Alignment in Spoken Interaction

We study how large language models can interpret and adapt to the way people coordinate understanding through speech, including pauses, disfluencies, timing, and repair. Drawing on the view of language as joint action, this work examines how conversational alignment shapes meaning and how AI systems can model these dynamics. The goal is to develop speech technologies that interact more like humans do—responsive, adaptive, and grounded in real communicative behavior. (read more here: click here )