Rethinking Full Connectivity in Recurrent Neural Networks

Abstract

Recurrent neural networks (RNNs) are omnipresent in sequence modeling tasks. Practical models usually consist of several layers of hundreds or thousands of neurons which are fully connected. This places a heavy computational and memory burden on hardware, restricting adoption in practical low-cost and low-power devices. Compared to fully convolutional models, the costly sequential operation of RNNs severely hinders performance on parallel hardware. This paper challenges the convention of full connectivity in RNNs. We study structurally sparse RNNs, showing that they are well suited for acceleration on parallel hardware, with a greatly reduced cost of the recurrent operations as well as orders of magnitude less recurrent weights. Extensive experiments on challenging tasks ranging from language modeling and speech recognition to video action recognition reveal that structurally sparse RNNs achieve competitive performance as compared to fully-connected networks. This allows for using large sparse RNNs for a wide range of real-world tasks that previously were too costly with fully connected networks.

References in corpus (9)

• Two-Stream Convolutional Networks for Action Recognition in Videos
• UCF101: A Dataset of 101 Human Actions Classes From Videos in The Wild
• An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling
• WaveNet: A Generative Model for Raw Audio
• Deep Speech 2: End-to-End Speech Recognition in English and Mandarin
• Quasi-Recurrent Neural Networks
• An Analysis of Neural Language Modeling at Multiple Scales
• Block-Sparse Recurrent Neural Networks
• Going Wider: Recurrent Neural Network With Parallel Cells

Cited by in corpus (3)

• Block-wise Dynamic Sparseness
• DARB: A Density-Aware Regular-Block Pruning for Deep Neural Networks
• Efficient Micro-Structured Weight Unification and Pruning for Neural Network Compression