Partner Personas Generation for Diverse Dialogue Generation

TL;DR

This paper introduces a novel framework that automatically generates partner personas to improve dialogue response diversity and relevance, utilizing reinforcement learning and a critic network, with demonstrated superior performance over baselines.

Contribution

The paper presents a new method for automatic partner persona generation in dialogue systems, enhancing response quality without relying on ground truth partner personas.

Findings

Generated partner personas are relevant, informative, and coherent.

Partner personas improve response informativeness and engagement.

The critic network effectively reinforces the framework.

Abstract

Incorporating personas information allows diverse and engaging responses in dialogue response generation. Unfortunately, prior works have primarily focused on self personas and have overlooked the value of partner personas. Moreover, in practical applications, the availability of ground truth partner personas is often not the case. This paper attempts to tackle these issues by offering a novel framework that leverages automatic partner personas generation to enhance the succeeding dialogue generation. We incorporate reinforcement learning with a dedicatedly designed critic network for reward judgement. Experimental results from both automatic and human evaluation demonstrate a) Our framework is capable of generating relevant, informative and coherent partner personas, even compared to the ground truth partner personas. b) Generated partner personas enhance the succeeding response generation, thus surpassing our baselines and comparison model when partner personas are missing during the inference stage. c) Our framework generates responses that are more informative and engaging than our baseline conditioned on the ground truth partner personas during inference. d) Our dedicatedly designed critic network reinforces our framework effectively. Finally, our framework gives better explainability and reduces the demands for external databases for partner personas.