R$^2$-Net: Relation of Relation Learning Network for Sentence Semantic Matching

TL;DR

R2-Net is a novel neural network model that enhances sentence semantic matching by integrating global and local sentence encoding with label-guided relation learning and triplet loss for finer relation discrimination.

Contribution

The paper introduces R2-Net, which leverages label-guided relation of relation classification and triplet loss to improve semantic matching accuracy over existing models.

Findings

R2-Net outperforms baseline models on two semantic matching tasks.

The use of relation of relation classification improves label relation understanding.

Triplet loss enhances intra- and inter-class relation discrimination.

Abstract

Sentence semantic matching is one of the fundamental tasks in natural language processing, which requires an agent to determine the semantic relation among input sentences. Recently, deep neural networks have achieved impressive performance in this area, especially BERT. Despite the effectiveness of these models, most of them treat output labels as meaningless one-hot vectors, underestimating the semantic information and guidance of relations that these labels reveal, especially for tasks with a small number of labels. To address this problem, we propose a Relation of Relation Learning Network (R2-Net) for sentence semantic matching. Specifically, we first employ BERT to encode the input sentences from a global perspective. Then a CNN-based encoder is designed to capture keywords and phrase information from a local perspective. To fully leverage labels for better relation information extraction, we introduce a self-supervised relation of relation classification task for guiding R2-Net to consider more about labels. Meanwhile, a triplet loss is employed to distinguish the intra-class and inter-class relations in a finer granularity. Empirical experiments on two sentence semantic matching tasks demonstrate the superiority of our proposed model. As a byproduct, we have released the codes to facilitate other researches.