news/2025-01-21/ Tue, 21 Jan 2025 00:00:00 +0000 https://amyx.lu/news/2025-01-21/ [Debora Marks Lab](https://www.deboramarkslab.com/) Journal Club, Harvard Medical School https://amyx.lu/talks/2025-marks-lab/ Fri, 17 Jan 2025 00:00:00 +0000 https://amyx.lu/talks/2025-marks-lab/ I'll be giving a remote talk at the [Marks Lab](https://www.deboramarkslab.com/) at Harvard Medical School. https://amyx.lu/news/2025-01-17/ Fri, 17 Jan 2025 00:00:00 +0000 https://amyx.lu/news/2025-01-17/ [Machine Learning for Structural Biology](https://www.mlsb.io/) (MLSB) 2024 workshop https://amyx.lu/talks/2024-mlsb/ Tue, 10 Dec 2024 00:00:00 +0000 https://amyx.lu/talks/2024-mlsb/ I'll be in Vancouver for [NeurIPS 2024](https://neurips.cc/Conferences/2024). Come say hi! https://amyx.lu/news/2024-12-09/ Mon, 09 Dec 2024 00:00:00 +0000 https://amyx.lu/news/2024-12-09/ Our [preprint](https://www.biorxiv.org/content/10.1101/2024.12.02.626353v1) and [code](https://github.com/amyxlu/plaid) on all-atom co-generation with latent diffusion is released! https://amyx.lu/news/2024-12-06/ Fri, 06 Dec 2024 00:00:00 +0000 https://amyx.lu/news/2024-12-06/ All-Atom Protein Generation with Latent Diffusion https://amyx.lu/publications/2024-plaid/ Mon, 02 Dec 2024 00:00:00 +0000 https://amyx.lu/publications/2024-plaid/ PLAID is a multimodal protein generation model that generates all-atom protein structures from function and organism prompts, but requires only sequence training data. Excited to give an invited talk at the [Stanford AI + Biomedicine series](https://snap.stanford.edu/ai-bio-seminar/). https://amyx.lu/news/2024-10-22/ Tue, 22 Oct 2024 00:00:00 +0000 https://amyx.lu/news/2024-10-22/ [Stanford AI + Biomedicine Seminar](https://snap.stanford.edu/ai-bio-seminar/) https://amyx.lu/talks/2024-stanford/ Sun, 20 Oct 2024 00:00:00 +0000 https://amyx.lu/talks/2024-stanford/ [ML Protein Engineering Seminar Series](https://www.ml4proteinengineering.com/) https://amyx.lu/talks/2024-mlproteins/ Tue, 15 Oct 2024 00:00:00 +0000 https://amyx.lu/talks/2024-mlproteins/ [Model weights](https://huggingface.co/amyxlu/cheap-proteins) for CHEAP are now released. https://amyx.lu/news/2024-10-11/ Fri, 11 Oct 2024 00:00:00 +0000 https://amyx.lu/news/2024-10-11/ Very excited to have two papers accepted as an oral presentation at [MLSB 2024](https://www.mlsb.io/)! https://amyx.lu/news/2024-10-08/ Tue, 08 Oct 2024 00:00:00 +0000 https://amyx.lu/news/2024-10-08/ Checkout our [preprint](https://www.biorxiv.org/content/10.1101/2024.10.03.616542v1) on understanding how training data affects protein language model likelihoods! https://amyx.lu/news/2024-10-03/ Thu, 03 Oct 2024 00:00:00 +0000 https://amyx.lu/news/2024-10-03/ Protein Language Model Fitness Is a Matter of Preference https://amyx.lu/publications/2024-preference/ Thu, 03 Oct 2024 00:00:00 +0000 https://amyx.lu/publications/2024-preference/ Enabled by a one-pass pseudolikelihood algorithm, we find that pLMs capture artifacts of training data selection rather than true fitness landscape via influence functions. Excited to return to the [ML Protein Engineering Seminar Series](https://www.ml4proteinengineering.com/) for an invited talk. https://amyx.lu/news/2024-10-01/ Tue, 01 Oct 2024 00:00:00 +0000 https://amyx.lu/news/2024-10-01/ [South Park Commons](https://www.southparkcommons.com/) Demo Night on Interpretability and Steerability https://amyx.lu/talks/2024-spc/ Sun, 15 Sep 2024 00:00:00 +0000 https://amyx.lu/talks/2024-spc/ Lightning talk at [South Park Commons](https://www.southparkcommons.com/) for the interpretability and steerability series. https://amyx.lu/news/2024-09-05/ Thu, 05 Sep 2024 00:00:00 +0000 https://amyx.lu/news/2024-09-05/ [Learning on Graphs and Geometry (LoGG) Seminar Series](https://portal.valencelabs.com/events/post/tokenized-and-continuous-embedding-compressions-of-protein-sequence-and-Uq8Nm5HEcHopMrX) https://amyx.lu/talks/2024-logg/ Thu, 15 Aug 2024 00:00:00 +0000 https://amyx.lu/talks/2024-logg/ New [preprint](https://www.biorxiv.org/content/10.1101/2024.08.06.606920v2) on the unreasonable compressibility of protein folding model latent