ITU

Automatic Evolution of Multimodal Behavior with Multi-Brain HyperNEAT

Research output: Conference Article in Proceeding or Book/Report chapter › Article in proceedings › Research › peer-review

Standard

Automatic Evolution of Multimodal Behavior with Multi-Brain HyperNEAT. / Schrum, Jacob; Lehman, Joel; Risi, Sebastian.

Proceedings of the 2016 on Genetic and Evolutionary Computation Conference Companion. New York, NY, USA : Association for Computing Machinery, 2016. p. 21-22.

Research output: Conference Article in Proceeding or Book/Report chapter › Article in proceedings › Research › peer-review

Harvard

Schrum, J, Lehman, J & Risi, S 2016, Automatic Evolution of Multimodal Behavior with Multi-Brain HyperNEAT. in Proceedings of the 2016 on Genetic and Evolutionary Computation Conference Companion. Association for Computing Machinery, New York, NY, USA, pp. 21-22. https://doi.org/10.1145/2908961.2908965

APA

Schrum, J., Lehman, J., & Risi, S. (2016). Automatic Evolution of Multimodal Behavior with Multi-Brain HyperNEAT. In Proceedings of the 2016 on Genetic and Evolutionary Computation Conference Companion (pp. 21-22). Association for Computing Machinery. https://doi.org/10.1145/2908961.2908965

Vancouver

Schrum J, Lehman J, Risi S. Automatic Evolution of Multimodal Behavior with Multi-Brain HyperNEAT. In Proceedings of the 2016 on Genetic and Evolutionary Computation Conference Companion. New York, NY, USA: Association for Computing Machinery. 2016. p. 21-22 https://doi.org/10.1145/2908961.2908965

Author

Schrum, Jacob ; Lehman, Joel ; Risi, Sebastian. / Automatic Evolution of Multimodal Behavior with Multi-Brain HyperNEAT. Proceedings of the 2016 on Genetic and Evolutionary Computation Conference Companion. New York, NY, USA : Association for Computing Machinery, 2016. pp. 21-22

Bibtex

@inproceedings{8b0217953a8941c2a74e7177060e2393,
title = "Automatic Evolution of Multimodal Behavior with Multi-Brain HyperNEAT",
abstract = "An important challenge in neuroevolution is to evolve multimodal behavior. Indirect network encodings can potentially answer this challenge. Yet in practice, indirect encodings do not yield effective multimodal controllers. This paper introduces novel multimodal extensions to HyperNEAT, a popular indirect encoding. A previous multimodal approach called situational policy geometry assumes that multiple brains benefit from being embedded within an explicit geometric space. However, this paper introduces HyperNEAT extensions for evolving many brains without assuming geometric relationships between them. The resulting Multi-Brain HyperNEAT can exploit human-specified task divisions, or can automatically discover when brains should be used, and how many to use. Experiments show that multi-brain approaches are more effective than HyperNEAT without multimodal extensions, and that brains without a geometric relation to each other are superior.",
keywords = "indirect encoding, modularity, multimodal behavior",
author = "Jacob Schrum and Joel Lehman and Sebastian Risi",
year = "2016",
doi = "10.1145/2908961.2908965",
language = "Udefineret/Ukendt",
isbn = "978-1-4503-4323-7",
pages = "21--22",
booktitle = "Proceedings of the 2016 on Genetic and Evolutionary Computation Conference Companion",
publisher = "Association for Computing Machinery",
address = "USA",

}

RIS

TY - GEN

T1 - Automatic Evolution of Multimodal Behavior with Multi-Brain HyperNEAT

AU - Schrum, Jacob

AU - Lehman, Joel

AU - Risi, Sebastian

PY - 2016

Y1 - 2016

N2 - An important challenge in neuroevolution is to evolve multimodal behavior. Indirect network encodings can potentially answer this challenge. Yet in practice, indirect encodings do not yield effective multimodal controllers. This paper introduces novel multimodal extensions to HyperNEAT, a popular indirect encoding. A previous multimodal approach called situational policy geometry assumes that multiple brains benefit from being embedded within an explicit geometric space. However, this paper introduces HyperNEAT extensions for evolving many brains without assuming geometric relationships between them. The resulting Multi-Brain HyperNEAT can exploit human-specified task divisions, or can automatically discover when brains should be used, and how many to use. Experiments show that multi-brain approaches are more effective than HyperNEAT without multimodal extensions, and that brains without a geometric relation to each other are superior.

AB - An important challenge in neuroevolution is to evolve multimodal behavior. Indirect network encodings can potentially answer this challenge. Yet in practice, indirect encodings do not yield effective multimodal controllers. This paper introduces novel multimodal extensions to HyperNEAT, a popular indirect encoding. A previous multimodal approach called situational policy geometry assumes that multiple brains benefit from being embedded within an explicit geometric space. However, this paper introduces HyperNEAT extensions for evolving many brains without assuming geometric relationships between them. The resulting Multi-Brain HyperNEAT can exploit human-specified task divisions, or can automatically discover when brains should be used, and how many to use. Experiments show that multi-brain approaches are more effective than HyperNEAT without multimodal extensions, and that brains without a geometric relation to each other are superior.

KW - indirect encoding, modularity, multimodal behavior

U2 - 10.1145/2908961.2908965

DO - 10.1145/2908961.2908965

M3 - Konferencebidrag i proceedings

SN - 978-1-4503-4323-7

SP - 21

EP - 22

BT - Proceedings of the 2016 on Genetic and Evolutionary Computation Conference Companion

PB - Association for Computing Machinery

CY - New York, NY, USA

ER -

ID: 81555074