Abstract:A task-oriented dialogue system (TOD) is an important application of artificial intelligence. In the past few years, works on multi-domain TODs have attracted increased research attention and have seen much progress. A main challenge of such dialogue systems is finding ways to deal with cross-domain slot sharing and dialogue act temporal planning. However, existing studies seldom consider the models' reasoning ability over the dialogue history; moreover, existing methods overlook the structure information of the ontology schema, which makes them inadequate for handling multi-domain TODs. In this paper, we present a multi-task learning framework equipped with graph attention networks (GATs) to probe the above two challenges. In the method, we explore a dialogue state GAT consisting of a dialogue context subgraph and an ontology schema subgraph to alleviate the cross-domain slot sharing issue. We further construct a GAT-enhanced memory network using the updated nodes in the ontology subgraph to filter out the irrelevant nodes to acquire the needed dialogue states. For dialogue act temporal planning, a similar GAT and corresponding memory network are proposed to obtain fine-grained dialogue act representation. Moreover, we design an entity detection task to improve the capability of soft gate, which determines whether the generated tokens are from the vocabulary or knowledge base. In the training phase, four training tasks are combined and optimized simultaneously to facilitate the response generation process. The experimental results for automatic and human evaluations show that the proposed model achieves superior results compared to the state-of-the-art models on the MultiWOZ 2.0 and MultiWOZ 2.1 datasets.

End-to-End Task-Oriented Dialog Modeling with Semi-Structured Knowledge Management

HyKnow: End-to-End Task-Oriented Dialog Modeling with Hybrid Knowledge Management

Read Key Points: Dialogue-Grounded Knowledge Points Generation with Multi-Level Salience-Aware Mixture

SKANet - Structured Knowledge-Aware Network for Visual Dialog.

Knowledge-Retrieval Task-Oriented Dialog Systems with Semi-Supervision

"What do others think?": Task-Oriented Conversational Modeling with Subjective Knowledge

Dual Semantic Knowledge Composed Multimodal Dialog Systems

Is MultiWOZ a Solved Task? An Interactive TOD Evaluation Framework with User Simulator

Towards Complex Scenarios: Building End-to-End Task-Oriented Dialogue System across Multiple Knowledge Bases

Training Zero-Shot Generalizable End-to-End Task-Oriented Dialog System Without Turn-level Dialog Annotations

Many Hands Make Light Work: Task-Oriented Dialogue System with Module-Based Mixture-of-Experts

HierTOD: A Task-Oriented Dialogue System Driven by Hierarchical Goals

Using Textual Interface to Align External Knowledge for End-to-End Task-Oriented Dialogue Systems

Simulating Task-Oriented Dialogues with State Transition Graphs and Large Language Models

Task-Optimized Adapters for an End-to-End Task-Oriented Dialogue System

Can Current Task-oriented Dialogue Models Automate Real-world Scenarios in the Wild?

InstructTODS: Large Language Models for End-to-End Task-Oriented Dialogue Systems

Multi-Task Pre-Training for Plug-and-Play Task-Oriented Dialogue System

Dialog History Management for End-to-end Task-Oriented Dialog Systems

Multi-task learning with graph attention networks for multi-domain task-oriented dialogue systems

Task-Oriented Dialog Systems that Consider Multiple Appropriate Responses under the Same Context