Psych-E: Configurable Response Generation using Personality Traits and Pragmatics

Anonymous

Abstract: Personality traits influence human actions and thoughts, which is manifested in day to day conversations. Although glimpses of personality traits are observable in existing open domain conversation corpora, leveraging generic language modelling for response generation overlooks the interlocutor idiosyncrasies, resulting in non-customizable personality agnostic responses. With the motivation of enabling configurable response generators, in this paper we experiment with ways to ground neural response generators based on both (i) interlocutor Big-5 personality traits, and (ii) discourse intent as control codes, training an end-to-end dialogue agent that can not only leverage the control codes as policy for nuanced response generation, but also predict and decide the generation policy to be utilized by the generator. Since most of the existing large scale open domain chat corpora do not include Big-5 personality traits and discourse intent, we employ automatic annotation schemes to enrich the corpora with policy consisting of noisy estimates of these features as control codes, and leverage automatic evaluation metrics along with ablation studies, to assess the impact of using control codes for response generation. Additionally, we leverage human judgement to demonstrate the effectiveness of using such personality and pragmatics based policy for response generation. Our experiments illustrate the effectiveness of this strategy resulting in improvements to existing benchmarks.