Highway Value Iteration Networks

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Wang, Yuhui, Li, Weida, Faccio, Francesco, Wu, Qingyuan, Schmidhuber, Jürgen
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917685443428352
author Wang, Yuhui
Li, Weida
Faccio, Francesco
Wu, Qingyuan
Schmidhuber, Jürgen
author_facet Wang, Yuhui
Li, Weida
Faccio, Francesco
Wu, Qingyuan
Schmidhuber, Jürgen
contents Value iteration networks (VINs) enable end-to-end learning for planning tasks by employing a differentiable "planning module" that approximates the value iteration algorithm. However, long-term planning remains a challenge because training very deep VINs is difficult. To address this problem, we embed highway value iteration -- a recent algorithm designed to facilitate long-term credit assignment -- into the structure of VINs. This improvement augments the "planning module" of the VIN with three additional components: 1) an "aggregate gate," which constructs skip connections to improve information flow across many layers; 2) an "exploration module," crafted to increase the diversity of information and gradient flow in spatial dimensions; 3) a "filter gate" designed to ensure safe exploration. The resulting novel highway VIN can be trained effectively with hundreds of layers using standard backpropagation. In long-term planning tasks requiring hundreds of planning steps, deep highway VINs outperform both traditional VINs and several advanced, very deep NNs.
format Preprint
id arxiv_https___arxiv_org_abs_2406_03485
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Highway Value Iteration Networks
Wang, Yuhui
Li, Weida
Faccio, Francesco
Wu, Qingyuan
Schmidhuber, Jürgen
Machine Learning
Artificial Intelligence
Value iteration networks (VINs) enable end-to-end learning for planning tasks by employing a differentiable "planning module" that approximates the value iteration algorithm. However, long-term planning remains a challenge because training very deep VINs is difficult. To address this problem, we embed highway value iteration -- a recent algorithm designed to facilitate long-term credit assignment -- into the structure of VINs. This improvement augments the "planning module" of the VIN with three additional components: 1) an "aggregate gate," which constructs skip connections to improve information flow across many layers; 2) an "exploration module," crafted to increase the diversity of information and gradient flow in spatial dimensions; 3) a "filter gate" designed to ensure safe exploration. The resulting novel highway VIN can be trained effectively with hundreds of layers using standard backpropagation. In long-term planning tasks requiring hundreds of planning steps, deep highway VINs outperform both traditional VINs and several advanced, very deep NNs.
title Highway Value Iteration Networks
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2406.03485