A neural framework for robot motor learning based on memory consolidation

Heni Ben Amor; Shuhei Ikemoto; Takashi Minato; Bernhard Jung; Hiroshi Ishiguro

doi:10.1007/978-3-540-71629-7_72

A neural framework for robot motor learning based on memory consolidation

Heni Ben Amor, Shuhei Ikemoto, Takashi Minato, Bernhard Jung, Hiroshi Ishiguro

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

2 Scopus citations

Abstract

Neural networks are a popular technique for learning the adaptive control of non-linear plants. When applied to the complex control of android robots, however, they suffer from serious limitations such as the moving target problem, i.e. the interference between old and newly learned knowledge. However, in order to achieve lifelong learning, it is important that robots are able to acquire new motor skills without for-getting previously learned ones. To overcome these problems, we propose a new framework for motor learning, which is based on consolidation. The framework contains a new rehearsal algorithm for retaining previously acquired knowledge and a growing neural network. In experiments, the framework was successfully applied to an artifical benchmark problem and a real-world android robot.

Original language	English (US)
Title of host publication	Adaptive and Natural Computing Algorithms - 8th International Conference, ICANNGA 2007, Proceedings
Publisher	Springer Verlag
Pages	641-648
Number of pages	8
Edition	PART 2
ISBN (Print)	9783540715900
DOIs	https://doi.org/10.1007/978-3-540-71629-7_72
State	Published - 2007
Externally published	Yes
Event	8th International Conference on Adaptive and Natural Computing Algorithms, ICANNGA 2007 - Warsaw, Poland Duration: Apr 11 2007 → Apr 14 2007

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Number	PART 2
Volume	4432 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Other

Other	8th International Conference on Adaptive and Natural Computing Algorithms, ICANNGA 2007
Country	Poland
City	Warsaw
Period	4/11/07 → 4/14/07

ASJC Scopus subject areas

Theoretical Computer Science
Computer Science(all)

Access to Document

10.1007/978-3-540-71629-7_72

Fingerprint Dive into the research topics of 'A neural framework for robot motor learning based on memory consolidation'. Together they form a unique fingerprint.

Cite this

Ben Amor, H., Ikemoto, S., Minato, T., Jung, B., & Ishiguro, H. (2007). A neural framework for robot motor learning based on memory consolidation. In Adaptive and Natural Computing Algorithms - 8th International Conference, ICANNGA 2007, Proceedings (PART 2 ed., pp. 641-648). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4432 LNCS, No. PART 2). Springer Verlag. https://doi.org/10.1007/978-3-540-71629-7_72

A neural framework for robot motor learning based on memory consolidation. / Ben Amor, Heni; Ikemoto, Shuhei; Minato, Takashi; Jung, Bernhard; Ishiguro, Hiroshi.

Adaptive and Natural Computing Algorithms - 8th International Conference, ICANNGA 2007, Proceedings. PART 2. ed. Springer Verlag, 2007. p. 641-648 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4432 LNCS, No. PART 2).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Ben Amor, H, Ikemoto, S, Minato, T, Jung, B & Ishiguro, H 2007, A neural framework for robot motor learning based on memory consolidation. in Adaptive and Natural Computing Algorithms - 8th International Conference, ICANNGA 2007, Proceedings. PART 2 edn, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 2, vol. 4432 LNCS, Springer Verlag, pp. 641-648, 8th International Conference on Adaptive and Natural Computing Algorithms, ICANNGA 2007, Warsaw, Poland, 4/11/07. https://doi.org/10.1007/978-3-540-71629-7_72

Ben Amor H, Ikemoto S, Minato T, Jung B, Ishiguro H. A neural framework for robot motor learning based on memory consolidation. In Adaptive and Natural Computing Algorithms - 8th International Conference, ICANNGA 2007, Proceedings. PART 2 ed. Springer Verlag. 2007. p. 641-648. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2). https://doi.org/10.1007/978-3-540-71629-7_72

Ben Amor, Heni ; Ikemoto, Shuhei ; Minato, Takashi ; Jung, Bernhard ; Ishiguro, Hiroshi. / A neural framework for robot motor learning based on memory consolidation. Adaptive and Natural Computing Algorithms - 8th International Conference, ICANNGA 2007, Proceedings. PART 2. ed. Springer Verlag, 2007. pp. 641-648 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2).

