A generalized product rule and weak independence based on Bregman divergence

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

Abstract

To describe the relation between some values, arithmetic operations like multiplication and division are important and conventional tools. These arithmetic operations for probabilities are characterized by the KL divergence, and hence, they can be generalized by using the Bregman divergence instead of the KL divergence. With this idea, independence of random variables is modified by generalized product rule, and a joint probability model is proposed based on this modified weak independence. Effectiveness of weak independent models is shown by numerical experiments on toy examples, and discussed from a geometrical viewpoint.

Original languageEnglish
Title of host publicationWMSCI 2008 - The 12th World Multi-Conference on Systemics, Cybernetics and Informatics, Jointly with the 14th International Conference on Information Systems Analysis and Synthesis, ISAS 2008 - Proc.
Pages248-253
Number of pages6
Volume5
Publication statusPublished - 2008
Event12th World Multi-Conference on Systemics, Cybernetics and Informatics, WMSCI 2008, Jointly with the 14th International Conference on Information Systems Analysis and Synthesis, ISAS 2008 - Orlando, FL
Duration: 2008 Jun 292008 Jul 2

Other

Other12th World Multi-Conference on Systemics, Cybernetics and Informatics, WMSCI 2008, Jointly with the 14th International Conference on Information Systems Analysis and Synthesis, ISAS 2008
CityOrlando, FL
Period08/6/2908/7/2

Fingerprint

Random variables
Experiments

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Networks and Communications

Cite this

Fujimoto, Y., & Murata, N. (2008). A generalized product rule and weak independence based on Bregman divergence. In WMSCI 2008 - The 12th World Multi-Conference on Systemics, Cybernetics and Informatics, Jointly with the 14th International Conference on Information Systems Analysis and Synthesis, ISAS 2008 - Proc. (Vol. 5, pp. 248-253)

A generalized product rule and weak independence based on Bregman divergence. / Fujimoto, Yu; Murata, Noboru.

WMSCI 2008 - The 12th World Multi-Conference on Systemics, Cybernetics and Informatics, Jointly with the 14th International Conference on Information Systems Analysis and Synthesis, ISAS 2008 - Proc.. Vol. 5 2008. p. 248-253.

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

Fujimoto, Y & Murata, N 2008, A generalized product rule and weak independence based on Bregman divergence. in WMSCI 2008 - The 12th World Multi-Conference on Systemics, Cybernetics and Informatics, Jointly with the 14th International Conference on Information Systems Analysis and Synthesis, ISAS 2008 - Proc.. vol. 5, pp. 248-253, 12th World Multi-Conference on Systemics, Cybernetics and Informatics, WMSCI 2008, Jointly with the 14th International Conference on Information Systems Analysis and Synthesis, ISAS 2008, Orlando, FL, 08/6/29.
Fujimoto Y, Murata N. A generalized product rule and weak independence based on Bregman divergence. In WMSCI 2008 - The 12th World Multi-Conference on Systemics, Cybernetics and Informatics, Jointly with the 14th International Conference on Information Systems Analysis and Synthesis, ISAS 2008 - Proc.. Vol. 5. 2008. p. 248-253
Fujimoto, Yu ; Murata, Noboru. / A generalized product rule and weak independence based on Bregman divergence. WMSCI 2008 - The 12th World Multi-Conference on Systemics, Cybernetics and Informatics, Jointly with the 14th International Conference on Information Systems Analysis and Synthesis, ISAS 2008 - Proc.. Vol. 5 2008. pp. 248-253
@inproceedings{aeac4c16105543649590326db9f15542,
title = "A generalized product rule and weak independence based on Bregman divergence",
abstract = "To describe the relation between some values, arithmetic operations like multiplication and division are important and conventional tools. These arithmetic operations for probabilities are characterized by the KL divergence, and hence, they can be generalized by using the Bregman divergence instead of the KL divergence. With this idea, independence of random variables is modified by generalized product rule, and a joint probability model is proposed based on this modified weak independence. Effectiveness of weak independent models is shown by numerical experiments on toy examples, and discussed from a geometrical viewpoint.",
author = "Yu Fujimoto and Noboru Murata",
year = "2008",
language = "English",
isbn = "1934272353",
volume = "5",
pages = "248--253",
booktitle = "WMSCI 2008 - The 12th World Multi-Conference on Systemics, Cybernetics and Informatics, Jointly with the 14th International Conference on Information Systems Analysis and Synthesis, ISAS 2008 - Proc.",

}

TY - GEN

T1 - A generalized product rule and weak independence based on Bregman divergence

AU - Fujimoto, Yu

AU - Murata, Noboru

PY - 2008

Y1 - 2008

N2 - To describe the relation between some values, arithmetic operations like multiplication and division are important and conventional tools. These arithmetic operations for probabilities are characterized by the KL divergence, and hence, they can be generalized by using the Bregman divergence instead of the KL divergence. With this idea, independence of random variables is modified by generalized product rule, and a joint probability model is proposed based on this modified weak independence. Effectiveness of weak independent models is shown by numerical experiments on toy examples, and discussed from a geometrical viewpoint.

AB - To describe the relation between some values, arithmetic operations like multiplication and division are important and conventional tools. These arithmetic operations for probabilities are characterized by the KL divergence, and hence, they can be generalized by using the Bregman divergence instead of the KL divergence. With this idea, independence of random variables is modified by generalized product rule, and a joint probability model is proposed based on this modified weak independence. Effectiveness of weak independent models is shown by numerical experiments on toy examples, and discussed from a geometrical viewpoint.

UR - http://www.scopus.com/inward/record.url?scp=84869451599&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84869451599&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84869451599

SN - 1934272353

SN - 9781934272350

VL - 5

SP - 248

EP - 253

BT - WMSCI 2008 - The 12th World Multi-Conference on Systemics, Cybernetics and Informatics, Jointly with the 14th International Conference on Information Systems Analysis and Synthesis, ISAS 2008 - Proc.

ER -