Conditioned adaptive behavior from Kalman filter trained recurrent networks
We demonstrate that a fixed-weight neural network can be trained with Kalman filter methods to exhibit input-output behavior that depends on which of two conditioning tasks had been performed a substantial number of time steps in the past. This behavior can also be made to survive an intervening int...
Saved in:
Published in | Proceedings of the International Joint Conference on Neural Networks, 2003 Vol. 4; pp. 3017 - 3021 vol.4 |
---|---|
Main Authors | , , |
Format | Conference Proceeding |
Language | English |
Published |
IEEE
2003
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | We demonstrate that a fixed-weight neural network can be trained with Kalman filter methods to exhibit input-output behavior that depends on which of two conditioning tasks had been performed a substantial number of time steps in the past. This behavior can also be made to survive an intervening interference task. |
---|---|
ISBN: | 9780780378988 0780378989 |
ISSN: | 1098-7576 1558-3902 |
DOI: | 10.1109/IJCNN.2003.1224052 |