IMMERSE Seminar: Ismini Lourentzou - Generative Models for Interactive Worlds

This event is hybrid, in person at 2405 Siebel Center for Computer Sceicne and on Zoom via: https://go.illinois.edu/immerseseminarlourentzou

Abstract:

Generative models offer a foundation for virtual worlds in which humans and agents can act, explore, and learn. Building these worlds requires representations that preserve meaningful structure, predictions that capture the consequences of actions, and mechanisms for adapting to feedback. In this talk, I will present recent work from the PLAN Lab on generative models for physical interaction. I will discuss their potential as interactive environments for learning, planning, and evaluation, and the open challenges in bringing these capabilities to embodied agents and immersive experiences.

Speaker Bio: 

Ismini Lourentzou is an Assistant Professor in the School of Information Sciences at the University of Illinois Urbana–Champaign, with affiliations in the Siebel School of Computing and Data Science, Electrical and Computer Engineering, IMMERSE, and NCSA. She leads the Perception and LANguage (PLAN) Lab with research focusing on physical intelligence and embodied AI. Her work aims to develop intelligent multimodal systems that can perceive, reason, anticipate, and act in the physical world. Before joining Illinois, she was a faculty member at Virginia Tech and a research scientist at IBM Research. She earned her Ph.D. in Computer Science from the University of Illinois Urbana–Champaign.

Refreshments will be provided.