Retrieval-Enhanced Adversarial Training for Neural Response Generation

TL;DR

This paper introduces Retrieval-Enhanced Adversarial Training (REAT), a novel method that combines retrieval-based and generative dialogue models using adversarial training to improve neural response generation.

Contribution

The paper proposes a new REAT framework that integrates retrieval-based candidates into adversarial training for neural response generation, enhancing dialogue system performance.

Findings

REAT significantly outperforms vanilla Seq2Seq models.

REAT surpasses conventional adversarial training approaches.

Empirical results on large benchmark datasets validate effectiveness.

Abstract

Dialogue systems are usually built on either generation-based or retrieval-based approaches, yet they do not benefit from the advantages of different models. In this paper, we propose a Retrieval-Enhanced Adversarial Training (REAT) method for neural response generation. Distinct from existing approaches, the REAT method leverages an encoder-decoder framework in terms of an adversarial training paradigm, while taking advantage of N-best response candidates from a retrieval-based system to construct the discriminator. An empirical study on a large scale public available benchmark dataset shows that the REAT method significantly outperforms the vanilla Seq2Seq model as well as the conventional adversarial training approach.