@inproceedings{1f94e11a92934250b54fe05b12f56ff5,

title = "A neural framework for robot motor learning based on memory consolidation",

abstract = "Neural networks are a popular technique for learning the adaptive control of non-linear plants. When applied to the complex control of android robots, however, they suffer from serious limitations such as the moving target problem, i.e. the interference between old and newly learned knowledge. However, in order to achieve lifelong learning, it is important that robots are able to acquire new motor skills without for-getting previously learned ones. To overcome these problems, we propose a new framework for motor learning, which is based on consolidation. The framework contains a new rehearsal algorithm for retaining previously acquired knowledge and a growing neural network. In experiments, the framework was successfully applied to an artifical benchmark problem and a real-world android robot.",

author = "{Ben Amor}, Heni and Shuhei Ikemoto and Takashi Minato and Bernhard Jung and Hiroshi Ishiguro",

note = "Copyright: Copyright 2020 Elsevier B.V., All rights reserved.; 8th International Conference on Adaptive and Natural Computing Algorithms, ICANNGA 2007 ; Conference date: 11-04-2007 Through 14-04-2007",

year = "2007",

doi = "10.1007/978-3-540-71629-7_72",

language = "English (US)",

isbn = "9783540715900",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

number = "PART 2",

pages = "641--648",

booktitle = "Adaptive and Natural Computing Algorithms - 8th International Conference, ICANNGA 2007, Proceedings",

edition = "PART 2",

}

TY - GEN

T1 - A neural framework for robot motor learning based on memory consolidation

AU - Ben Amor, Heni

AU - Ikemoto, Shuhei

AU - Minato, Takashi

AU - Jung, Bernhard

AU - Ishiguro, Hiroshi

PY - 2007

Y1 - 2007

N2 - Neural networks are a popular technique for learning the adaptive control of non-linear plants. When applied to the complex control of android robots, however, they suffer from serious limitations such as the moving target problem, i.e. the interference between old and newly learned knowledge. However, in order to achieve lifelong learning, it is important that robots are able to acquire new motor skills without for-getting previously learned ones. To overcome these problems, we propose a new framework for motor learning, which is based on consolidation. The framework contains a new rehearsal algorithm for retaining previously acquired knowledge and a growing neural network. In experiments, the framework was successfully applied to an artifical benchmark problem and a real-world android robot.

AB - Neural networks are a popular technique for learning the adaptive control of non-linear plants. When applied to the complex control of android robots, however, they suffer from serious limitations such as the moving target problem, i.e. the interference between old and newly learned knowledge. However, in order to achieve lifelong learning, it is important that robots are able to acquire new motor skills without for-getting previously learned ones. To overcome these problems, we propose a new framework for motor learning, which is based on consolidation. The framework contains a new rehearsal algorithm for retaining previously acquired knowledge and a growing neural network. In experiments, the framework was successfully applied to an artifical benchmark problem and a real-world android robot.

UR - http://www.scopus.com/inward/record.url?scp=38049092662&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=38049092662&partnerID=8YFLogxK

U2 - 10.1007/978-3-540-71629-7_72

DO - 10.1007/978-3-540-71629-7_72

M3 - Conference contribution

AN - SCOPUS:38049092662

SN - 9783540715900

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 641

EP - 648

BT - Adaptive and Natural Computing Algorithms - 8th International Conference, ICANNGA 2007, Proceedings

PB - Springer Verlag

T2 - 8th International Conference on Adaptive and Natural Computing Algorithms, ICANNGA 2007

Y2 - 11 April 2007 through 14 April 2007

ER -