Optimization transfer for computational learning: A hierarchy from f-ICA and alpha-EM to their offsprings

Yasuo Matsuyama; Shuichiro Imahara; Naoto Katsumata

Optimization transfer for computational learning: A hierarchy from f-ICA and alpha-EM to their offsprings

Yasuo Matsuyama^*, Shuichiro Imahara, Naoto Katsumata

^*Corresponding author for this work

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

3 Citations (Scopus)

Abstract

Likelihood optimization methods for learning algorithms are generalized and faster algorithms are provided. The idea is to transfer the optimization to a general class of convex divergences between two probability density functions. The first part explains why such optimization transfer is significant. The second part contains derivation of the generalized ICA (Independent Component Analysis). Experiments on brain fMRI maps are reported. The third part discusses this optimization transfer in the generalized EM algorithm (Expectation-Maximization). Hierarchical descendants to this algorithm such as vector quantization and self-organization are also explained.

Original language	English
Title of host publication	Proceedings of the International Joint Conference on Neural Networks
Pages	1883-1888
Number of pages	6
Volume	2
Publication status	Published - 2002
Event	2002 International Joint Conference on Neural Networks (IJCNN '02) - Honolulu, HI Duration: 2002 May 12 → 2002 May 17

Other

Other	2002 International Joint Conference on Neural Networks (IJCNN '02)
City	Honolulu, HI
Period	02/5/12 → 02/5/17

ASJC Scopus subject areas

Software

Cite this

@inproceedings{eb2a293771e0488d9a3b71c5eadccba3,

title = "Optimization transfer for computational learning: A hierarchy from f-ICA and alpha-EM to their offsprings",

abstract = "Likelihood optimization methods for learning algorithms are generalized and faster algorithms are provided. The idea is to transfer the optimization to a general class of convex divergences between two probability density functions. The first part explains why such optimization transfer is significant. The second part contains derivation of the generalized ICA (Independent Component Analysis). Experiments on brain fMRI maps are reported. The third part discusses this optimization transfer in the generalized EM algorithm (Expectation-Maximization). Hierarchical descendants to this algorithm such as vector quantization and self-organization are also explained.",

author = "Yasuo Matsuyama and Shuichiro Imahara and Naoto Katsumata",

year = "2002",

language = "English",

volume = "2",

pages = "1883--1888",

booktitle = "Proceedings of the International Joint Conference on Neural Networks",

note = "2002 International Joint Conference on Neural Networks (IJCNN '02) ; Conference date: 12-05-2002 Through 17-05-2002",

}

TY - GEN

T1 - Optimization transfer for computational learning

T2 - 2002 International Joint Conference on Neural Networks (IJCNN '02)

AU - Matsuyama, Yasuo

AU - Imahara, Shuichiro

AU - Katsumata, Naoto

PY - 2002

Y1 - 2002

N2 - Likelihood optimization methods for learning algorithms are generalized and faster algorithms are provided. The idea is to transfer the optimization to a general class of convex divergences between two probability density functions. The first part explains why such optimization transfer is significant. The second part contains derivation of the generalized ICA (Independent Component Analysis). Experiments on brain fMRI maps are reported. The third part discusses this optimization transfer in the generalized EM algorithm (Expectation-Maximization). Hierarchical descendants to this algorithm such as vector quantization and self-organization are also explained.

AB - Likelihood optimization methods for learning algorithms are generalized and faster algorithms are provided. The idea is to transfer the optimization to a general class of convex divergences between two probability density functions. The first part explains why such optimization transfer is significant. The second part contains derivation of the generalized ICA (Independent Component Analysis). Experiments on brain fMRI maps are reported. The third part discusses this optimization transfer in the generalized EM algorithm (Expectation-Maximization). Hierarchical descendants to this algorithm such as vector quantization and self-organization are also explained.

UR - http://www.scopus.com/inward/record.url?scp=0036088054&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0036088054&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:0036088054

VL - 2

SP - 1883

EP - 1888

BT - Proceedings of the International Joint Conference on Neural Networks

Y2 - 12 May 2002 through 17 May 2002

ER -

Optimization transfer for computational learning: A hierarchy from f-ICA and alpha-EM to their offsprings

Abstract

Other

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this