Dual Goal Representations

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Park, Seohong, Mann, Deepinder, Levine, Sergey
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866915799073030144
author Park, Seohong
Mann, Deepinder
Levine, Sergey
author_facet Park, Seohong
Mann, Deepinder
Levine, Sergey
contents In this work, we introduce dual goal representations for goal-conditioned reinforcement learning (GCRL). A dual goal representation characterizes a state by "the set of temporal distances from all other states"; in other words, it encodes a state through its relations to every other state, measured by temporal distance. This representation provides several appealing theoretical properties. First, it depends only on the intrinsic dynamics of the environment and is invariant to the original state representation. Second, it contains provably sufficient information to recover an optimal goal-reaching policy, while being able to filter out exogenous noise. Based on this concept, we develop a practical goal representation learning method that can be combined with any existing GCRL algorithm. Through diverse experiments on the OGBench task suite, we empirically show that dual goal representations consistently improve offline goal-reaching performance across 20 state- and pixel-based tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2510_06714
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Dual Goal Representations
Park, Seohong
Mann, Deepinder
Levine, Sergey
Machine Learning
Artificial Intelligence
In this work, we introduce dual goal representations for goal-conditioned reinforcement learning (GCRL). A dual goal representation characterizes a state by "the set of temporal distances from all other states"; in other words, it encodes a state through its relations to every other state, measured by temporal distance. This representation provides several appealing theoretical properties. First, it depends only on the intrinsic dynamics of the environment and is invariant to the original state representation. Second, it contains provably sufficient information to recover an optimal goal-reaching policy, while being able to filter out exogenous noise. Based on this concept, we develop a practical goal representation learning method that can be combined with any existing GCRL algorithm. Through diverse experiments on the OGBench task suite, we empirically show that dual goal representations consistently improve offline goal-reaching performance across 20 state- and pixel-based tasks.
title Dual Goal Representations
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2510.06714