Abstract

The goal-oriented dialogue system needs to be optimized for tracking the dialogue flow and carrying out an effective conversation under various situations to meet the user goal. The traditional approach to build such a dialogue system is to take a pipelined modular architecture, where its modules are optimized individually. However, such an optimization scheme does not necessarily yield the overall performance improvement of the whole system. On the other hand, end-to-end dialogue systems with monolithic neural architecture are often trained only with input-output utterances, without taking into account the entire annotations available in the corpus. This scheme makes it difficult for goal-oriented dialogues where the system needs to integrate with external systems or to provide interpretable information about why the system generated a particular response. In this paper, we present an end-to-end neural architecture for dialogue systems that addresses both challenges above. In the human evaluation, our dialogue system achieved the success rate of 68.32%, the language understanding score of 4.149, and the response appropriateness score of 4.287, which ranked the system at the top position in the end-to-end multi-domain dialogue system task in the 8th dialogue systems technology challenge (DSTC8).

Highlights

  • The goal-oriented dialogue system helps users achieve their goals such as requesting information or executing commands via natural language conversations

  • The traditional approach to building a goaloriented dialogue system mostly adopts a pipelined modular architecture, with the natural language understanding (NLU) module (Kim et al, 2017; Lee et al, 2019b) that first recognizes and comprehends user’s intent and extracts values for slots, the dialogue state tracking (DST) module (Williams et al, 2013) that tracks the values of slots, the dialogue policy (POL) module that decides the system action, and the natural language generation (NLG) module (Wen et al, 2015) that generates the utterance that corresponds to the system action

  • We show that our model is competitive to other state-of-the-art models specialized for two sub-tasks in the dialogue management, i.e. Dialogue State Tracking and Dialogue-Context-to-Text Generation tasks, our model was not tuned for those sub-tasks

Read more

Summary

Introduction

The goal-oriented dialogue system helps users achieve their goals such as requesting information or executing commands via natural language conversations. Multiple modules are combined together, e.g. the Word-level DST (Ramadan et al, 2018; Wu et al, 2019; Lee et al, 2019a) which maps the dialogue history to the dialogue state (the composite function of NLU and DST), and the Word-level POL (Budzianowski et al, 2018; Pei et al, 2019; Chen et al, 2019; Mehri et al, 2019; Zhao et al, 2019) which maps the previous utterance and dialogue state to the system response (the composite function of POL and NLG) These modules are usually optimized separately, which does not necessarily lead to an overall optimized performance for successful task completion.

The MultiWOZ Dataset
ConvLab
End-to-End Neural Pipeline for Goal-Oriented Dialogue System
Input Representation
Delexicalization
Training Objective
Decoding Strategy
Handling Empty Query Result
Related Work
Training Details
Evaluation Metrics
Automatic Evaluation
Human Evaluation
Attention Weights
MultiWOZ Benchmarks Performance
Dialogue State Tracking
Dialogue-Context-to-Text Generation
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call