MRS: Multi-Resolution Skills for HRL Agents

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Sharma, Shashank, Hoffmann, Janina, Namboodiri, Vinay
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908981381824512
author Sharma, Shashank
Hoffmann, Janina
Namboodiri, Vinay
author_facet Sharma, Shashank
Hoffmann, Janina
Namboodiri, Vinay
contents Hierarchical reinforcement learning (HRL) decomposes the policy into a manager and a worker, enabling long-horizon planning but introducing a performance gap on tasks requiring agility. We identify a root cause: in subgoal-based HRL, the manager's goal representation is typically learned without constraints on reachability or temporal distance from the current state, preventing precise local subgoal selection. We further show that the optimal subgoal distance is both task- and state-dependent: nearby subgoals enable precise control but amplify prediction noise, while distant subgoals produce smoother motion at the cost of geometric precision. We propose Multi-Resolution Skills (MRS), which learns multiple goal-prediction modules each specialized to a fixed temporal horizon, with a jointly trained meta-controller that selects among them based on the current state. MRS consistently outperforms fixed-resolution baselines and significantly reduces the performance gap between HRL and non-HRL state-of-the-art on DeepMind Control Suite, Gym-Robotics, and long-horizon AntMaze tasks. [Project page: https://sites.google.com/view/multi-res-skills/home]
format Preprint
id arxiv_https___arxiv_org_abs_2505_21410
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MRS: Multi-Resolution Skills for HRL Agents
Sharma, Shashank
Hoffmann, Janina
Namboodiri, Vinay
Artificial Intelligence
Machine Learning
Robotics
Hierarchical reinforcement learning (HRL) decomposes the policy into a manager and a worker, enabling long-horizon planning but introducing a performance gap on tasks requiring agility. We identify a root cause: in subgoal-based HRL, the manager's goal representation is typically learned without constraints on reachability or temporal distance from the current state, preventing precise local subgoal selection. We further show that the optimal subgoal distance is both task- and state-dependent: nearby subgoals enable precise control but amplify prediction noise, while distant subgoals produce smoother motion at the cost of geometric precision. We propose Multi-Resolution Skills (MRS), which learns multiple goal-prediction modules each specialized to a fixed temporal horizon, with a jointly trained meta-controller that selects among them based on the current state. MRS consistently outperforms fixed-resolution baselines and significantly reduces the performance gap between HRL and non-HRL state-of-the-art on DeepMind Control Suite, Gym-Robotics, and long-horizon AntMaze tasks. [Project page: https://sites.google.com/view/multi-res-skills/home]
title MRS: Multi-Resolution Skills for HRL Agents
topic Artificial Intelligence
Machine Learning
Robotics
url https://arxiv.org/abs/2505.21410