• Search Research Projects
  • Search Researchers
  • How to Use
  1. Back to project page

1993 Fiscal Year Final Research Report Summary

Mamalian-like neural networks for dynamic information processing and its learning algorithm

Research Project

Project/Area Number 04805032
Research Category

Grant-in-Aid for General Scientific Research (C)

Allocation TypeSingle-year Grants
Research Field 電子通信系統工学
Research InstitutionThe University of Electro-Communications

Principal Investigator

TAKAHASHI Haruhisa  The Univ.of Electro-Comms, Comms & Syst, Associate Prof., 電気通信学部, 助教授 (90135418)

Co-Investigator(Kenkyū-buntansha) TAKEDA Mitsuo  The Univ.of Electro-Comms, Comms & Syst, Professor, 電気通信学部, 教授 (00114926)
TOMITA Etsuji  The Univ.of Electro-Comms, Comms & Syst, Professor, 電気通信学部, 教授 (40016598)
Project Period (FY) 1992 – 1993
KeywordsPAC learning / Recurrent network / VC dimension / Speech recognition / Neural network
Research Abstract

(1)It is mathematically investigated as to what kind of internal representations are separable by a single output unit of a three layr feednext neural network. A topologically described necessary and sufficient condition is shown for partitions of input spaces to be classified by the output unit. Then an efficient algorithm is proposed for checking if a given partition of the input space is resulted in linear separation at the output unit.
(2)(3)These papers improves the sample complexity needed for reliable generalization in the PAC learnability in machine learning. By introducing an ill-posed learning algorithm which gives error worse over the candidates of network realizarions that are attained by minimizing empirical error, we can refine the order of the sample complexity, whereas the previous methods seek the uniform error over the whole configuration space. Essential VC dimension of concept classes, which is smaller than or equal to the number of modifiable system parameters, is introduced for calculating the generalization error instead of the traditional VC dimension analysis. Noisy learning is also treated.
(4)In this paper we propose a very simple recurrent neural network(VSRN)architecture which is a three-layr network and contains only self-loop recurrent connections in the hidden layr. The role of the recurrent connection is explained by the network dynamic and its function will be acquired by learning from finite examples like a mamalian action. Through the learning process some characteristic functions observed in the mamalian auditory systems are found automatically acquired by the network. These contain on-neuron, off-neuron and on-off-neuron. This architecture can perform phoneme spotting in real time by utilizing these characteristic functions. Some simulation experiments are done to investigate the recognition performance.

  • Research Products

    (10 results)

All Other

All Publications (10 results)

  • [Publications] 柳谷尚寿: "リカレントネットワークを用いた連続音声認識" 電子情報通信学会技術研究報告. S P93-111. 55-62 (199)

    • Description
      「研究成果報告書概要(和文)」より
  • [Publications] 山田義朗: "近似最大クリークを抽出する確率アルゴリズムとその実験的評価" 電子情報通信学会論文誌. J76-D-I,2. 46-53 (1993)

    • Description
      「研究成果報告書概要(和文)」より
  • [Publications] 高橋治久: "Estimation of learning Curve in Learning Neural Networks From Noisy Sample" International Symposium on Nonlinear Theory and its Applications HAWAII. Vol1,1.2-1. 47-50 (1993)

    • Description
      「研究成果報告書概要(和文)」より
  • [Publications] 高橋治久: "Estimating Learning Curves by PAC-Learnability Criterion" International Joint Conference on Neural Networks,Nagoya. Vol.2. 1641-1644 (1991)

    • Description
      「研究成果報告書概要(和文)」より
  • [Publications] 高橋治久: "Separability of Internal Representations in Multilayer Perceptrons With Application to Learning" Neural Networks. Vol6. 689-703 (1993)

    • Description
      「研究成果報告書概要(和文)」より
  • [Publications] 武田光夫: "Dynamics of Complex Neural Fields with an Analogy to Optical Fields Generated in a Phase-Conjugate Resonator" Proc.SPIE,San Diego. Vol.2039. 314-322 (1991)

    • Description
      「研究成果報告書概要(和文)」より
  • [Publications] Yanagiya N., Takahashi H.: ""A model for the function of recurrent connections in speech perception"" Technical Report UEC-CAS 94-1, Dept.Comm and Sys, Univ.Electro-Comms.(1994)

    • Description
      「研究成果報告書概要(欧文)」より
  • [Publications] Takahashi, H and Tomita, E.: ""Estimation of learning Curve in Learning Neural Networks From Noisy Sample."" International Symposium on Nonlinear Theory and its Applications HAWAII. (1993)

    • Description
      「研究成果報告書概要(欧文)」より
  • [Publications] Takahashi, H and Tomita, E.: ""Estimating Learning Curves by PAC-Learnability Criterion."" International Joint Conference on Neural Networks Nagoya. Vol.2. 1641-1644 (1993)

    • Description
      「研究成果報告書概要(欧文)」より
  • [Publications] Takahashi, H and Tomita, E.: ""Separability of Internal Representations in Multilayr Perceptrons With Application to Learning."" Neural Networks. 6. 689-703 (1993)

    • Description
      「研究成果報告書概要(欧文)」より

URL: 

Published: 1995-03-27  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi