Wider and Deeper, Cheaper and Faster: Tensorized LSTMs for Sequence Learning

Abstract

Long Short-Term Memory (LSTM) is a popular approach to boosting the ability of Recurrent Neural Networks to store longer term temporal information. The capacity of an LSTM network can be increased by widening and adding layers. However, usually the former introduces additional parameters, while the latter increases the runtime. As an alternative we propose the Tensorized LSTM in which the hidden states are represented by tensors and updated via a cross-layer convolution. By increasing the tensor size, the network can be widened efficiently without additional parameters since the parameters are shared across different locations in the tensor; by delaying the output, the network can be deepened implicitly with little additional runtime since deep computations for each timestep are merged into temporal computations of the sequence. Experiments conducted on five challenging sequence learning tasks show the potential of the proposed model.

Accepted by NIPS 2017

References in corpus (8)

• A Simple Way to Initialize Recurrent Networks of Rectified Linear Units
• Adaptive Computation Time for Recurrent Neural Networks
• Learning feed-forward one-shot learners
• Spatio-temporal video autoencoder with differentiable memory
• Combining Fully Convolutional and Recurrent Neural Networks for 3D Biomedical Image Segmentation
• Dilated Recurrent Neural Networks
• Optimizing Performance of Recurrent Neural Networks on GPUs
• Fast-Slow Recurrent Neural Networks

Cited by in corpus (4)

• The unreasonable effectiveness of the forget gate
• Explainable Tensorized Neural Ordinary Differential Equations forArbitrary-step Time Series Prediction
• Multi-variable LSTM neural network for autoregressive exogenous model
• Long Short-Term Attention