Welcome to Brachio Lab.

Our research focuses on developing techniques that enable reliable and explainable machine learning systems in a verifiable or provable way. We work closely with cosmologists and surgeons to ensure that our algorithms produce results and guarantees that are understandable and usable by experts.

News: We are at NeurIPS! Come find us at Poster #1625 on Tuesday, Dec 12 at 10:45AM CST to chat with Anton about stability guarantees for feature attributions. Shreya, Weiqiu, and Anton will also be at the XAI in Action workshop poster session on Saturday from 4:30PM-5:30PM to talk about this and other explainability work on faithful attributions and visual topics.



To learn more about our research, check out our blog at debugml.github.io