A theoretically grounded application of dropout in recurrent neural networks

Conference item

Recurrent neural networks (RNNs) stand at the forefront of many recent developments in deep learning. Yet a major difficulty with these models is their tendency to overfit, with dropout shown to fail when applied to recurrent layers. Recent results at the intersection of Bayesian modelling and deep learning offer a Bayesian interpretation of common deep learning techniques such as dropout. This grounding of dropout in approximate Bayesian inference suggests an extension of the theoretical results, offering insights into the use of dropout with RNN models. We apply this new variational inference based dropout technique in LSTM and GRU models, assessing it on language modelling and sentiment analysis tasks. The new approach outperforms existing techniques, and to the best of our knowledge improves on the single model state-of-the-art in language modelling with the Penn Treebank (73.4 test perplexity). This extends our arsenal of variational tools in deep learning.

Authors: Gal, Y; Ghahramani, Z

• Publication status: Published
• Peer review status: Peer reviewed
• Publisher: Massachusetts Institute of Technology Press
• Host title: Advances in Neural Information Processing Systems 29 (NIPS 2016)
• Journal: NeurIPS Proceedings 2016
• Volume: 29
• Pages: 1019-1027
• Publication date: 2016-12-01
• Acceptance date: 2016-08-12
• Event title: 30th Conference on Neural Information Processing Systems (NIPS 2016)
• Event location: Barcelona, Spain
• Event start date: 2016-12-05
• Event end date: 2016-12-10
• ISSN: 1049-5258
• Language: English