smitha milli




I want to make machine learning systems more robust and reliable. I am interested in how interaction with humans can help ensure that a system has learned the correct objective or behavior.

I am a 1st-year PhD student at Berkeley advised by Anca Dragan and Moritz Hardt. In Summer 2017 I was a research intern at OpenAI. I am a part of the Berkeley Artificial Intelligence Research Lab and the Center for Human-Compatible AI .

I introspect a lot on my own styles of thinking and productivity (aim -> phronesis). Here is some collected research advice. I like books (history, philosophy, econ, Haruki Murakami, children’s books with deep themes, …), art, music, and performances (improv, plays, poetry slams, storytelling, stand-up comedy, …).

If you want to chat, shoot me an email at smitha dot milli at gmail dot com.

people providing me validation

teaching

workshops co-organized

Reliable Machine Learning in the Wild. ICML 2017. website