BioML Seminar 2.1 - Mechanistic Interpretability for BioML with Elana Simon

​[IN PERSON EVENT IN BERKELEY]

​Join us on Monday night at 7 PM for an insightful talk by Elana Simon on mechanistic interpretability for bioML!

​Elana Simon is a PhD student at Stanford in James Zou's lab. Her research focuses on reverse engineering machine learning models of proteins to extract biological insights and advance model interpretability. She has previously worked on a range of computational biology problems including cancer bioinformatics, protein language models, and small molecule drug design.

​She'll be discussing InterPLM, a recent project using mechanistic interpretability techniques to identify interpretable features in protein language models. The talk will cover methods for analyzing and interpreting these features, and explore practical applications including protein annotations and design.