Skip RNN: Learning to Skip State Updates in Recurrent Neural Networks

TL;DR

This paper introduces Skip RNN, a model that learns to skip state updates in recurrent neural networks, reducing computational cost while maintaining or improving performance on sequence tasks.

Contribution

The paper proposes a novel Skip RNN model that learns to skip state updates, effectively reducing computational complexity and addressing training challenges in long sequence modeling.

Findings

Reduces the number of RNN updates needed

Maintains or improves performance on sequence tasks

Demonstrates efficiency gains in various experiments

Abstract

Recurrent Neural Networks (RNNs) continue to show outstanding performance in sequence modeling tasks. However, training RNNs on long sequences often face challenges like slow inference, vanishing gradients and difficulty in capturing long term dependencies. In backpropagation through time settings, these issues are tightly coupled with the large, sequential computational graph resulting from unfolding the RNN in time. We introduce the Skip RNN model which extends existing RNN models by learning to skip state updates and shortens the effective size of the computational graph. This model can also be encouraged to perform fewer state updates through a budget constraint. We evaluate the proposed model on various tasks and show how it can reduce the number of required RNN updates while preserving, and sometimes even improving, the performance of the baseline RNN models. Source code is publicly available at https://imatge-upc.github.io/skiprnn-2017-telecombcn/ .