Call for Customized Conversation: Customized Conversation Grounding Persona and Knowledge

Abstract

Humans usually have conversations by making use of prior knowledge about a topic and background information of the people whom they are talking to. However, existing conversational agents and datasets do not consider such comprehensive information, and thus they have a limitation in generating the utterances where the knowledge and persona are fused properly. To address this issue, we introduce a call For Customized conversation (FoCus) dataset where the customized answers are built with the user's persona and Wikipedia knowledge. To evaluate the abilities to make informative and customized utterances of pre-trained language models, we utilize BART and GPT-2 as well as transformer-based models. We assess their generation abilities with automatic scores and conduct human evaluations for qualitative results. We examine whether the model reflects adequate persona and knowledge with our proposed two sub-tasks, persona grounding (PG) and knowledge grounding (KG). Moreover, we show that the utterances of our data are constructed with the proper knowledge and persona through grounding quality assessment.

You Truly Understand What I Need: Intellectual and Friendly Dialogue Agents grounding Knowledge and Persona

“…In addition, the incoherent topics of the dialogues lead to shallow levels of conversation between the interlocutors. To elaborate on this chit-chat conversation supported by external knowledge, Jang et al (2022) presents a novel persona-knowledge chat with a generative model that considers persona information and world knowledge altogether. Despite obtaining the knowledge and persona when generating the answers, the generative models' responses still exhibit both hallucination and lesser engagingness as in Table 1.…”

“…Dataset FoCus (Jang et al, 2022) is the dataset for customized dialogue benchmark, where each conversation is directly grounded with knowledge and persona. The dataset includes knowledgeaware dialogue with personal profiles between humans and machines.…”

“…We implement the baselines from previous study (Jang et al, 2022) and we conduct experiments with GPT-2 (Radford et al, 2019) and BART (Lewis et al, 2020a) as well. For a fair comparison, we demonstrate the results on GPT-2 small , which has 12 layers, and BART base , which has 6 encoders and 6 decoder layers.…”

Beyond Candidates : Adaptive Dialogue Agent Utilizing Persona and Knowledge

“…It seems that Jang et al (2022a) and Lim et al (2022) adhere to this human-like approach on the conversation by referring to persona and knowledge. However, it neglects the humans' semantic concept reconstruction and retrieval capability by requiring pre-defined candidate sets to ground as in Figure 1 (b).…”

“…For the pre-training process, we first build the persona pool with the collections of unique persona sentences from FoCus (Jang et al, 2022a) and PersonaChat (Zhang et al, 2018). Then, we pre-train the persona retriever using DPR (Karpukhin et al, 2020) and regard highly ranked persona sentences from BM25 (Robertson and Zaragoza, 2009)…”