Abstract:The goal of sentence matching is to determine the semantic relation between two sentences, which is the basis of many downstream tasks in natural language processing, such as question answering and information retrieval. Recent studies using attention mechanism to align the elements of two sentences have shown promising results in capturing semantic similarity/relevance. Most existing methods mainly focus on the design of multi-layer attention network, however, some critical issues have not been dealt with well: 1) the higher attention layer is easily affected by error propagation because it relies on the alignment results of preceding attentions; 2) models have the risk of losing low-layer semantic features with the increase of network depth; and 3) the approach of capturing global matching information brings about large computing complexity for model training. To this end, we propose a Deep Bi-Directional Interaction Network (DBDIN) to solve these issues, which captures semantic relatedness from two directions and each direction employs multiple attention-based interaction units. To be specific, the attention of each interaction unit will repeatedly focus on the original sentence representation of another one for semantic alignment, which alleviates the error propagation problem by attending to a fixed semantic representation. Then we design deep fusion to aggregate and propagate attention information from low layers to high layers, which effectively retains low-layer semantic features for subsequential interactions. Moreover, we introduce a self-attention mechanism at last to enhance global matching information with smaller model complexity. We conduct experiments on natural language inference and paraphrase identification tasks with three benchmark datasets SNLI, SciTail and Quora. Experimental results demonstrate that our proposed method can achieve significant improvements over baseline systems without using any external knowledge. Additionally, we conduct interpretable study to disclose how our deep interaction network with attention can benefit sentence matching, which provides a reference for future model design. Ablation studies and visualization analyses further verify that our model can better capture interactive information between two sentences, and the proposed components are indeed able to help modeling semantic relation more precisely.

Dependent Multilevel Interaction Network For Natural Language Inference

Asynchronous Deep Interaction Network for Natural Language Inference.

Attention-Fused Deep Matching Network for Natural Language Inference

Convolutional Interaction Network for Natural Language Inference

Deep Bi-Directional Interaction Network for Sentence Matching

A Sentence Interaction Network for Modeling Dependence Between Sentences.

Natural Language Inference Using Lstm Model With Sentence Fusion

Multi-turn Inference Matching Network for Natural Language Inference

Research on Multi-granularity Sentence Interaction Natural Language Inference Based on At-tention Mechanism

Collaborative Attention Network for Natural Language Inference

Enhanced Lstm For Natural Language Inference

SDF-NN: A Deep Neural Network with Semantic Dropping and Fusion for Natural Language Inference

Multilevel Image-Enhanced Sentence Representation Net for Natural Language Inference

Multi-level Interaction Network for Multi-Modal Rumor Detection.

MLI: A Multi-level Inference Mechanism for User Attributes in Social Networks

Multi-Feature Fusion Transformer for Natural Language Inference

Research on Attention Memory Networks As a Model for Learning Natural Language Inference.

Abductive natural language inference by interactive model with structural loss

Enhancing and Combining Sequential and Tree LSTM for Natural Language Inference.

MPII: Multi-Level Mutual Promotion for Inference and Interpretation

Asymmetric feature interaction for interpreting model predictions