Exploiting Persona Information for Diverse Generation of Conversational Responses

Abstract

In human conversations, due to their personalities in mind, people can easily carry out and maintain the conversations. Giving conversational context with persona information to a chatbot, how to exploit the information to generate diverse and sustainable conversations is still a non-trivial task. Previous work on persona-based conversational models successfully make use of predefined persona information and have shown great promise in delivering more realistic responses. And they all learn with the assumption that given a source input, there is only one target response. However, in human conversations, there are massive appropriate responses to a given input message. In this paper, we propose a memory-augmented architecture to exploit persona information from context and incorporate a conditional variational autoencoder model together to generate diverse and sustainable conversations. We evaluate the proposed model on a benchmark persona-chat dataset. Both automatic and human evaluations show that our model can deliver more diverse and more engaging personabased responses than baseline approaches.

You Impress Me: Dialogue Generation via Mutual Persona Perception

“…Furthermore, researchers explored to generate diverse responses conditioned on persona (Song et al, 2019(Song et al, , 2020. Personalization in goal-oriented di-alogue systems has also received some attention (Joshi et al, 2017;Luo et al, 2019).…”

“…chit-chat) systems have made great progress towards mimicking human-like responses. Nevertheless, there still exist some serious challenges in building personalized chatbots that can deliver engaging conversations and gain user trust (Song et al, 2019). For example, current chit-chat systems tend to generate uninformative responses (Li et al, 2016b).…”

Guiding Variational Response Generator to Exploit Persona

“…Persona-based neural conversation models can be categorized into two major research directions. One is to directly train a model from conversational data by considering the persona information (Li et al, 2016b;Kottur et al, 2017;Madotto et al, 2019), while the other approach makes use of the profiles or sideinformation of users to generate the aligned responses (Chu et al, 2018;Qian et al, 2018;Mazare et al, 2018;Song et al, 2019). The work described in this paper belongs to the first research direction.…”

“…Thus, we have introduced the architecture proposed by Zhao et al (2017) and modified it to adapt to the persona-aware generation, for the meaningful comparison. Especially, Song et al (2019) have utilized persona information into the CVAE architecture, except they focus on modeling and copying users' explicit profiles.…”

A Multi-Persona Chatbot for Hotline Counselor Training

“…There has recently been a surge of interest in generating coherent and consistent dialogues grounded on pre-defined persona profile information from the PersonaChat dataset (Zhang et al, 2018;. Approaches to enforce consistent personas on this dataset have included retrieving relevant profile facts (Zhang et al, 2018), retrieving and refining relevant utterances , increasing the probability of copying a word from the profile (Yavuz et al, 2019), tuning to discourage inconsistent responses (Li et al, 2019a), reranking candidate responses (Welleck et al, 2019), and combining natural language inference with reinforcement learning (Song et al, 2019). Unfortunately, these methods fall short of generating responses that are as grammatical, diverse, engaging, and descriptive as natural human generated conversation (See et al, 2019;Roller et al, 2020).…”