Language Modeling with Gated Convolutional Networks

TL;DR

This paper introduces a convolutional neural network approach with gating mechanisms for language modeling, achieving state-of-the-art results and significantly faster inference compared to recurrent models.

Contribution

It develops a finite context convolutional model with a novel gating mechanism, outperforming previous non-recurrent models on large-scale benchmarks.

Findings

Achieved state-of-the-art on WikiText-103

Competitive results on Google Billion Words

Reduced sentence scoring latency by an order of magnitude

Abstract

The pre-dominant approach to language modeling to date is based on recurrent neural networks. Their success on this task is often linked to their ability to capture unbounded context. In this paper we develop a finite context approach through stacked convolutions, which can be more efficient since they allow parallelization over sequential tokens. We propose a novel simplified gating mechanism that outperforms Oord et al (2016) and investigate the impact of key architectural decisions. The proposed approach achieves state-of-the-art on the WikiText-103 benchmark, even though it features long-term dependencies, as well as competitive results on the Google Billion Words benchmark. Our model reduces the latency to score a sentence by an order of magnitude compared to a recurrent baseline. To our knowledge, this is the first time a non-recurrent approach is competitive with strong recurrent models on these large scale language tasks.