Equivariant Action Sampling for Reinforcement Learning and Planning

We study the problem of action sampling and propose a method to incorporate equivariance properties to the action sampling procedure.

Learning to Navigate in Mazes with Novel Layouts Using Abstract Top-down Maps

This work learns to navigate end-to-end with map input, aiming to generalize to novel map layouts in zero-shot.

Practice Makes Perfect: Planning to Learn Skill Parameter Policies

We enable a robot to rapidly and autonomously specialize parameterized skills by planning to practice them. The robot decides what skills to practice and how to practice them. The robot is left alone for hours, repeatedly practicing and improving.

Sample Efficient Modeling of Drag Coefficients for Satellites with Symmetry

We study sample-efficient modeling of invariant drag coefficients with equivariant networks.

Can Euclidean Symmetry Help in Reinforcement Learning and Planning?

We study whether Euclidean symmetry can help in reinforcement learning and planning, which models the geometric transformations between reference frames of robots.

Model-based Navigation in Environments with Novel Layouts Using Abstract 2-D Maps

This work aims to improve the zero-shot generalization performance in map-based navigation on novel layouts.

InterFact: Towards Interactive Factorization of Actionable Entities

This work proposes a reinforcement learning framework for compositional object-oriented environments.