Yamaguchi’s: Oscillation network model of pattern recognition (1982- )

Mizuno-BIONOLONICS project was carried out under supervision of Hiroshi Shimizu, University of Tokyo as the first project of ERATO JST in 1982-1987. This project was devoted to inovative research of information system based on a biological principle of selforganization.

A visual pattern recognition model, Holovision, was reported in 1985, based on the hypothesis, “pre-understanding and sensory signals are unified through synchronization to give presentation in the brain”.

This model was later developed to provide the ability of figure-ground separation and metastability of ambiguous pattern recogniion (1995, 1996).

The same principle of synchronization between perception and memory was extended to vowel recognition in the auditory system (1994).

The last paragraph is given to show that wave propagation is necessary to process temporal sequence in real time, which was inspired by collaboration with Tanituchi lab at Tokyo Medical and Dental University and students at Tokyo Denki University (1993-2000). This principle continued to later hippocampus study.

Selected papers:

Visual pattern recognition

H. Shimizu, Y. Yamaguchi, I. Tsuda and M. Yano: “Pattern Recognition Based on Holonic Information Dynamics towards Synergetic Computers”, in Complex Systems – Operational Approaches in Neurobiology, Physics, and Computers. ed. H. Haken, Springer, 225-239, 1985.

H. Shimizu and Y. Yamaguchi: Synergetic Computers and Holonics – Information Dynamics of a Semantic Computers, Physica Scripta 36, 970-985, 1987.

Y. Yamaguchi and H. Shimizu: Pattern recognition with Figure-Ground Separation by generation of coherent oscillations, Neural Networks,7(1) pp.49-63,1994. (This paper won Best Paper Award in JNNS in 1995.)

Y. Hirakura, Y. Yamaguchi , H. Shimizu and S. Nagai: Dynamic Linking among Neural Oscillators leads to Flexible Pattern Recognition with Figure Ground Separation, Neural Networks , 9(2) pp.189-209,1996.

Additional references

H. Shimizu, Y. Yamaguchi and K. Satoh: “Holovision: a Semantic Information Processor for Visual Perception”, in Dynamic Patterns in Complex Systems. JAS Kelso, AJ Mandel and MF Shlesinger eds., World Scientific, 42-76, 1988.

H. Shimizu and Y. Yamaguchi: How animals understand the meaning of indefinite information from environments? Prog. Theor. Phys. Supplement No. 99, 404-424, 1989.

H. Shimizu, Y.Yamaguchi and Y.Liu: “A Neural Network System to Discover a Relevant Figure in Indefinite Information”, Annual International conference of IEEE Engineering in Medicine and Biology Society. 12(4) 1774-1775, 1990.

Y. Yamaguchi, Y. Hirakura and H. Shimizu, Analyses of Dynamical Coherence in the Pattern Recognition Model of Oscillator Neural Networks, Holovision, Proceedings of 1993 IJCNN, 159-162 1993.

Y. Yamaguchi and H. Shimizu: “Self-organization of Information in the Brain”, in Physics of Living State., T. Musha and Y.Sawada eds. Ohm-Sya, Tokyo, 253-286, 1994.

Y. Yamaguchi: Synchronization in the Brain – Unification of Information as the Emergence-, in Complexity and Diversity, E.R.Nakamura Editor-in-Chief, Springer, 147-153, 1997.

Vowel recogition through synchronization

F. Liu, Y. Yamaguchi and H. Shimizu: Flexible vowel recognition by the generation of dynamic coherence in oscillator neural networks: speaker-independent vowel recognition, Biol. Cybern.71:105-114, 1994.

F. Liu, Y.Yamaguchi and H. Shimizu, Recognition of Ambiguous Vowels by Generation of Dynamical Coherence in Oscillator Neural Networks, Proceedings of 1993 IJCNN, 1155-1158, 1993.

Auditory recognition as synchronization of wave propagations

Y. Yamaguchi,Y. Kanzaka, F. Tanaka, J. Horikawa and I. Taniguchi: Bifurcation of Wave Propagation as a Base of Dynamical Linking in Auditory Recognition. The Fifth International Conference on Neural Informaion Processing, 1426-1429, 1998

Yamaguchi Y., Horikawa J. and Taniguchi I. : Population Network Dynamics and Modeling of Vocal Processing in the Cortex in Biophysical Neural Networks – Foundations of analytical neuroscience, ed. Poznanski RR, Mary Ann Liebert, Inc., New York 343-362, 2001.


メールアドレスが公開されることはありません。 * が付いている欄は必須項目です