Contextualized Non-local Neural Networks for Sequence Learning

TL;DR

This paper introduces contextualized non-local neural networks (CN3) that combine self-attention and graph neural networks to improve sequence learning by dynamically modeling task-specific structures and local dependencies.

Contribution

The paper presents CN3, a novel model that integrates the strengths of self-attention and GNNs for enhanced sequence learning and interpretability.

Findings

Outperforms competitive baselines on ten NLP tasks.

Effectively models task-specific dependency structures.

Provides improved interpretability of learned structures.

Abstract

Recently, a large number of neural mechanisms and models have been proposed for sequence learning, of which self-attention, as exemplified by the Transformer model, and graph neural networks (GNNs) have attracted much attention. In this paper, we propose an approach that combines and draws on the complementary strengths of these two methods. Specifically, we propose contextualized non-local neural networks (CN$^{\textbf{3}}$), which can both dynamically construct a task-specific structure of a sentence and leverage rich local dependencies within a particular neighborhood. Experimental results on ten NLP tasks in text classification, semantic matching, and sequence labeling show that our proposed model outperforms competitive baselines and discovers task-specific dependency structures, thus providing better interpretability to users.