Diffusion Reward: Learning Rewards via Conditional Video Diffusion

TL;DR

Diffusion Reward introduces a novel approach to learning reward functions from expert videos using conditional video diffusion models, enabling effective reinforcement learning in complex visual tasks and generalizing to unseen challenges.

Contribution

The paper presents a new framework that leverages conditional video diffusion models to learn rewards from expert videos, improving exploration and generalization in visual RL tasks.

Findings

Outperforms baselines in robotic manipulation tasks

Successfully generalizes to unseen tasks

Effective in both simulation and real-world environments

Abstract

Learning rewards from expert videos offers an affordable and effective solution to specify the intended behaviors for reinforcement learning (RL) tasks. In this work, we propose Diffusion Reward, a novel framework that learns rewards from expert videos via conditional video diffusion models for solving complex visual RL problems. Our key insight is that lower generative diversity is exhibited when conditioning diffusion on expert trajectories. Diffusion Reward is accordingly formalized by the negative of conditional entropy that encourages productive exploration of expert behaviors. We show the efficacy of our method over robotic manipulation tasks in both simulation platforms and the real world with visual input. Moreover, Diffusion Reward can even solve unseen tasks successfully and effectively, largely surpassing baseline methods. Project page and code: https://diffusion-reward.github.io.