Daniel Filan

Intern with Owain Evans

Daniel Filan is working with Owain Evans to study algorithms that learn preferences from human behavior (including when that behavior is not fully rational). His interests are in technical AI safety research, in particular value learning, algorithmic information theory, and general reinforcement learning.

Daniel recently graduated from the Australian National University, where he studied mathematics and physics as an undergraduate, and wrote his honours thesis on the speed prior and a variant thereof, studying its computability and predictive power.