Learning to Perceive the World Through Control: Empowerment-Based Representation Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Bastankhah, Mahsa, Broderick, Sophie, Eysenbach, Benjamin
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911730791088128
author Bastankhah, Mahsa
Broderick, Sophie
Eysenbach, Benjamin
author_facet Bastankhah, Mahsa
Broderick, Sophie
Eysenbach, Benjamin
contents In many practical reinforcement learning environments, observations are far higher-dimensional than the variables that matter for control. In this work, we ask: can we learn representations that capture only control-relevant features of the environment? We study this question through the empowerment objective, which maximizes an agent's influence over the environment and is widely used for unsupervised skill learning. We show that empowerment agents induce two distinct representations -- forward and backward -- that capture complementary aspects of the state, and both of which are invariant to control-irrelevant features. Thus, empowerment maximization leads agents to learn an implicit, control-centric model of the world. Our analysis highlights the importance of learning representations through interaction rather than from passive datasets: interaction aimed at maximizing control is essential for learning useful invariance properties, a perspective that aligns closely with the causal learning literature.
format Preprint
id arxiv_https___arxiv_org_abs_2605_30656
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Learning to Perceive the World Through Control: Empowerment-Based Representation Learning
Bastankhah, Mahsa
Broderick, Sophie
Eysenbach, Benjamin
Machine Learning
In many practical reinforcement learning environments, observations are far higher-dimensional than the variables that matter for control. In this work, we ask: can we learn representations that capture only control-relevant features of the environment? We study this question through the empowerment objective, which maximizes an agent's influence over the environment and is widely used for unsupervised skill learning. We show that empowerment agents induce two distinct representations -- forward and backward -- that capture complementary aspects of the state, and both of which are invariant to control-irrelevant features. Thus, empowerment maximization leads agents to learn an implicit, control-centric model of the world. Our analysis highlights the importance of learning representations through interaction rather than from passive datasets: interaction aimed at maximizing control is essential for learning useful invariance properties, a perspective that aligns closely with the causal learning literature.
title Learning to Perceive the World Through Control: Empowerment-Based Representation Learning
topic Machine Learning
url https://arxiv.org/abs/2605.30656