Zoom: https://illinois.zoom.us/j/88521857435?pwd=LkBMPM4uQkONOAc4J2cg2RLSOTxa6P.1
Abstract:
Multimodal machine perception has transformed the ability of today's AI systems to understand and interact with the world. However, the integration of various sensory modalities presents significant challenges, including efficient adaptation to downstream tasks, robust cross-modal alignment, and fine-grained visual reasoning. This talk will cover my lab’s research in contrastive self-supervision, cross-modal retrieval, and scene-graph representations, aimed at improving multimodal robustness and accuracy. We will conclude with a discussion on open research directions in human-agent collaboration and embodied intelligence.
Bio:
Ismini Lourentzou is an Assistant Professor in the School of Information Sciences at the University of Illinois Urbana-Champaign, where she leads the Perception and LANguage (PLAN) Lab. Her research focus is multimodal machine learning, primarily the intersection of vision and language in settings with limited supervision, and its applications in healthcare, embodied AI, and other fields. She has served on the NeurIPS'23 and NeurIPS'22 organizing committees, as a co-organizer for the NeurIPS'22 Gaze Meets ML workshop, and in editorial and area chair roles for PLOS Digital Health, ECCV, ACL, IEEE BigData, and MICCAI. Her research has been supported by NSF, DARPA, Commonwealth Cyber Initiative (CCI), and Amazon. Lourentzou completed her PhD in Computer Science at the University of Illinois Urbana-Champaign. Before joining iSchool@Illinois, she was an Assistant Professor of Computer Science at Virginia Tech, where she received a 2023 VT College of Engineering Dean’s Award for Excellence as an Outstanding New Assistant Professor.
Part of the Illinois Computer Science Speakers Series. Faculty Host: Jiawei Han
Meeting ID: 825 1375 0293
Passcode: csillinois
If accommodation is required, please email <erink@illinois.edu> or <communications@cs.illinois.edu>. Someone from our staff will contact you to discuss your specific needs
