Thanos: Enhancing Conversational Agents with Skill-of-Mind-Infused Large Language Model
Young-Jun Lee, Dokyong Lee, Junyoung Youn, Kyeongjin Oh, Ho-Jin Choi
2024-11-08
Summary
This paper discusses Thanos, a new conversational agent that enhances communication by using a special dataset to teach the model how to respond appropriately in different social situations.
What's the problem?
Conversational agents, like chatbots, often struggle to respond in a way that feels natural or appropriate during conversations. This is because they lack the ability to choose the right conversational skills based on the context, which makes their responses less engaging and relevant.
What's the solution?
To solve this problem, the researchers created a dataset called Multifaceted Skill-of-Mind, which includes around 100,000 conversations that showcase various conversational skills needed for different scenarios (like counseling or task-oriented discussions). They then developed Thanos, a series of large language models (LLMs) that use this dataset to learn how to respond better in conversations. Thanos can handle different types of interactions and demonstrates improved understanding and response quality compared to previous models.
Why it matters?
This research is important because it helps make conversational agents more effective and relatable. By teaching these models how to use appropriate conversational skills, Thanos can improve user experiences in applications such as customer service, mental health support, and social interactions, ultimately leading to better communication between humans and machines.
Abstract
To increase social bonding with interlocutors, humans naturally acquire the ability to respond appropriately in a given situation by considering which conversational skill is most suitable for the response - a process we call skill-of-mind. For large language model (LLM)-based conversational agents, planning appropriate conversational skills, as humans do, is challenging due to the complexity of social dialogue, especially in interactive scenarios. To address this, we propose a skill-of-mind-annotated conversation dataset, named Multifaceted Skill-of-Mind, which includes multi-turn and multifaceted conversational skills across various interactive scenarios (e.g., long-term, counseling, task-oriented), grounded in diverse social contexts (e.g., demographics, persona, rules of thumb). This dataset consists of roughly 100K conversations. Using this dataset, we introduce a new family of skill-of-mind-infused LLMs, named Thanos, with model sizes of 1B, 3B, and 8B parameters. With extensive experiments, these models successfully demonstrate the skill-of-mind process and exhibit strong generalizability in inferring multifaceted skills across a variety of domains. Moreover, we show that Thanos significantly enhances the quality of responses generated by LLM-based conversational agents and promotes prosocial behavior in human evaluations.