Simple and Effective Text Matching with Richer Alignment Features

Abstract

In this paper, we present a fast and strong neural approach for general purpose text matching applications. We explore what is sufficient to build a fast and well-performed text matching model and propose to keep three key features available for inter-sequence alignment: original point-wise features, previous aligned features, and contextual features while simplifying all the remaining components. We conduct experiments on four well-studied benchmark datasets across tasks of natural language inference, paraphrase identification and answer selection. The performance of our model is on par with the state-of-the-art on all datasets with much fewer parameters and the inference speed is at least 6 times faster compared with similarly performed ones.

arXiv comment11 pages, 7 tables, 3 figures, accepted by ACL 2019

References in corpus (5)

• Natural Language Processing (almost) from Scratch
• Enhanced LSTM for Natural Language Inference
• Deep Learning for Answer Sentence Selection
• Natural Language Inference over Interaction Space
• Stochastic Answer Networks for Natural Language Inference

Cited by in corpus (9)

• Match-Ignition: Plugging PageRank into Transformer for Long-form Text Matching
• AliCoCo: Alibaba E-commerce Cognitive Concept Net
• Learning Fine-grained Fact-Article Correspondence in Legal Cases
• Matching Text with Deep Mutual Information Estimation
• A multi-perspective combined recall and rank framework for Chinese procedure terminology normalization
• Combining word embeddings and convolutional neural networks to detect duplicated questions
• Answer Ranking for Product-Related Questions via Multiple Semantic Relations Modeling
• K-AID: Enhancing Pre-trained Language Models with Domain Knowledge for Question Answering
• Query-bag Matching with Mutual Coverage for Information-seeking Conversations in E-commerce