Hierarchical Reinforcement Learning with Uncertainty-Guided Diffusional Subgoals

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wang, Vivienne Huiling, Wang, Tinghuai, Pajarinen, Joni
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912399216345088
author Wang, Vivienne Huiling
Wang, Tinghuai
Pajarinen, Joni
author_facet Wang, Vivienne Huiling
Wang, Tinghuai
Pajarinen, Joni
contents Hierarchical reinforcement learning (HRL) learns to make decisions on multiple levels of temporal abstraction. A key challenge in HRL is that the low-level policy changes over time, making it difficult for the high-level policy to generate effective subgoals. To address this issue, the high-level policy must capture a complex subgoal distribution while also accounting for uncertainty in its estimates. We propose an approach that trains a conditional diffusion model regularized by a Gaussian Process (GP) prior to generate a complex variety of subgoals while leveraging principled GP uncertainty quantification. Building on this framework, we develop a strategy that selects subgoals from both the diffusion policy and GP's predictive mean. Our approach outperforms prior HRL methods in both sample efficiency and performance on challenging continuous control benchmarks.
format Preprint
id arxiv_https___arxiv_org_abs_2505_21750
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Hierarchical Reinforcement Learning with Uncertainty-Guided Diffusional Subgoals
Wang, Vivienne Huiling
Wang, Tinghuai
Pajarinen, Joni
Machine Learning
Hierarchical reinforcement learning (HRL) learns to make decisions on multiple levels of temporal abstraction. A key challenge in HRL is that the low-level policy changes over time, making it difficult for the high-level policy to generate effective subgoals. To address this issue, the high-level policy must capture a complex subgoal distribution while also accounting for uncertainty in its estimates. We propose an approach that trains a conditional diffusion model regularized by a Gaussian Process (GP) prior to generate a complex variety of subgoals while leveraging principled GP uncertainty quantification. Building on this framework, we develop a strategy that selects subgoals from both the diffusion policy and GP's predictive mean. Our approach outperforms prior HRL methods in both sample efficiency and performance on challenging continuous control benchmarks.
title Hierarchical Reinforcement Learning with Uncertainty-Guided Diffusional Subgoals
topic Machine Learning
url https://arxiv.org/abs/2505.21750