A Simple Language Model for Task-Oriented Dialogue

Abstract

Task-oriented dialogue is often decomposed into three tasks: understanding user input, deciding actions, and generating a response. While such decomposition might suggest a dedicated model for each sub-task, we find a simple, unified approach leads to state-of-the-art performance on the MultiWOZ dataset. SimpleTOD is a simple approach to task-oriented dialogue that uses a single, causal language model trained on all sub-tasks recast as a single sequence prediction problem. This allows SimpleTOD to fully leverage transfer learning from pre-trained, open domain, causal language models such as GPT-2. SimpleTOD improves over the prior state-of-the-art in joint goal accuracy for dialogue state tracking, and our analysis reveals robustness to noisy annotations in this setting. SimpleTOD also improves the main metrics used to evaluate action decisions and response generation in an end-to-end setting: inform rate by 8.1 points, success rate by 9.7 points, and combined score by 7.2 points.

arXiv comment22 Pages, 2 figures, 16 tables

References in corpus (20)

• HuggingFace's Transformers: State-of-the-art Natural Language Processing
• The Curious Case of Neural Text Degeneration
• SQuAD: 100,000+ Questions for Machine Comprehension of Text
• CTRL: A Conditional Transformer Language Model for Controllable Generation
• Layer Normalization
• Generating Long Sequences with Sparse Transformers
• Towards a Human-like Open-Domain Chatbot
• A Network-based End-to-End Trainable Task-oriented Dialogue System
• Semantically Conditioned LSTM-based Natural Language Generation for Spoken Dialogue Systems
• DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation
• Multi-domain Dialogue State Tracking as Dynamic Knowledge Graph Enhanced Question Answering
• Transferable Multi-Domain State Generator for Task-Oriented Dialogue Systems
• Global-to-local Memory Pointer Networks for Task-Oriented Dialogue
• Alternating Recurrent Dialog Model with Large-scale Pre-trained Language Models
• Task-Oriented Dialog Systems that Consider Multiple Appropriate Responses under the Same Context
• Semantically Conditioned Dialog Response Generation via Hierarchical Disentangled Self-Attention
• Rethinking Action Spaces for Reinforcement Learning in End-to-end Dialog Agents with Latent Variable Models
• Neural Assistant: Joint Action Prediction, Response Generation, and Latent Knowledge Reasoning
• Paraphrase Augmented Task-Oriented Dialog Generation
• MOSS: End-to-End Dialog System Framework with Modular Supervision

Cited by in corpus (51)

• Recent Advances in Natural Language Processing via Large Pre-Trained Language Models: A Survey
• SOLOIST: Building Task Bots at Scale with Transfer Learning and Machine Teaching
• DialoGLUE: A Natural Language Understanding Benchmark for Task-Oriented Dialogue
• CoCo: Controllable Counterfactuals for Evaluating Dialogue State Trackers
• Structured Prediction as Translation between Augmented Natural Languages
• BiToD: A Bilingual Multi-Domain Dataset For Task-Oriented Dialogue Modeling
• Task-Oriented Dialogue System as Natural Language Generation
• UBAR: Towards Fully End-to-End Task-Oriented Dialog Systems with GPT-2
• Modelling Hierarchical Structure between Dialogue Policy and Natural Language Generator with Option Framework for Task-oriented Dialogue System
• A Survey on Dialog Management: Recent Advances and Challenges
• Jointly Optimizing State Operation Prediction and Value Generation for Dialogue State Tracking
• Robust Conversational AI with Grounded Text Generation
• Domain State Tracking for a Simplified Dialogue System
• MinTL: Minimalist Transfer Learning for Task-Oriented Dialogue Systems
• Modeling Text-visual Mutual Dependency for Multi-modal Dialog Generation
• Personalizing Task-oriented Dialog Systems via Zero-shot Generalizable Reward Function
• Truthful AI: Developing and governing AI that does not lie
• SUMBT+LaRL: Effective Multi-domain End-to-end Neural Task-oriented Dialog System
• RADDLE: An Evaluation Benchmark and Analysis Platform for Robust Task-oriented Dialog Systems
• Action-Based Conversations Dataset: A Corpus for Building More In-Depth Task-Oriented Dialogue Systems
• A Sequence-to-Sequence Approach to Dialogue State Tracking
• MultiWOZ 2.4: A Multi-Domain Task-Oriented Dialogue Dataset with Essential Annotation Corrections to Improve State Tracking Evaluation
• Zero-shot Generalization in Dialog State Tracking through Generative Question Answering
• Multi-Task Learning for Situated Multi-Domain End-to-End Dialogue Systems
• Teaching Models new APIs: Domain-Agnostic Simulators for Task Oriented Dialogue
• Joint System-Wise Optimization for Pipeline Goal-Oriented Dialog System
• A Probabilistic End-To-End Task-Oriented Dialog Model with Latent Belief States towards Semi-Supervised Learning
• Robustness through Data Augmentation Loss Consistency
• Improving Longer-range Dialogue State Tracking
• Leveraging Slot Descriptions for Zero-Shot Cross-Domain Dialogue State Tracking
• DORA: Toward Policy Optimization for Task-oriented Dialogue System with Efficient Context
• Slot Self-Attentive Dialogue State Tracking
• Continual Learning in Task-Oriented Dialogue Systems
• Resource Constrained Dialog Policy Learning via Differentiable Inductive Logic Programming
• SYNERGY: Building Task Bots at Scale Using Symbolic Knowledge and Machine Teaching
• Zero-Shot Dialogue State Tracking via Cross-Task Transfer
• Transferable Dialogue Systems and User Simulators
• Variational Latent-State GPT for Semi-Supervised Task-Oriented Dialog Systems
• Towards a Universal NLG for Dialogue Systems and Simulators with Future Bridging
• DLGNet-Task: An End-to-end Neural Network Framework for Modeling Multi-turn Multi-domain Task-Oriented Dialogue
• Oh My Mistake!: Toward Realistic Dialogue State Tracking including Turnback Utterances
• Reinforced Natural Language Interfaces via Entropy Decomposition
• MEDCOD: A Medically-Accurate, Emotive, Diverse, and Controllable Dialog System
• An Asynchronous Updating Reinforcement Learning Framework for Task-oriented Dialog System
• Doc2Dict: Information Extraction as Text Generation
• SOCCER: An Information-Sparse Discourse State Tracking Collection in the Sports Commentary Domain
• Learn to Focus: Hierarchical Dynamic Copy Network for Dialogue State Tracking
• Self-training Improves Pre-training for Few-shot Learning in Task-oriented Dialog Systems
• Dialogue State Tracking with a Language Model using Schema-Driven Prompting
• NUANCED: Natural Utterance Annotation for Nuanced Conversation with Estimated Distributions
• SIMMC 2.0: A Task-oriented Dialog Dataset for Immersive Multimodal Conversations