Several variants of recurrent neural networks (RNNs) with orthogonal or unitary recurrent matrices have recently been developed to mitigate the vanishing/exploding gradient problem and to model long-term dependencies of sequences. However, with the eigenvalues of the recurrent matrix on the unit circle, the recurrent state retains all input information which may unnecessarily consume model capacity. In this paper, we address this issue by proposing an architecture that expands upon an orthogonal/unitary RNN with a state that is generated by a recurrent matrix with eigenvalues in the unit disc. Any input to this state dissipates in time and is replaced with new inputs, simulating short-term memory. A gradient descent algorithm is derived for learning such a recurrent matrix. The resulting method, called the Eigenvalue Normalized RNN (ENRNN), is shown to be highly competitive in several experiments.
|Title of host publication||AAAI 2020 - 34th AAAI Conference on Artificial Intelligence|
|Number of pages||8|
|State||Published - 2020|
|Event||34th AAAI Conference on Artificial Intelligence, AAAI 2020 - New York, United States|
Duration: Feb 7 2020 → Feb 12 2020
|Name||AAAI 2020 - 34th AAAI Conference on Artificial Intelligence|
|Conference||34th AAAI Conference on Artificial Intelligence, AAAI 2020|
|Period||2/7/20 → 2/12/20|
Bibliographical noteFunding Information:
Acknowledgments. This research was supported in part by NSF under grants DMS-1821144 and DMS-1620082.
© 2020, Association for the Advancement of Artificial Intelligence.
ASJC Scopus subject areas
- Artificial Intelligence