Generating Informative and Diverse Conversational Responses via Adversarial Information Maximization

Abstract

Responses generated by neural conversational models tend to lack informativeness and diversity. We present Adversarial Information Maximization (AIM), an adversarial learning strategy that addresses these two related but distinct problems. To foster response diversity, we leverage adversarial training that allows distributional matching of synthetic and real responses. To improve informativeness, our framework explicitly optimizes a variational lower bound on pairwise mutual information between query and response. Empirical results from automatic and human evaluations demonstrate that our methods significantly boost informativeness and diversity.

NIPS 2018

Cited by in corpus (67)

• Towards a Human-like Open-Domain Chatbot
• DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation
• Meta-evaluation of Conversational Search Evaluation Metrics
• Improving Sequence-to-Sequence Learning via Optimal Transport
• Neural Language Generation: Formulation, Methods, and Evaluation
• An Empirical Investigation of Pre-Trained Transformer Language Models for Open-Domain Dialogue Generation
• Trading Off Diversity and Quality in Natural Language Generation
• Progressive Memory Banks for Incremental Domain Adaptation
• Conversing by Reading: Contentful Neural Conversation with On-demand Machine Reading
• Improving Background Based Conversation with Context-aware Knowledge Pre-selection
• Deep Retrieval-Based Dialogue Systems: A Short Review
• Grounded Conversation Generation as Guided Traverses in Commonsense Knowledge Graphs
• A Survey of Document Grounded Dialogue Systems (DGDS)
• Towards Generating Long and Coherent Text with Multi-Level Latent Variable Models
• DAL: Dual Adversarial Learning for Dialogue Generation
• POINTER: Constrained Progressive Text Generation via Insertion-based Generative Pre-training
• On the Generation of Medical Dialogues for COVID-19
• Contrastive Learning with Adversarial Perturbations for Conditional Text Generation
• Thinking Globally, Acting Locally: Distantly Supervised Global-to-Local Knowledge Selection for Background Based Conversation
• Knowledge Injection into Dialogue Generation via Language Models
• Jointly Optimizing Diversity and Relevance in Neural Response Generation
• Self-Supervised Dialogue Learning
• Neural Response Generation with Meta-Words
• Generate, Prune, Select: A Pipeline for Counterspeech Generation against Online Hate Speech
• Retrieval-Enhanced Adversarial Training for Neural Response Generation
• Conversations Are Not Flat: Modeling the Dynamic Information Flow across Dialogue Utterances
• Negative Training for Neural Dialogue Response Generation
• DLGNet: A Transformer-based Model for Dialogue Response Generation
• Jointly Learning Semantic Parser and Natural Language Generator via Dual Information Maximization
• Regularizing Dialogue Generation by Imitating Implicit Scenarios
• Diversifying Dialogue Generation with Non-Conversational Text
• Narrative Incoherence Detection
• SMRT Chatbots: Improving Non-Task-Oriented Dialog with Simulated Multiple Reference Training
• Diversifying Task-oriented Dialogue Response Generation with Prototype Guided Paraphrasing
• CliniQG4QA: Generating Diverse Questions for Domain Adaptation of Clinical Question Answering
• Improving the Robustness of QA Models to Challenge Sets with Variational Question-Answer Pair Generation
• Towards Unique and Informative Captioning of Images
• A Discrete CVAE for Response Generation on Short-Text Conversation
• Knowledge-Grounded Response Generation with Deep Attentional Latent-Variable Model
• An Adversarially-Learned Turing Test for Dialog Generation Models
• Towards Coherent and Cohesive Long-form Text Generation
• Adversarial Bootstrapping for Dialogue Model Training
• Guiding Topic Flows in the Generative Chatbot by Enhancing the ConceptNet with the Conversation Corpora
• Controllable and Diverse Text Generation in E-commerce
• Promoting Diversity for End-to-End Conversation Response Generation
• Group-wise Contrastive Learning for Neural Dialogue Generation
• Modeling Explicit Concerning States for Reinforcement Learning in Visual Dialogue
• Keyword-Guided Neural Conversational Model
• HSCJN: A Holistic Semantic Constraint Joint Network for Diverse Response Generation
• Proposal Towards a Personalized Knowledge-powered Self-play Based Ensemble Dialog System
• Controllable Dialogue Generation with Disentangled Multi-grained Style Specification and Attribute Consistency Reward
• Modeling Psychotherapy Dialogues with Kernelized Hashcode Representations: A Nonparametric Information-Theoretic Approach
• Lightweight Decoding Strategies for Increasing Specificity
• Enhancing Visual Dialog Questioner with Entity-based Strategy Learning and Augmented Guesser
• Identifying Untrustworthy Samples: Data Filtering for Open-domain Dialogues with Bayesian Optimization
• Data-Efficient Methods for Dialogue Systems
• A Dataset and Baselines for Multilingual Reply Suggestion
• Multi-Modal Dialogue State Tracking for Playing GuessWhich Game
• Exposure Bias versus Self-Recovery: Are Distortions Really Incremental for Autoregressive Text Generation?
• Sentence-Permuted Paragraph Generation
• Cross-modal Image Retrieval with Deep Mutual Information Maximization
• Learning from Easy to Complex: Adaptive Multi-curricula Learning for Neural Dialogue Generation
• Towards Multimodal Response Generation with Exemplar Augmentation and Curriculum Optimization
• Towards Efficiently Diversifying Dialogue Generation via Embedding Augmentation
• Sequential Sentence Matching Network for Multi-turn Response Selection in Retrieval-based Chatbots
• On Estimating the Training Cost of Conversational Recommendation Systems
• Building A User-Centric and Content-Driven Socialbot