Bidirectional Understanding In Human-AI Interaction
Abstract: Intelligent agents hold much promise for improving people’s lives, from driving assistance systems to home robots for older adults. Such assistive interactions inherently involve bidirectional communication, in which humans and agents both try to interpret and respond to their partner’s actions. As intelligent systems become more complex, their decision making becomes simultaneously more useful and harder to understand. In this talk, I will present several examples from my lab’s research on how people interpret the behavior of AI agents, and how agents can interpret the behavior of people. I’ll describe how we can use cognitive science concepts like theory of mind to improve the communication and explainability of AI systems, making them more useful for assistive and collaborative tasks with people.
Bio: Dr. Henny Admoni is an Associate Professor in the Robotics Institute at Carnegie Mellon University, where she leads the Human And Robot Partners (HARP) Lab. Dr. Admoni’s research interests include human-robot interaction, assistive robotics, and nonverbal communication. Dr. Admoni holds a PhD in Computer Science from Yale University, and a BA/MA joint degree in Computer Science from Wesleyan University.