Abstract: While Deep RL tackle long-horizon tasks by representation learning, most emergent latent spaces are not interpretable. In this talk, we will introduce two methods in deep reinforcement learning. STAR is a deep hierarchical reinforcement learning algorithm for learning sequences of actions that uses reachability analysis to devise a latent space grounded in the environment both useful for more efficient learning by reinforcement learning, and interpretable for interaction with humans. Further than sequences of actions, we explore complex activities with a richer temporal relationship between primitives. We also use deep reinforcement learning as a framework to learn temporal rules describing long-horizon activities with algorithm AlphaRule. These neurosymbolic methods combine trial-and-error paradigm in continuous settings with formal verification to devise an interpretable representation for human verification.

Bio: Sao Mai Nguyen is an associate professor at Ensta-Paris. She is interested in robotics, artificial intelligence and cognitive science. She tries to bridge these 3 aspects in her work, and started working in cognitive developmental robotics. She has a special interest in linking research of machine learning, robotic embodiment, child development psychology and neuroscience.