TRAINING A NEURAL NETWORK TO CONTROL AN AGENT USING TASK-RELEVANT ADVERSARIAL IMITATION LEARNING

A method is proposed of training a neural network to generate action data for controlling an agent to perform a task in an environment. The method includes obtaining, for each of a plurality of performances of the task, one or more first tuple datasets, each first tuple dataset comprising state data...

Ausführliche Beschreibung

Bibliographische Detailangaben

Hauptverfasser:	Colmenarejo, Sergio Gomez, Gomes de Freitas, Joao Ferdinando, Cabi, Serkan, Reed, Scott Ellison, Novikov, Alexander, Wang, Ziyu, Zolna, Konrad, Budden, David
Format:	Patent
Sprache:	eng
Schlagworte:	CALCULATING COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS COMPUTING COUNTING PHYSICS
Online-Zugang:	Volltext bestellen
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!