Title
Neural Network Based Reinforcement Learning for Audio-Visual Gaze Control in Human-Robot Interaction
Abstract
•We address the problem of robot gaze control in human–robot social interactions.•The robot learns without human supervision, via a reinforcement learning approach.•Our reinforcement learning method uses neural networks on multimodal data.•Pretraining on a simulated environment avoids long training periods with humans.•We test our proposal on simulated data, and on offline and online real data.
Year
DOI
Venue
2019
10.1016/j.patrec.2018.05.023
Pattern Recognition Letters
Keywords
Field
DocType
Reinforcement learning,Human–robot interaction,Robot gaze control,Neural networks,Transfer learning,Multimodal data fusion
Robot learning,Social robot,Gaze,Recurrent neural network,Psychology,Artificial intelligence,Robot,Artificial neural network,Human–robot interaction,Reinforcement learning
Journal
Volume
ISSN
Citations 
118
0167-8655
5
PageRank 
References 
Authors
0.42
18
4
Name
Order
Citations
PageRank
Stéphane Lathuilière1335.98
B.R Mâsse2368.46
Pablo Mesejo3163.01
Radu Horaud42776261.99