A Fully Tensorized Recurrent Neural Network

Abstract

Recurrent neural networks (RNNs) are powerful tools for sequential modeling, but typically require significant overparameterization and regularization to achieve optimal performance. This leads to difficulties in the deployment of large RNNs in resource-limited settings, while also introducing complications in hyperparameter selection and training. To address these issues, we introduce a "fully tensorized" RNN architecture which jointly encodes the separate weight matrices within each recurrent cell using a lightweight tensor-train (TT) factorization. This approach represents a novel form of weight sharing which reduces model size by several orders of magnitude, while still maintaining similar or better performance compared to standard RNNs. Experiments on image classification and speaker verification tasks demonstrate further benefits for reducing inference times and stabilizing model training and hyperparameter selection.

References in corpus (9)

• PyTorch: An Imperative Style, High-Performance Deep Learning Library
• Distilling the Knowledge in a Neural Network
• UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction
• Tensorizing Neural Networks
• Compressing Recurrent Neural Network with Tensor Train
• Ultimate tensorization: compressing convolutional and FC layers alike
• Tensor-Train Recurrent Neural Networks for Video Classification
• Tensorized Embedding Layers for Efficient Model Compression
• Compressing Recurrent Neural Networks Using Hierarchical Tucker Tensor Decomposition