LSTM-based Deep Learning Models for Non-factoid Answer Selection

Abstract

In this paper, we apply a general deep learning (DL) framework for the answer selection task, which does not depend on manually defined features or linguistic tools. The basic framework is to build the embeddings of questions and answers based on bidirectional long short-term memory (biLSTM) models, and measure their closeness by cosine similarity. We further extend this basic model in two directions. One direction is to define a more composite representation for questions and answers by combining convolutional neural network with the basic framework. The other direction is to utilize a simple but efficient attention mechanism in order to generate the answer representation according to the question context. Several variations of models are provided. The models are examined by two datasets, including TREC-QA and InsuranceQA. Experimental results demonstrate that the proposed models substantially outperform several strong baselines.

added new experiments on TREC-QA

References in corpus (4)

• Sequence to Sequence Learning with Neural Networks
• Theano: new features and speed improvements
• Convolutional Neural Network Architectures for Matching Natural Language Sentences
• Deep Learning for Answer Sentence Selection

Cited by in corpus (83)

• Deep Learning Based Text Classification: A Comprehensive Review
• Attentive Pooling Networks
• Learning Natural Language Inference using Bidirectional LSTM model and Inner-Attention
• Information Retrieval: Recent Advances and Beyond
• Semantic Models for the First-stage Retrieval: A Comprehensive Review
• Bilateral Multi-Perspective Matching for Natural Language Sentences
• Rethinking Search: Making Domain Experts out of Dilettantes
• Modeling Multi-turn Conversation with Deep Utterance Aggregation
• ABCNN: Attention-Based Convolutional Neural Network for Modeling Sentence Pairs
• EKT: Exercise-aware Knowledge Tracing for Student Performance Prediction
• Quantum Language Model with Entanglement Embedding for Question Answering
• Res3ATN -- Deep 3D Residual Attention Network for Hand Gesture Recognition in Videos
• Sentence-State LSTM for Text Representation
• One-shot Learning for Question-Answering in Gaokao History Challenge
• CNM: An Interpretable Complex-valued Network for Matching
• Sequential Attention-based Network for Noetic End-to-End Response Selection
• Enhance word representation for out-of-vocabulary on Ubuntu dialogue corpus
• Query-Based Abstractive Summarization Using Neural Networks
• On the Benefit of Combining Neural, Statistical and External Features for Fake News Identification
• Deep Retrieval-Based Dialogue Systems: A Short Review
• Do Neural Nets Learn Statistical Laws behind Natural Language?
• Avoiding Echo-Responses in a Retrieval-Based Conversation System
• Deep Relevance Ranking Using Enhanced Document-Query Interactions
• Towards Understanding and Answering Multi-Sentence Recommendation Questions on Tourism
• A Study of BERT for Non-Factoid Question-Answering under Passage Length Constraints
• Adaptability of Neural Networks on Varying Granularity IR Tasks
• WikiPassageQA: A Benchmark Collection for Research on Non-factoid Answer Passage Retrieval
• An Attention-Based Word-Level Interaction Model: Relation Detection for Knowledge Base Question Answering
• A survey of Community Question Answering
• Sequential Matching Network: A New Architecture for Multi-turn Response Selection in Retrieval-based Chatbots
• Mining Implicit Relevance Feedback from User Behavior for Web Question Answering
• Advances in Natural Language Question Answering: A Review
• Recognizing Implicit Discourse Relations via Repeated Reading: Neural Networks with Multi-Level Attention
• Character-based NMT with Transformer
• A Sequential Matching Framework for Multi-turn Response Selection in Retrieval-based Chatbots
• Multi-modality Sensor Data Classification with Selective Attention
• Open Domain Question Answering Using Web Tables
• FAQ Retrieval using Query-Question Similarity and BERT-Based Query-Answer Relevance
• The World is Not Binary: Learning to Rank with Grayscale Data for Dialogue Response Selection
• M${}^2$S-Net: Multi-Modal Similarity Metric Learning based Deep Convolutional Network for Answer Selection
• Gated Group Self-Attention for Answer Selection
• Channel-aware Decoupling Network for Multi-turn Dialogue Comprehension
• Adversarial Training for Community Question Answer Selection Based on Multi-scale Matching
• Knowledge Distillation in Document Retrieval
• RAP-Net: Recurrent Attention Pooling Networks for Dialogue Response Selection
• SemEval-2017 Task 3: Community Question Answering
• Topic-Aware Multi-turn Dialogue Modeling
• Convolutional Neural Networks with Recurrent Neural Filters
• Why and How to Pay Different Attention to Phrase Alignments of Different Intensities
• A Multi-Size Neural Network with Attention Mechanism for Answer Selection
• Dialogue-oriented Pre-training
• An Audio-enriched BERT-based Framework for Spoken Multiple-choice Question Answering
• Chinese Medical Question Answer Matching Based on Interactive Sentence Representation Learning
• Joint Multitask Learning for Community Question Answering Using Task-Specific Embeddings
• Explainable Student Performance Prediction With Personalized Attention for Explaining Why A Student Fails
• Cross-Thought for Sentence Encoder Pre-training
• Explain by Evidence: An Explainable Memory-based Neural Network for Question Answering
• Review-based Question Generation with Adaptive Instance Transfer and Augmentation
• Generative Question Refinement with Deep Reinforcement Learning in Retrieval-based QA System
• Entity Synonym Discovery via Multipiece Bilateral Context Matching
• Grounded Textual Entailment
• Can AI Generate Love Advice?: Toward Neural Answer Generation for Non-Factoid Questions
• Machine Translation Evaluation Meets Community Question Answering
• Hierarchical Gated Recurrent Neural Tensor Network for Answer Triggering
• Cavs: A Vertex-centric Programming Interface for Dynamic Neural Networks
• Sequential Sentence Matching Network for Multi-turn Response Selection in Retrieval-based Chatbots
• Dual Attention Network for Product Compatibility and Function Satisfiability Analysis
• Hierarchical Ranking for Answer Selection
• Discriminative Information Retrieval for Knowledge Discovery
• Beyond [CLS] through Ranking by Generation
• Data-Efficient Methods for Dialogue Systems
• Multi-level Head-wise Match and Aggregation in Transformer for Textual Sequence Matching
• Semi-interactive Attention Network for Answer Understanding in Reverse-QA
• Hashing based Answer Selection
• Automatic Question-Answering Using A Deep Similarity Neural Network
• Image2song: Song Retrieval via Bridging Image Content and Lyric Words
• Representation Learning Models for Entity Search
• Indirect Supervision for Relation Extraction using Question-Answer Pairs
• Context-Aware Dialog Re-Ranking for Task-Oriented Dialog Systems
• Customizing Contextualized Language Models forLegal Document Reviews
• User-Inspired Posterior Network for Recommendation Reason Generation
• Convexification of Neural Graph
• Learning to Focus when Ranking Answers