Ankur Taly

Ankur Taly works on tools for debugging and understanding predictions made by deep neural networks. Additionally, he is also interested in the blending rules and machine learnt models to obtain robust systems with strong guarantees. In a past life, Ankur worked on distributed authorization, language-based security, program analysis and programming language semantics. See http://theory.stanford.edu/~ataly/ for a complete list of publications.

Google Publications

Previous Publications