Simple and conditioned adaptive behavior from Kalman filter trained recurrent networks

被引:46
作者
Feldkamp, LA [1 ]
Prokhorov, DV [1 ]
Feldkamp, TA [1 ]
机构
[1] Ford Motor Co, Rea & Adv Engn, Dearborn, MI 48121 USA
关键词
recurrent neural network; input-output; adaptive behavior;
D O I
10.1016/S0893-6080(03)00127-8
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
We illustrate the ability of a fixed-weight neural network, trained with Kalman filter methods, to perform tasks that are usually entrusted to an explicitly adaptive system. Following a simple example, we demonstrate that such a network can be trained to exhibit input-output behavior that depends on which of two conditioning tasks was performed a substantial number of time steps in the past. This behavior can also be made to survive an intervening interference task. (C) 2003 Published by Elsevier Science Ltd.
引用
收藏
页码:683 / 689
页数:7
相关论文
共 14 条
[1]  
ANNASWAMY A, 1996, UNPUB ABILITY RECURR
[2]  
Feldkamp LA, 1996, IEEE IJCNN, P155, DOI 10.1109/ICNN.1996.548883
[3]   A signal processing framework based on dynamic neural networks with application to problems in adaptation, filtering, and classification [J].
Feldkamp, LA ;
Puskorius, GV .
PROCEEDINGS OF THE IEEE, 1998, 86 (11) :2259-2277
[4]   Recurrent neural network training by nprKF joint estimation [J].
Feldkamp, LA ;
Feldkamp, TM ;
Prokhorov, DV .
PROCEEDING OF THE 2002 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, VOLS 1-3, 2002, :2086-2091
[5]  
Feldkamp LA, 2001, IEEE IJCNN, P109, DOI 10.1109/IJCNN.2001.939001
[6]   Adaptive behavior from fixed weight networks [J].
Feldkamp, LA ;
Puskorius, GV ;
Moore, PC .
INFORMATION SCIENCES, 1997, 98 (1-4) :217-235
[7]  
Feldkamp LA, 1997, 1997 IEEE INTERNATIONAL CONFERENCE ON NEURAL NETWORKS, VOLS 1-4, P773, DOI 10.1109/ICNN.1997.616120
[8]  
FELDKAMP LA, 2001, P 11 YAL WORKSH AD L, P78
[9]  
Hochreiter S, 2001, LECT NOTES COMPUT SC, V2130, P87
[10]   Adaptive behavior with fixed weights in RNN: An overview [J].
Prokhorov, DV ;
Feldkamp, LA ;
Tyukin, IY .
PROCEEDING OF THE 2002 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, VOLS 1-3, 2002, :2018-2022