The Human Interaction Lab investigates the computational foundations of human social interaction, focusing on both human-to-human communication and interactions between humans and artificial agents, such as Large Language Models (LLMs). While spoken and written language remain our primary focus, we also study other communicative modalities, including gesture, intonation, and prosody.
Our methodological approach is rigorously eclectic, prioritizing the most effective tools to address our research questions. We employ experimental methods, qualitative techniques such as Conversation Analysis, quantitative corpus analysis, computational modeling, and other approaches as needed to deepen our understanding of social interaction.
Recent research highlights include examining the strengths and limitations of Large Language Models as conversational agents. Additionally, we develop datasets and algorithms aimed at enhancing social communication between humans and machines.
Recent Publications
2024
Muhammad Umair, Vasanth Sarathy, and Jan Ruiter. 2024. Large Language Models Know What To Say But Not When To Speak. In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 15503–15514, Miami, Florida, USA. Association for Computational Linguistics.
Umair M, Mertens JB, Warnke L, de Ruiter JP. Can Language Models Trained on Written Monologue Learn to Predict Spoken Dialogue? Cogn Sci. 2024 Nov;48(11):e70013. doi: 10.1111/cogs.70013. PMID: 39587969.
2023