Benutzer: Gast  Login
Titel:

Analysis and Optimisation of Bellman Residual Errors with Neural Function Approximation

Dokumenttyp:
Zeitschriftenaufsatz
Autor(en):
Gottwald, Martin; Gronauer, Sven; Shen, Hao; Diepold, Klaus
Abstract:
Recent development of Deep Reinforcement Learning (DRL) has demonstrated superior performance of neural networks in solving challenging problems with large or even continuous state spaces. One specific approach is to deploy neural networks to approximate value functions by minimising the Mean Squared Bellman Error (MSBE) function. Despite great successes of DRL, development of reliable and efficient numerical algorithms to minimise the MSBE is still of great scientific interest and practical dem...     »
Stichworte:
Critical Point Analysis, Dynamic Programming, Gauss Newton Algorithm, Local Quadratic Convergence, Mean Squared Bellman Error, Residual Gradient
Dewey Dezimalklassifikation:
620 Ingenieurwissenschaften
Zeitschriftentitel:
CoRR
Jahr:
2021
Band / Volume:
abs/2106.08774
Jahr / Monat:
2021-10
Quartal:
4. Quartal
Monat:
Oct
Reviewed:
nein
Sprache:
en
WWW:
https://arxiv.org/abs/2106.08774
TUM Einrichtung:
Lehrstuhl für Datenverarbeitung
Letzte Änderung:
02.08.2022
CC-Lizenz:
by-nc-sa, http://creativecommons.org/licenses/by-nc-sa/3.0/de
 BibTeX