A Deep Architecture for Semantic Matching with Multiple Positional Sentence Representations

Abstract

Matching natural language sentences is central for many applications such as information retrieval and question answering. Existing deep models rely on a single sentence representation or multiple granularity representations for matching. However, such methods cannot well capture the contextualized local information in the matching process. To tackle this problem, we present a new deep architecture to match two sentences with multiple positional sentence representations. Specifically, each positional sentence representation is a sentence representation at this position, generated by a bidirectional long short term memory (Bi-LSTM). The matching score is finally produced by aggregating interactions between these different positional sentence representations, through $k$-Max pooling and a multi-layer perceptron. Our model has several advantages: (1) By using Bi-LSTM, rich context of the whole sentence is leveraged to capture the contextualized local information in each positional sentence representation; (2) By matching with multiple positional sentence representations, it is flexible to aggregate different important contextualized local information in a sentence to support the matching; (3) Experiments on different tasks such as question answering and sentence completion demonstrate the superiority of our model.

Accepted by AAAI-2016

References in corpus (2)

• LSTM: A Search Space Odyssey
• Convolutional Neural Network Architectures for Matching Natural Language Sentences

Cited by in corpus (58)

• Deep Learning Based Text Classification: A Comprehensive Review
• Advances and Challenges in Conversational Recommender Systems: A Survey
• Distilling Task-Specific Knowledge from BERT into Simple Neural Networks
• Match-SRNN: Modeling the Recursive Matching Structure with Spatial RNN
• Graph Neural Networks for Natural Language Processing: A Survey
• Neural ranking models for document retrieval
• Multiresolution Graph Attention Networks for Relevance Matching
• MatchZoo: A Toolkit for Deep Text Matching
• CAIL2019-SCM: A Dataset of Similar Case Matching in Legal Domain
• Towards Scalable and Reliable Capsule Networks for Challenging NLP Applications
• Sequential Attention-based Network for Noetic End-to-End Response Selection
• Neural Matching Models for Question Retrieval and Next Question Prediction in Conversation
• Multi-Cast Attention Networks for Retrieval-based Question Answering and Response Prediction
• Match-Ignition: Plugging PageRank into Transformer for Long-form Text Matching
• Induction Networks for Few-Shot Text Classification
• Learning To Retrieve: How to Train a Dense Retrieval Model Effectively and Efficiently
• A Deep Look into Neural Ranking Models for Information Retrieval
• Learning Contextualized Document Representations for Healthcare Answer Retrieval
• Modelling Interaction of Sentence Pair with coupled-LSTMs
• Fine-grained Event Categorization with Heterogeneous Graph Convolutional Networks
• Transfer Learning for Context-Aware Question Matching in Information-seeking Conversations in E-commerce
• InferCode: Self-Supervised Learning of Code Representations by Predicting Subtrees
• Matching Algorithms: Fundamentals, Applications and Challenges
• A Context-Aware User-Item Representation Learning for Item Recommendation
• Promotion of Answer Value Measurement with Domain Effects in Community Question Answering Systems
• Improving Retrieval Modeling Using Cross Convolution Networks And Multi Frequency Word Embedding
• Matching Article Pairs with Graphical Decomposition and Convolutions
• Gated Group Self-Attention for Answer Selection
• Lattice CNNs for Matching Based Chinese Question Answering
• Finding Similar Medical Questions from Question Answering Websites
• LATTE: Latent Type Modeling for Biomedical Entity Linking
• Pre-trained Language Model for Web-scale Retrieval in Baidu Search
• RLTM: An Efficient Neural IR Framework for Long Documents
• IART: Intent-aware Response Ranking with Transformers in Information-seeking Conversation Systems
• Learning Fine-grained Fact-Article Correspondence in Legal Cases
• Heterogeneous Network Embedding for Deep Semantic Relevance Match in E-commerce Search
• Word Embedding based Edit Distance
• Bangla Text Classification using Transformers
• Curriculum Learning Strategies for IR: An Empirical Study on Conversation Response Ranking
• Biomedical Evidence Generation Engine
• Match${}^2$: A Matching over Matching Model for Similar Question Identification
• Impact of Training Dataset Size on Neural Answer Selection Models
• Toward the Understanding of Deep Text Matching Models for Information Retrieval
• Review-guided Helpful Answer Identification in E-commerce
• An In-depth Analysis of Passage-Level Label Transfer for Contextual Document Ranking
• BiTe-GCN: A New GCN Architecture via BidirectionalConvolution of Topology and Features on Text-Rich Networks
• Neural Relation Prediction for Simple Question Answering over Knowledge Graph
• An Efficient Approach for Geo-Multimedia Cross-Modal Retrieval
• Sentence transition matrix: An efficient approach that preserves sentence semantics
• Learning to Expand: Reinforced Pseudo-relevance Feedback Selection for Information-seeking Conversations
• Hashing based Answer Selection
• Locate Who You Are: Matching Geo-location to Text for User Identity Linkage
• Streaming Social Event Detection and Evolution Discovery in Heterogeneous Information Networks
• Retrieving and ranking short medical questions with two stages neural matching model
• Learning Fast Matching Models from Weak Annotations
• YES SIR!Optimizing Semantic Space of Negatives with Self-Involvement Ranker
• Place Deduplication with Embeddings
• ScriptWriter: Narrative-Guided Script Generation