The impacts of diffusion kernels on recognition-primed multi-agent decision making

Xiaocong Fan, Meng Su

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

Novel perspectives on multiscale information analysis are highly demanded in several areas of multi-agent research, including large-scale agent organization and experience-based decision making. A recent breakthrough in harmonic analysis is diffusion geometry and diffusion wavelets, which offers a general framework for multiscale analysis of massive data sets. In this paper, we investigate the impacts of various diffusion kernel functions on the performance of solution synthesis in experience-based multi-agent decision making. In particular, we take a two-phase approach to conduct our experiment. In phase one (learning), we apply four commonly-used kernel functions on a data set including a large collection of battlefield decision experiences, identifying the kernel functions that can outperform the others. In phase two, we apply the kernels identified in phase one to an independent data set for testing. It is shown that Cosine exponential outperformed the other kernel functions. In general, this study indicates that, in order to achieve the best possible performance in diffusion multiscale analysis, it is critical to identify kernel functions that are applicable to the massive data set under concern.

Original languageEnglish (US)
Title of host publicationProceedings - 2010 IEEE/WIC/ACM International Conference on Intelligent Agent Technology, IAT 2010
Pages117-124
Number of pages8
DOIs
StatePublished - Dec 13 2010
Event2010 IEEE/WIC/ACM International Conference on Intelligent Agent Technology, IAT 2010 - Toronto, ON, Canada
Duration: Aug 31 2010Sep 3 2010

Publication series

NameProceedings - 2010 IEEE/WIC/ACM International Conference on Intelligent Agent Technology, IAT 2010
Volume2

Other

Other2010 IEEE/WIC/ACM International Conference on Intelligent Agent Technology, IAT 2010
CountryCanada
CityToronto, ON
Period8/31/109/3/10

Fingerprint

Decision making
Information analysis
Harmonic analysis
Geometry
Testing
Experiments

All Science Journal Classification (ASJC) codes

  • Artificial Intelligence
  • Computer Networks and Communications
  • Software

Cite this

Fan, X., & Su, M. (2010). The impacts of diffusion kernels on recognition-primed multi-agent decision making. In Proceedings - 2010 IEEE/WIC/ACM International Conference on Intelligent Agent Technology, IAT 2010 (pp. 117-124). [5614250] (Proceedings - 2010 IEEE/WIC/ACM International Conference on Intelligent Agent Technology, IAT 2010; Vol. 2). https://doi.org/10.1109/WI-IAT.2010.17
Fan, Xiaocong ; Su, Meng. / The impacts of diffusion kernels on recognition-primed multi-agent decision making. Proceedings - 2010 IEEE/WIC/ACM International Conference on Intelligent Agent Technology, IAT 2010. 2010. pp. 117-124 (Proceedings - 2010 IEEE/WIC/ACM International Conference on Intelligent Agent Technology, IAT 2010).
@inproceedings{f53412476ecd4d0ba26565a70652cd52,
title = "The impacts of diffusion kernels on recognition-primed multi-agent decision making",
abstract = "Novel perspectives on multiscale information analysis are highly demanded in several areas of multi-agent research, including large-scale agent organization and experience-based decision making. A recent breakthrough in harmonic analysis is diffusion geometry and diffusion wavelets, which offers a general framework for multiscale analysis of massive data sets. In this paper, we investigate the impacts of various diffusion kernel functions on the performance of solution synthesis in experience-based multi-agent decision making. In particular, we take a two-phase approach to conduct our experiment. In phase one (learning), we apply four commonly-used kernel functions on a data set including a large collection of battlefield decision experiences, identifying the kernel functions that can outperform the others. In phase two, we apply the kernels identified in phase one to an independent data set for testing. It is shown that Cosine exponential outperformed the other kernel functions. In general, this study indicates that, in order to achieve the best possible performance in diffusion multiscale analysis, it is critical to identify kernel functions that are applicable to the massive data set under concern.",
author = "Xiaocong Fan and Meng Su",
year = "2010",
month = "12",
day = "13",
doi = "10.1109/WI-IAT.2010.17",
language = "English (US)",
isbn = "9780769541914",
series = "Proceedings - 2010 IEEE/WIC/ACM International Conference on Intelligent Agent Technology, IAT 2010",
pages = "117--124",
booktitle = "Proceedings - 2010 IEEE/WIC/ACM International Conference on Intelligent Agent Technology, IAT 2010",

