Autonomous parsing of behavior in a multi-agent setting

D. Vanderelst; E.I. Barakova

doi:10.1109/IS.2008.4670490

Autonomous parsing of behavior in a multi-agent setting

D. Vanderelst
, E.I. Barakova

Onderzoeksoutput: Hoofdstuk in Boek/Rapport/Congresprocedure › Conferentiebijdrage › Academic › peer review

2 Citaten (Scopus)

Samenvatting

Imitation learning is a promising route to instruct robotic multi-agent systems. However, imitating agents should be able to decide autonomously what behavior, observed in others, is interesting to copy. Here we investigate whether a simple recurrent network (Elman net) can be used to extract meaningful chunks from a continuous sequence of observed actions. Results suggest that, even in spite of the high level of task specific noise, Elman nets can be used for isolating re-occurring action patterns in robots. Limitations and future directions are discussed.

Originele taal-2	Engels
Titel	Intelligent Systems 2008 : IS '08 ; 4th International IEEE Conference, 6-8 September, 2008, Varna, Bulgaria
Plaats van productie	Piscataway
Uitgeverij	Institute of Electrical and Electronics Engineers
Pagina's	10-7-10-13
Volume	2
ISBN van geprinte versie	978-1-4244-1739-1
DOI's	https://doi.org/10.1109/IS.2008.4670490
Status	Gepubliceerd - 2008
Evenement	conference; IS '08 : Intelligent Systems 2008 ; 4th International IEEE Conference, 6-8 September, 2008, Varna, Bulgaria - Duur: 1 jan. 2008 → …

Congres

Congres	conference; IS '08 : Intelligent Systems 2008 ; 4th International IEEE Conference, 6-8 September, 2008, Varna, Bulgaria
Periode	1/01/08 → …
Ander	IS '08 : Intelligent Systems 2008 ; 4th International IEEE Conference, 6-8 September, 2008, Varna, Bulgaria

Toegang tot document

10.1109/IS.2008.4670490

Citeer dit

@inproceedings{02b606dc5276417d92a99eaeeb24924e,

title = "Autonomous parsing of behavior in a multi-agent setting",

abstract = "Imitation learning is a promising route to instruct robotic multi-agent systems. However, imitating agents should be able to decide autonomously what behavior, observed in others, is interesting to copy. Here we investigate whether a simple recurrent network (Elman net) can be used to extract meaningful chunks from a continuous sequence of observed actions. Results suggest that, even in spite of the high level of task specific noise, Elman nets can be used for isolating re-occurring action patterns in robots. Limitations and future directions are discussed.",

author = "D. Vanderelst and E.I. Barakova",

year = "2008",

doi = "10.1109/IS.2008.4670490",

language = "English",

isbn = "978-1-4244-1739-1",

volume = "2",

pages = "10--7--10--13",

booktitle = "Intelligent Systems 2008 : IS '08 ; 4th International IEEE Conference, 6-8 September, 2008, Varna, Bulgaria",

publisher = "Institute of Electrical and Electronics Engineers",

address = "United States",

note = "conference; IS \'08 : Intelligent Systems 2008 ; 4th International IEEE Conference, 6-8 September, 2008, Varna, Bulgaria ; Conference date: 01-01-2008",

}

Vanderelst, D & Barakova, EI 2008, Autonomous parsing of behavior in a multi-agent setting. in Intelligent Systems 2008 : IS '08 ; 4th International IEEE Conference, 6-8 September, 2008, Varna, Bulgaria. vol. 2, Institute of Electrical and Electronics Engineers, Piscataway, blz. 10-7-10-13, conference; IS '08 : Intelligent Systems 2008 ; 4th International IEEE Conference, 6-8 September, 2008, Varna, Bulgaria, 1/01/08. https://doi.org/10.1109/IS.2008.4670490

Autonomous parsing of behavior in a multi-agent setting. / Vanderelst, D.; Barakova, E.I.
Intelligent Systems 2008 : IS '08 ; 4th International IEEE Conference, 6-8 September, 2008, Varna, Bulgaria. Vol. 2 Piscataway: Institute of Electrical and Electronics Engineers, 2008. blz. 10-7-10-13.

Onderzoeksoutput: Hoofdstuk in Boek/Rapport/Congresprocedure › Conferentiebijdrage › Academic › peer review

TY - GEN

T1 - Autonomous parsing of behavior in a multi-agent setting

AU - Vanderelst, D.

AU - Barakova, E.I.

PY - 2008

Y1 - 2008

N2 - Imitation learning is a promising route to instruct robotic multi-agent systems. However, imitating agents should be able to decide autonomously what behavior, observed in others, is interesting to copy. Here we investigate whether a simple recurrent network (Elman net) can be used to extract meaningful chunks from a continuous sequence of observed actions. Results suggest that, even in spite of the high level of task specific noise, Elman nets can be used for isolating re-occurring action patterns in robots. Limitations and future directions are discussed.

AB - Imitation learning is a promising route to instruct robotic multi-agent systems. However, imitating agents should be able to decide autonomously what behavior, observed in others, is interesting to copy. Here we investigate whether a simple recurrent network (Elman net) can be used to extract meaningful chunks from a continuous sequence of observed actions. Results suggest that, even in spite of the high level of task specific noise, Elman nets can be used for isolating re-occurring action patterns in robots. Limitations and future directions are discussed.

U2 - 10.1109/IS.2008.4670490

DO - 10.1109/IS.2008.4670490

M3 - Conference contribution

SN - 978-1-4244-1739-1

VL - 2

SP - 10-7-10-13

BT - Intelligent Systems 2008 : IS '08 ; 4th International IEEE Conference, 6-8 September, 2008, Varna, Bulgaria

PB - Institute of Electrical and Electronics Engineers

CY - Piscataway

T2 - conference; IS '08 : Intelligent Systems 2008 ; 4th International IEEE Conference, 6-8 September, 2008, Varna, Bulgaria

Y2 - 1 January 2008

ER -

Autonomous parsing of behavior in a multi-agent setting

Samenvatting

Congres

Toegang tot document

Vingerafdruk

Citeer dit