Timezone: »

Variational Inverse Control with Events: A General Framework for Data-Driven Reward Definition
Justin Fu · Avi Singh · Dibya Ghosh · Larry Yang · Sergey Levine

Wed Dec 05 07:45 AM -- 09:45 AM (PST) @ Room 517 AB #166

The design of a reward function often poses a major practical challenge to real-world applications of reinforcement learning. Approaches such as inverse reinforcement learning attempt to overcome this challenge, but require expert demonstrations, which can be difficult or expensive to obtain in practice. We propose inverse event-based control, which generalizes inverse reinforcement learning methods to cases where full demonstrations are not needed, such as when only samples of desired goal states are available. Our method is grounded in an alternative perspective on control and reinforcement learning, where an agent's goal is to maximize the probability that one or more events will happen at some point in the future, rather than maximizing cumulative rewards. We demonstrate the effectiveness of our methods on continuous control tasks, with a focus on high-dimensional observations like images where rewards are hard or even impossible to specify.

Author Information

Justin Fu (UC Berkeley)
Avi Singh (UC Berkeley)
Dibya Ghosh (UC Berkeley)
Larry Yang (UC Berkeley)
Sergey Levine (UC Berkeley)

More from the Same Authors