Explicit Sparse Transformer: Concentrated Attention Through Explicit Selection

TL;DR

The paper introduces Explicit Sparse Transformer, a model that enhances attention focus on relevant segments, improving performance and efficiency across NLP and vision tasks.

Contribution

It proposes a novel explicit selection mechanism for sparse attention, leading to better focus and faster inference compared to previous methods.

Findings

Improved model performance on NLP and vision tasks.

Achieved faster inference speed, twice as fast as sparsemax.

Comparable or better results than existing sparse attention methods.

Abstract

Self-attention based Transformer has demonstrated the state-of-the-art performances in a number of natural language processing tasks. Self-attention is able to model long-term dependencies, but it may suffer from the extraction of irrelevant information in the context. To tackle the problem, we propose a novel model called \textbf{Explicit Sparse Transformer}. Explicit Sparse Transformer is able to improve the concentration of attention on the global context through an explicit selection of the most relevant segments. Extensive experimental results on a series of natural language processing and computer vision tasks, including neural machine translation, image captioning, and language modeling, all demonstrate the advantages of Explicit Sparse Transformer in model performance. We also show that our proposed sparse attention method achieves comparable or better results than the previous sparse attention method, but significantly reduces training and testing time. For example, the inference speed is twice that of sparsemax in Transformer model. Code will be available at \url{https://github.com/lancopku/Explicit-Sparse-Transformer}