| 摘要 |
<p>A method for the supervised teaching of a recurrent neutral network (RNN) is disclosed. A typical embodiment of the method utilizes a large (50 units or more), randomly initialized RNN with a globally stable dynamics. During the training period, the output units of this RNN are teacher-forced to follow the desired output signal. During this period, activations from all hidden units are recorded. At the end of the teaching period, these recorded data are used as input for a method which computes new weights of those connections that feed into the output units. The method is distinguished from existing training methods for RNNs through the following characteristics: (1) Only the weights of connections to output units are changed by learning - existing methods for teaching recurrent networks adjust all network weights. (2) The internal dynamics of large networks are used as a 'reservoir' of dynamical components which are not changed, but only newly combined by the learning procedure - existing methods use small networks, whose internal dynamics are themselves competely re-shaped through learning.</p> |
