Saved in:
Bibliographic Details
Main Authors: Lange, Moritz, Engelhardt, Raphael C., Konen, Wolfgang, Wiskott, Laurenz
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2402.12067
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916131278684160
author Lange, Moritz
Engelhardt, Raphael C.
Konen, Wolfgang
Wiskott, Laurenz
author_facet Lange, Moritz
Engelhardt, Raphael C.
Konen, Wolfgang
Wiskott, Laurenz
contents Visual navigation requires a whole range of capabilities. A crucial one of these is the ability of an agent to determine its own location and heading in an environment. Prior works commonly assume this information as given, or use methods which lack a suitable inductive bias and accumulate error over time. In this work, we show how the method of slow feature analysis (SFA), inspired by neuroscience research, overcomes both limitations by generating interpretable representations of visual data that encode location and heading of an agent. We employ SFA in a modern reinforcement learning context, analyse and compare representations and illustrate where hierarchical SFA can outperform other feature extractors on navigation tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2402_12067
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Interpretable Brain-Inspired Representations Improve RL Performance on Visual Navigation Tasks
Lange, Moritz
Engelhardt, Raphael C.
Konen, Wolfgang
Wiskott, Laurenz
Machine Learning
Neural and Evolutionary Computing
Robotics
Visual navigation requires a whole range of capabilities. A crucial one of these is the ability of an agent to determine its own location and heading in an environment. Prior works commonly assume this information as given, or use methods which lack a suitable inductive bias and accumulate error over time. In this work, we show how the method of slow feature analysis (SFA), inspired by neuroscience research, overcomes both limitations by generating interpretable representations of visual data that encode location and heading of an agent. We employ SFA in a modern reinforcement learning context, analyse and compare representations and illustrate where hierarchical SFA can outperform other feature extractors on navigation tasks.
title Interpretable Brain-Inspired Representations Improve RL Performance on Visual Navigation Tasks
topic Machine Learning
Neural and Evolutionary Computing
Robotics
url https://arxiv.org/abs/2402.12067