Owain Evans

About

I'm a postdoc at the University of Oxford with Nick Bostrom's group. I also lead a collaborative project on "Inferring Human Preferences" with Andreas Stuhlmueller, Jessica Taylor and Noah Goodman. I'm working on techniques for learning about human beliefs, preferences and values from observing human behavior or interacting with humans. This draws on machine learning (e.g. inverse reinforcement learning and probabilistic programming), cognitive science, and analytic philosophy.

I was previously a PhD student at MIT. I worked on MIT's Probabilistic Computing Project on tools and applications for Venture. I also worked on Bayesian cognitive science with Josh Tenenbaum. My dissertation was "Bayesian Computational Models for Inferring Preferences" (supervisor: Roger White).

Email

owain at mit dot edu

Papers

Evans, O., Stuhlmueller A., Goodman N. Learning the Preferences of Ignorant, Inconsistent Agents.
To appear in AAAI-16.

Evans, O., Stuhlmueller A., Goodman N. Learning the Preferences of Bounded Agents.
Abstract for NIPS 2015 Workshop on Bounded Optimality.

Evans, O., Bergen, L., Tenenbaum, J.B. Learning Structured Preferences.
Proceedings of the Thirty-Second Annual Conference of the Cognitive Science Society.

Ullman, T.D., Baker, C.L., Macindoe, O., Evans, O., Goodman, N.D., & Tenenbaum, J.B. (2010). Help or hinder: Bayesian models of social goal inference.
Advances in Neural Information Processing Systems (Vol. 22, pp. 1874-1882).

Dissertation

"Bayesian Computational Models for Inferring Preferences" (2015). Available on request.