}

Fan, X & Su, M 2010, The impacts of diffusion kernels on recognition-primed multi-agent decision making. in Proceedings - 2010 IEEE/WIC/ACM International Conference on Intelligent Agent Technology, IAT 2010., 5614250, Proceedings - 2010 IEEE/WIC/ACM International Conference on Intelligent Agent Technology, IAT 2010, vol. 2, pp. 117-124, 2010 IEEE/WIC/ACM International Conference on Intelligent Agent Technology, IAT 2010, Toronto, ON, Canada, 8/31/10. https://doi.org/10.1109/WI-IAT.2010.17

The impacts of diffusion kernels on recognition-primed multi-agent decision making. / Fan, Xiaocong; Su, Meng.

Proceedings - 2010 IEEE/WIC/ACM International Conference on Intelligent Agent Technology, IAT 2010. 2010. p. 117-124 5614250 (Proceedings - 2010 IEEE/WIC/ACM International Conference on Intelligent Agent Technology, IAT 2010; Vol. 2).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - The impacts of diffusion kernels on recognition-primed multi-agent decision making

AU - Fan, Xiaocong

AU - Su, Meng

PY - 2010/12/13

Y1 - 2010/12/13

N2 - Novel perspectives on multiscale information analysis are highly demanded in several areas of multi-agent research, including large-scale agent organization and experience-based decision making. A recent breakthrough in harmonic analysis is diffusion geometry and diffusion wavelets, which offers a general framework for multiscale analysis of massive data sets. In this paper, we investigate the impacts of various diffusion kernel functions on the performance of solution synthesis in experience-based multi-agent decision making. In particular, we take a two-phase approach to conduct our experiment. In phase one (learning), we apply four commonly-used kernel functions on a data set including a large collection of battlefield decision experiences, identifying the kernel functions that can outperform the others. In phase two, we apply the kernels identified in phase one to an independent data set for testing. It is shown that Cosine exponential outperformed the other kernel functions. In general, this study indicates that, in order to achieve the best possible performance in diffusion multiscale analysis, it is critical to identify kernel functions that are applicable to the massive data set under concern.

AB - Novel perspectives on multiscale information analysis are highly demanded in several areas of multi-agent research, including large-scale agent organization and experience-based decision making. A recent breakthrough in harmonic analysis is diffusion geometry and diffusion wavelets, which offers a general framework for multiscale analysis of massive data sets. In this paper, we investigate the impacts of various diffusion kernel functions on the performance of solution synthesis in experience-based multi-agent decision making. In particular, we take a two-phase approach to conduct our experiment. In phase one (learning), we apply four commonly-used kernel functions on a data set including a large collection of battlefield decision experiences, identifying the kernel functions that can outperform the others. In phase two, we apply the kernels identified in phase one to an independent data set for testing. It is shown that Cosine exponential outperformed the other kernel functions. In general, this study indicates that, in order to achieve the best possible performance in diffusion multiscale analysis, it is critical to identify kernel functions that are applicable to the massive data set under concern.

UR - http://www.scopus.com/inward/record.url?scp=78649885231&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=78649885231&partnerID=8YFLogxK

U2 - 10.1109/WI-IAT.2010.17

DO - 10.1109/WI-IAT.2010.17

M3 - Conference contribution

AN - SCOPUS:78649885231

SN - 9780769541914

T3 - Proceedings - 2010 IEEE/WIC/ACM International Conference on Intelligent Agent Technology, IAT 2010

SP - 117

EP - 124

BT - Proceedings - 2010 IEEE/WIC/ACM International Conference on Intelligent Agent Technology, IAT 2010

ER -

Fan X, Su M. The impacts of diffusion kernels on recognition-primed multi-agent decision making. In Proceedings - 2010 IEEE/WIC/ACM International Conference on Intelligent Agent Technology, IAT 2010. 2010. p. 117-124. 5614250. (Proceedings - 2010 IEEE/WIC/ACM International Conference on Intelligent Agent Technology, IAT 2010). https://doi.org/10.1109/WI-IAT.2010.17