SPEAKER: Na Du, Assistant Professor, Department of Informatics and Networked Systems, University of Pittsburgh
 

Abstract: As robots increasingly enter homes, workplaces, and assistive settings, they must understand how people naturally communicate and interact in order to collaborate effectively with diverse users. This talk presents a program of research on multimodal and social human-robot interaction that investigates how humans and robots establish shared understanding and adapt their communication to one another. First, I describe our work on multimodal human-to-robot communication through speech, gaze, gesture, and embodied behavior. We show that people often rely on nonverbal cues, particularly gaze, to clarify underspecified verbal instructions. Building on these findings, we develop a gaze-aware multimodal instruction-understanding model to improve target-object grounding and intent understanding. Second, I present a set of studies on robot-to-human communication. These studies examine how robots can establish shared context, align task goals with diverse users, and adapt communication strategies to users' states and interaction contexts. I conclude with implications and open directions for building human-aware robots to support natural, effective, and inclusive human-robot collaboration.

Bio: Na Du is an Assistant Professor in the Department of Informatics and Networked Systems, School of Computing and Information at the University of Pittsburgh. Her research interests include human factors in intelligent systems, human-agent teaming, and interaction design, with applications spanning robotics, transportation, healthcare, and other intelligent systems. She is a recipient of several prestigious awards, including the NSF CAREER Award, HFE Women Rising Star Award, and Best Paper Awards from HRI and HFES international conferences. She earned her Ph.D. in Industrial & Operations Engineering from the University of Michigan in 2021.  

---

IPaT: GVU Lunch Lecture Series