Training Transition Policies via Distribution Matching for Complex Tasks

Humans decompose novel complex tasks into simpler ones to exploit previously learned skills. Analogously, hierarchical reinforcement learning seeks to leverage lower-level policies for simple tasks to solve complex ones. However, because each lower-level policy induces a different distribution of st...

Ausführliche Beschreibung

Bibliographische Detailangaben

Veröffentlicht in:	arXiv.org 2022-03
Hauptverfasser:	Ju-Seung Byun, Perrault, Andrew
Format:	Artikel
Sprache:	eng
Schlagworte:	Locomotion Machine learning Policies Skills Task complexity Training
Online-Zugang:	Volltext
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!