Optimizing Dialogue Management with Reinforcement Learning: Experiments with the NJFun System

Abstract

Designing the dialogue policy of a spoken dialogue system involves many nontrivial choices. This paper presents a reinforcement learning approach for automatically optimizing a dialogue policy, which addresses the technical challenges in applying reinforcement learning to a working dialogue system with human users. We report on the design, construction and empirical evaluation of NJFun, an experimental spoken dialogue system that provides users with access to information about fun things to do in New Jersey. Our results show that by optimizing its performance via reinforcement learning, NJFun measurably improves system performance.

References in corpus (2)

• Metrics for Evaluating Dialogue Strategies in a Spoken Language System
• Tracking Initiative in Collaborative Dialogue Interactions

Cited by in corpus (30)

• A Brief Survey of Deep Reinforcement Learning
• Deep Reinforcement Learning for Dialogue Generation
• A Personalized System for Conversational Recommendations
• A Deep Reinforcement Learning Chatbot
• A Survey of Available Corpora for Building Data-Driven Dialogue Systems
• Survey on reinforcement learning for language processing
• Towards End-to-End Learning for Dialog State Tracking and Management using Deep Reinforcement Learning
• Generative Deep Neural Networks for Dialogue: A Short Review
• A Survey on Recent Advances and Challenges in Reinforcement Learning Methods for Task-Oriented Dialogue Policy Learning
• Few-Shot Generalization Across Dialogue Tasks
• User Simulation for Evaluating Information Access Systems
• Learning Robust Dialog Policies in Noisy Environments
• Multimodal Hierarchical Reinforcement Learning Policy for Task-Oriented Visual Dialog
• Neural Assistant: Joint Action Prediction, Response Generation, and Latent Knowledge Reasoning
• Investigation of Error Simulation Techniques for Learning Dialog Policies for Conversational Error Recovery
• Sentiment Adaptive End-to-End Dialog Systems
• Deep Reinforcement Learning with Pre-training for Time-efficient Training of Automatic Speech Recognition
• An empirical assessment of deep learning approaches to task-oriented dialog management
• I love your chain mail! Making knights smile in a fantasy game world: Open-domain goal-oriented dialogue agents
• Improving Mild Cognitive Impairment Prediction via Reinforcement Learning and Dialogue Simulation
• Pre-training in Deep Reinforcement Learning for Automatic Speech Recognition
• Augmenting Knowledge through Statistical, Goal-oriented Human-Robot Dialog
• Document-editing Assistants and Model-based Reinforcement Learning as a Path to Conversational AI
• Towards Personalized Dialog Policies for Conversational Skill Discovery
• Dueling RL: Reinforcement Learning with Trajectory Preferences
• Domain Transfer in Dialogue Systems without Turn-Level Supervision
• Learning Natural Language Generation from Scratch
• Data-Efficient Methods for Dialogue Systems
• Natural Language Generation as Planning under Uncertainty Using Reinforcement Learning
• Building A User-Centric and Content-Driven Socialbot