Modeling Multi-turn Conversation with Deep Utterance Aggregation

Abstract

Multi-turn conversation understanding is a major challenge for building intelligent dialogue systems. This work focuses on retrieval-based response matching for multi-turn conversation whose related work simply concatenates the conversation utterances, ignoring the interactions among previous utterances for context modeling. In this paper, we formulate previous utterances into context using a proposed deep utterance aggregation model to form a fine-grained context representation. In detail, a self-matching attention is first introduced to route the vital information in each utterance. Then the model matches a response with each refined utterance and the final matching score is obtained after attentive turns aggregation. Experimental results show our model outperforms the state-of-the-art methods on three multi-turn conversation benchmarks, including a newly introduced e-commerce dialogue corpus.

arXiv commentProceedings of the 27th International Conference on Computational Linguistics (COLING 2018)

References in corpus (9)

• Efficient Estimation of Word Representations in Vector Space
• LSTM-based Deep Learning Models for Non-factoid Answer Selection
• Improved Deep Learning Baselines for Ubuntu Corpus Dialogs
• BBQ-Networks: Efficient Exploration in Deep Reinforcement Learning for Task-Oriented Dialogue Systems
• Deep Enhanced Representation for Implicit Discourse Relation Recognition
• Multiresolution Recurrent Neural Networks: An Application to Dialogue Response Generation
• Neural Response Generation with Dynamic Vocabularies
• Coherent Dialogue with Attention-based Language Models
• One-shot Learning for Question-Answering in Gaokao History Challenge

Cited by in corpus (37)

• Poly-encoders: Transformer Architectures and Pre-training Strategies for Fast and Accurate Multi-sentence Scoring
• DUMA: Reading Comprehension with Transposition Thinking
• Challenges in Building Intelligent Open-domain Dialog Systems
• The JDDC Corpus: A Large-Scale Multi-Turn Chinese Dialogue Dataset for E-commerce Customer Service
• One-shot Learning for Question-Answering in Gaokao History Challenge
• Sequential Attention-based Network for Noetic End-to-End Response Selection
• Explicit Contextual Semantics for Text Comprehension
• SG-Net: Syntax-Guided Machine Reading Comprehension
• Semantics-aware BERT for Language Understanding
• A comprehensive solution to retrieval-based chatbot construction
• Speaker-Aware BERT for Multi-Turn Response Selection in Retrieval-Based Chatbots
• Pre-Trained and Attention-Based Neural Networks for Building Noetic Task-Oriented Dialogue Systems
• A Survey of Document Grounded Dialogue Systems (DGDS)
• Learning an Effective Context-Response Matching Model with Self-Supervised Tasks for Retrieval-based Dialogues
• Interactive Matching Network for Multi-Turn Response Selection in Retrieval-Based Chatbots
• Matching Algorithms: Fundamentals, Applications and Challenges
• SJTU-NICT's Supervised and Unsupervised Neural Machine Translation Systems for the WMT20 News Translation Task
• Ultra-Fast, Low-Storage, Highly Effective Coarse-grained Selection in Retrieval-based Chatbot by Using Deep Semantic Hashing
• Effective Subword Segmentation for Text Comprehension
• Partner Matters! An Empirical Study on Fusing Personas for Personalized Response Selection in Retrieval-Based Chatbots
• Topic-Aware Multi-turn Dialogue Modeling
• Dialogue History Matters! Personalized Response Selectionin Multi-turn Retrieval-based Chatbots
• Curriculum Learning Strategies for IR: An Empirical Study on Conversation Response Ranking
• Advances in Multi-turn Dialogue Comprehension: A Survey
• Automatic Intent-Slot Induction for Dialogue Systems
• Self-attention Comparison Module for Boosting Performance on Retrieval-based Open-Domain Dialog Systems
• Multi-turn Dialog System on Single-turn Data in Medical Domain
• Multi-turn Dialogue Reading Comprehension with Pivot Turns and Knowledge
• Open Vocabulary Learning for Neural Chinese Pinyin IME
• Effective Representation for Easy-First Dependency Parsing
• Utterance-to-Utterance Interactive Matching Network for Multi-Turn Response Selection in Retrieval-Based Chatbots
• Recall and Learn: A Memory-augmented Solver for Math Word Problems
• Stylistic Retrieval-based Dialogue System with Unparallel Training Data
• Effective Character-augmented Word Embedding for Machine Reading Comprehension
• Open Named Entity Modeling from Embedding Distribution
• Accurate Word Representations with Universal Visual Guidance
• Improving Multi-Turn Response Selection Models with Complementary Last-Utterance Selection by Instance Weighting