Doina Precup, a research team lead at DeepMind Montreal and a professor at McGill University, joins us today. We talk about Doina’s recent research interests, including her work in hierarchical reinforcement learning, with the goal of agents learning abstract representations over time. We also look at her work on reward specification for RL agents, where she hypothesizes that a reward signal in a complex environment can lead to an agent developing intuitive intelligence attributes.
We also look at several of her papers, including On the Expressivity of Markov Reward, which received a NeruIPS 2021 outstanding paper award. Finally, we talk about the analogy between hierarchical RL and CNNs, her work in continuous RL, and her thoughts on the evolution of RL in the recent past and present, as well as the biggest challenges facing the field in the future.
