Improving Semantic Matching through Dependency-Enhanced Pre-trained Model with Adaptive Fusion

TL;DR

This paper introduces DAFA, a novel method that explicitly incorporates dependency structure into pre-trained models like BERT, enhancing semantic matching performance by adaptively fusing structural and semantic information.

Contribution

The paper proposes a new dependency-enhanced attention mechanism with adaptive fusion, improving semantic matching by integrating dependency prior knowledge into pre-trained models.

Findings

Achieves state-of-the-art results on 10 datasets.

Demonstrates the effectiveness of dependency-aware attention.

Provides better interpretability of semantic matching models.

Abstract

Transformer-based pre-trained models like BERT have achieved great progress on Semantic Sentence Matching. Meanwhile, dependency prior knowledge has also shown general benefits in multiple NLP tasks. However, how to efficiently integrate dependency prior structure into pre-trained models to better model complex semantic matching relations is still unsettled. In this paper, we propose the \textbf{D}ependency-Enhanced \textbf{A}daptive \textbf{F}usion \textbf{A}ttention (\textbf{DAFA}), which explicitly introduces dependency structure into pre-trained models and adaptively fuses it with semantic information. Specifically, \textbf{\emph{(i)}} DAFA first proposes a structure-sensitive paradigm to construct a dependency matrix for calibrating attention weights. It adopts an adaptive fusion module to integrate the obtained dependency information and the original semantic signals. Moreover, DAFA reconstructs the attention calculation flow and provides better interpretability. By applying it on BERT, our method achieves state-of-the-art or competitive performance on 10 public datasets, demonstrating the benefits of adaptively fusing dependency structure in semantic matching task.