spaces. https://amyx.lu/news/2024-08-08/ Thu, 08 Aug 2024 00:00:00 +0000 https://amyx.lu/news/2024-08-08/ Tokenized and Continuous Embedding Compressions of Protein Sequence and Structure https://amyx.lu/publications/2024-cheap/ Tue, 06 Aug 2024 00:00:00 +0000 https://amyx.lu/publications/2024-cheap/ CHEAP is a joint embedding of protein sequence and structure that can be obtained from sequence alone, and unveil insights into the compressibilitiy, tokenizability, and mechanistic interpretability of protein folding models. Will be in Vienna to share two ICML workshop papers at [AccMLBio](https://accml.bio/) and [ML4LMS](https://ml4lms.bio/). https://amyx.lu/news/2024-07-27/ Sat, 27 Jul 2024 00:00:00 +0000 https://amyx.lu/news/2024-07-27/ BIOE 145/245 Guest Lecture on AlphaFold2 https://amyx.lu/talks/2024-af2/ Mon, 15 Apr 2024 00:00:00 +0000 https://amyx.lu/talks/2024-af2/ TOPH: Adapting A Contrastive Question-Answering Framework for Protein Search https://amyx.lu/publications/2023-toph/ Sat, 01 Jul 2023 00:00:00 +0000 https://amyx.lu/publications/2023-toph/ We present a protein semantic similarity search method for RNA-Guided endonuclease discovery, inspired by dense retrieval methods in open-domain question answering, and introduce a new dataset of CRISPR-Cas and evolutionary-related nucleases. Pretraining strategies for effective promoter-driven gene expression prediction https://amyx.lu/publications/2023-promoter/ Mon, 27 Feb 2023 00:00:00 +0000 https://amyx.lu/publications/2023-promoter/ Pretraining and transfer learning strategies for improving model-based design of promoters for cell type-specific expression. Data-Driven Optimization for Protein Design: Workflows, Algorithms and Metrics https://amyx.lu/publications/2022-mldd/ Fri, 01 Apr 2022 00:00:00 +0000 https://amyx.lu/publications/2022-mldd/ Strategies for data curation, model-training, optimization, and evaluation heuristics for data-driven proposals of novel de novo proteins. Discovering molecular features of intrinsically disordered regions by using evolution for contrastive learning https://amyx.lu/publications/2022-reverse-homology/ Sat, 01 Jan 2022 00:00:00 +0000 https://amyx.lu/publications/2022-reverse-homology/ Reverse Homology is a self-supervised method which captures evolutionary information by contrastive learning to discover molecular features of intrinsically disordered regions. Learned embeddings from deep learning to visualize and predict protein sets https://amyx.lu/publications/2021-bioembeddings/ Fri, 01 Jan 2021 00:00:00 +0000 https://amyx.lu/publications/2021-bioembeddings/ Evolution Is All You Need: Phylogenetic Augmentation for Contrastive Learning https://amyx.lu/publications/2020-eiayn/ Wed, 23 Dec 2020 00:00:00 +0000 https://amyx.lu/publications/2020-eiayn/ We outline how viewing evolution as natural sequence augmentation for contrastive learning recapitulates comparative genomics, and maximizes the mutual information between sequence and function. Self-Supervised Contrastive Learning of Protein Representations by Mutual Information Maximization https://amyx.lu/publications/2020-cpcprot/ Fri, 04 Sep 2020 00:00:00 +0000 https://amyx.lu/publications/2020-cpcprot/ CPCProt uses contrastive learning to learn a parameter-efficient way of embedding proteins, and performs competitively with large language models. Hurtful Words: Quantifying Biases in Clinical Contextual Word Embeddings https://amyx.lu/publications/2020-hurtful-words/ Sun, 01 Mar 2020 00:00:00 +0000 https://amyx.lu/publications/2020-hurtful-words/ We apply fairness definitions to quantify the cross-group bias in BERT embeddings pretrained on medical notes, and find statistically significant differences in classifier performance. The Cells Out of Sample (COOS) dataset and benchmarks for measuring out-of-sample generalization of image classifiers https://amyx.lu/publications/2019-coos/ Sun, 01 Dec 2019 00:00:00 +0000 https://amyx.lu/publications/2019-coos/ Introduces the COOS-7 dataset to benchmark and evaluate the capacity of feature learning methods to generalize to natural distribution shifts in microscopy images.