论文信息 - Learning internal representations of pattern sequences in a neural network with adaptive time-delays

Learning internal representations of pattern sequences in a neural network with adaptive time-delays

The Tempo-Network is an artificial neural network with both adaptive weights and adaptive time delays. U Bodenhausen (see ibid., vol.1, p.597-600, 1990) showed that the network is able to work as a autoassociator for sequences of patterns which are fed into the network one after the other. The patterns of the sequence are stepped through in time. The learning rules and the internal representations in the hidden layer are discussed. Simulations with the Tempo-Network are discussed. Neural networks with adaptive time delays seem to be interesting for tasks where the sequence contains nonadjacent events

Ulrich Bodenhausen

[1] Kevin J. Lang. A time delay neural network architecture for speech recognition , 1989 .

[2] P. Dicke,et al. Feature linking via stimulus-evoked oscillations: experimental results from cat visual cortex and functional implications from a network model , 1989, International 1989 Joint Conference on Neural Networks.

[3] Elie Bienenstock,et al. A neural network for the retrieval of superimposed connection patterns , 1987 .

[4] Geoffrey E. Hinton,et al. Learning Representations by Recirculation , 1987, NIPS.

[5] Alexander H. Waibel,et al. Modular Construction of Time-Delay Neural Networks for Speech Recognition , 1989, Neural Computation.

[6] Erich J. Smythe. Temporal Representations in a Connectionist Speech System , 1988, NIPS.

[7] Geoffrey E. Hinton,et al. Phoneme recognition using time-delay neural networks , 1989, IEEE Trans. Acoust. Speech Signal Process..

[8] J J Hopfield,et al. Neural computation by concentrating information in time. , 1987, Proceedings of the National Academy of Sciences of the United States of America.