TY - GEN
T1 - Mixing actual and predicted sensory states based on uncertainty estimation for flexible and robust robot behavior
AU - Murata, Shingo
AU - Masuda, Wataru
AU - Tomioka, Saki
AU - Ogata, Tetsuya
AU - Sugano, Shigeki
N1 - Funding Information:
Acknowledgement. This work was supported in part by JST CREST Grant Number: JPMJCR15E3, Japan;JSPS KAKENHI Grant Numbers: 25220005, 17K12754, Japan and the “Fundamental Study for Intelligent Machine to Coexist with Nature” program of the Research Institute for Science and Engineering at Waseda University, Japan.
Publisher Copyright:
© Springer International Publishing AG 2017.
PY - 2017
Y1 - 2017
N2 - In this paper, we propose a method to dynamically modulate the input state of recurrent neural networks (RNNs) so as to realize flexible and robust robot behavior. We employ the so-called stochastic continuous-time RNN (S-CTRNN), which can learn to predict the mean and variance (or uncertainty) of subsequent sensorimotor information. Our proposed method uses this estimated uncertainty to determine a mixture ratio for combining actual and predicted sensory states of network input. The method is evaluated by conducting a robot learning experiment in which a robot is required to perform a sensory-dependent task and a sensory-independent task. The sensory-dependent task requires the robot to incorporate meaningful sensory information, and the sensory-independent task requires the robot to ignore irrelevant sensory information. Experimental results demonstrate that a robot controlled by our proposed method exhibits flexible and robust behavior, which results from dynamic modulation of the network input on the basis of the estimated uncertainty of actual sensory states.
AB - In this paper, we propose a method to dynamically modulate the input state of recurrent neural networks (RNNs) so as to realize flexible and robust robot behavior. We employ the so-called stochastic continuous-time RNN (S-CTRNN), which can learn to predict the mean and variance (or uncertainty) of subsequent sensorimotor information. Our proposed method uses this estimated uncertainty to determine a mixture ratio for combining actual and predicted sensory states of network input. The method is evaluated by conducting a robot learning experiment in which a robot is required to perform a sensory-dependent task and a sensory-independent task. The sensory-dependent task requires the robot to incorporate meaningful sensory information, and the sensory-independent task requires the robot to ignore irrelevant sensory information. Experimental results demonstrate that a robot controlled by our proposed method exhibits flexible and robust behavior, which results from dynamic modulation of the network input on the basis of the estimated uncertainty of actual sensory states.
KW - Neurorobotics
KW - Recurrent neural networks
KW - Robot
KW - Uncertainty
UR - http://www.scopus.com/inward/record.url?scp=85034265361&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85034265361&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-68600-4_2
DO - 10.1007/978-3-319-68600-4_2
M3 - Conference contribution
AN - SCOPUS:85034265361
SN - 9783319685991
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 11
EP - 18
BT - Artificial Neural Networks and Machine Learning – ICANN 2017 - 26th International Conference on Artificial Neural Networks, Proceedings
A2 - Verschure, Paul F.
A2 - Lintas, Alessandra
A2 - Villa, Alessandro E.
A2 - Rovetta, Stefano
PB - Springer Verlag
T2 - 26th International Conference on Artificial Neural Networks, ICANN 2017
Y2 - 11 September 2017 through 14 September 2017
ER -