Reward-free World Models for Online Imitation Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Shangzhe, Huang, Zhiao, Su, Hao
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908358457425920
author Li, Shangzhe
Huang, Zhiao
Su, Hao
author_facet Li, Shangzhe
Huang, Zhiao
Su, Hao
contents Imitation learning (IL) enables agents to acquire skills directly from expert demonstrations, providing a compelling alternative to reinforcement learning. However, prior online IL approaches struggle with complex tasks characterized by high-dimensional inputs and complex dynamics. In this work, we propose a novel approach to online imitation learning that leverages reward-free world models. Our method learns environmental dynamics entirely in latent spaces without reconstruction, enabling efficient and accurate modeling. We adopt the inverse soft-Q learning objective, reformulating the optimization process in the Q-policy space to mitigate the instability associated with traditional optimization in the reward-policy space. By employing a learned latent dynamics model and planning for control, our approach consistently achieves stable, expert-level performance in tasks with high-dimensional observation or action spaces and intricate dynamics. We evaluate our method on a diverse set of benchmarks, including DMControl, MyoSuite, and ManiSkill2, demonstrating superior empirical performance compared to existing approaches.
format Preprint
id arxiv_https___arxiv_org_abs_2410_14081
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Reward-free World Models for Online Imitation Learning
Li, Shangzhe
Huang, Zhiao
Su, Hao
Machine Learning
Imitation learning (IL) enables agents to acquire skills directly from expert demonstrations, providing a compelling alternative to reinforcement learning. However, prior online IL approaches struggle with complex tasks characterized by high-dimensional inputs and complex dynamics. In this work, we propose a novel approach to online imitation learning that leverages reward-free world models. Our method learns environmental dynamics entirely in latent spaces without reconstruction, enabling efficient and accurate modeling. We adopt the inverse soft-Q learning objective, reformulating the optimization process in the Q-policy space to mitigate the instability associated with traditional optimization in the reward-policy space. By employing a learned latent dynamics model and planning for control, our approach consistently achieves stable, expert-level performance in tasks with high-dimensional observation or action spaces and intricate dynamics. We evaluate our method on a diverse set of benchmarks, including DMControl, MyoSuite, and ManiSkill2, demonstrating superior empirical performance compared to existing approaches.
title Reward-free World Models for Online Imitation Learning
topic Machine Learning
url https://arxiv.org/abs/2410.14081