An adaptive penalty-based learning extension for the backpropagation family
スポンサーリンク
概要
- 論文の詳細を見る
Over the years, many improvements and refinements to the backpropagation learning algorithm have been reported. In this paper, a new adaptive penalty-based learning extension for the backpropagation learning algorithm and its variants is proposed. The new method initially puts pressure on artificial neural networks in order to get all outputs for all training patterns into the correct half of the output range, instead of mainly focusing on minimizing the difference between the target and actual output values. The upper bound of the penalty values is also controlled. The technique is easy to implement and computationally inexpensive. In this study, the new approach is applied to the backpropagation learning algorithm as well as the RPROP learning algorithm. The superiority of the new proposed method is demonstrated though many simulations. By applying the extension, the percentage of successful runs can be greatly increased and the average number of epochs to convergence can be well reduced on various problem instances. The behavior of the penalty values during training is also analyzed and their active role within the learning process is confirmed. Copyright © 2006 The Institute of Electronics, Information and Communication Engineers.
- 社団法人電子情報通信学会の論文
- 2006-08-01
著者
-
Jansen Boris
Graduate School Of Natural Science And Technology Kanazawa University
-
Nakayama Kenji
Graduate School Of Natural Science And Technology Kanazawa University
-
Nakayama Kenji
Kanazawa Univ. Ishikawa Jpn
-
Nakayama Kenji
Graduate School Of Natural Sci. & Tech. Kanazawa University
関連論文
- A stereo echo canceller with simultaneous input-sliding and sliding-period control (スマートインフォメディアシステム)
- A stereo echo canceller with simultaneous input-sliding and sliding-period control (信号処理)
- A distortion-free learning algorithm for feedforward multi-channel blind source separation
- A-4-41 A STEREO ECHO CANCELER WITH SIMULTANEOUS INPUT-SLIDING AND SLIDING-PERIOD CONTROL
- Performance of single and multi-reference NLMS noise canceller based on correlation between signal and noise
- An adaptive penalty-based learning extension for the backpropagation family
- A stable least square algorithm based on predictors and its application to fast Newton transversal filters
- A combined fast adaptive filter algorithm with an automatic switching method
- Metallothionein (MT) 1/2 expression in MT 1/2 and MT 3 knock-out mice and Long-Evans Cinnamon (LEC) rats