Study on fundamental characteristics of learning algorithms in artificial neural networks ニューラルネットワークにおける学習アルゴリズムの基本的性質に関する研究
この論文にアクセスする
この論文をさがす
著者
書誌事項
- タイトル
-
Study on fundamental characteristics of learning algorithms in artificial neural networks
- タイトル別名
-
ニューラルネットワークにおける学習アルゴリズムの基本的性質に関する研究
- 著者名
-
Kamruzzaman, Joarder
- 著者別名
-
カムルザマン, ジョアダー
- 学位授与大学
-
室蘭工業大学
- 取得学位
-
博士 (工学)
- 学位授与番号
-
甲第11号
- 学位授与年月日
-
1993-03-23
注記・抄録
博士論文
目次
- TABLE OF CONTENTS / p3 (0006.jp2)
- ABSTRACT / p1 (0004.jp2)
- LIST OF FIGURES / p5 (0008.jp2)
- LIST OF TABLES / p12 (0015.jp2)
- 1.INTRODUCTION / p1 (0016.jp2)
- 1.1 General introduction / p1 (0016.jp2)
- 1.2 Purpose and Objective / p14 (0029.jp2)
- 1.3 Organization of thesis / p16 (0031.jp2)
- 2.BRIEF HISTORY OF NEURAL NETWORKS / p18 (0033.jp2)
- 3.DESCRIPTION OF BACKPROPAGATION ALGORITHM / p24 (0039.jp2)
- 3.1 Network Architecture / p24 (0039.jp2)
- 3.2 Formulation of Learning Algorithm / p25 (0040.jp2)
- 3.3 Remarks / p29 (0044.jp2)
- 4.PERFORMANCE OF BACKPROPAGATION LEARNING WITH LINEAR OUTPUT UNITS / p32 (0047.jp2)
- 4.1 Introduction / p32 (0047.jp2)
- 4.2 Convergence Behavior in Learning with Linear Output Units / p33 (0048.jp2)
- 4.3 Computer Simulation and Discussions / p43 (0058.jp2)
- 4.4 Concluding Remarks / p65 (0080.jp2)
- 5.GENERALIZATION ABILITY AND INCREMENTAL LEARNING IN FAHLMAN-LEBIERE ALGORITHM / p66 (0081.jp2)
- 5.1 Introduction / p66 (0081.jp2)
- 5.2 Algorithms for Network Construction / p67 (0082.jp2)
- 5.3 Description of Fahlman-Lebiere Algorithm / p70 (0085.jp2)
- 5.4 Comparative Study with Backpropagation Algorithm / p75 (0090.jp2)
- 5.5 Concluding Remarks / p103 (0118.jp2)
- 6.A FEED-FORWARD CASCADED ARCHITECTURE / p105 (0120.jp2)
- 6.1 Introduction / p105 (0120.jp2)
- 6.2 Description of Proposed Architecture / p106 (0121.jp2)
- 6.3 Comparative Study with Backpropagation Network / p109 (0124.jp2)
- 6.4 Concluding Remarks / p137 (0152.jp2)
- 7.NOISE FILTERING USING IDENTITY MAPPING BY BACKPROPAGATION / p139 (0154.jp2)
- 7.1 Introduction / p139 (0154.jp2)
- 7.2 Network Architecture and Training Patterns / p140 (0155.jp2)
- 7.3 Feature Acquisition at the Third Layer Unit / p140 (0155.jp2)
- 7.4 Generalization Ability and Steepness of Sigmoid Function / p153 (0168.jp2)
- 7.5 Noise Cancellation / p160 (0175.jp2)
- 7.6 Concluding Remarks / p171 (0186.jp2)
- 8.FORMULATION AND REDUCTION OF CROSS TALK IN ASSOCIATIVE MEMORY / p172 (0187.jp2)
- 8.1 Introduction / p172 (0187.jp2)
- 8.2 Previous Works / p174 (0189.jp2)
- 8.3 Description of Associative Memory Architecture / p176 (0191.jp2)
- 8.4 Higher Order Correlation Analysis / p177 (0192.jp2)
- 8.5 A new Associative Memory Architecture / p192 (0207.jp2)
- 8.6 Probability of exact data retrieval / p193 (0208.jp2)
- 8.7 Computer Simulation and Discussions / p196 (0211.jp2)
- 8.8 Concluding Remarks / p208 (0223.jp2)
- 9.FURTHER REDUCTION OF CROSS TALK / p209 (0224.jp2)
- 9.1 Introduction / p209 (0224.jp2)
- 9.2 Further Reduction of Cross Talk / p210 (0225.jp2)
- 9.3 Proposed Architecture applied Hamming Code / p212 (0227.jp2)
- 9.4 Computer Simulation and Discussions / p217 (0232.jp2)
- 9.5 Concluding Remarks / p229 (0244.jp2)
- 10.CONCLUSIONS / p231 (0246.jp2)
- 10.1 Conclusions / p231 (0246.jp2)
- 10.2 Future Works / p236 (0251.jp2)
- APPENDIX-I / p238 (0253.jp2)
- APPENDIX-II / p239 (0254.jp2)
- APPENDIX-III / p241 (0256.jp2)
- REFERENCES / p246 (0261.jp2)