Agenda

Beyond Recognition: Multimodal AI for Reconstructing, Understanding, and Interacting with Complex Worlds

Speaker:
Elena Gavagnin, Zurich University of Applied Sciences (ZHAW)

Abstract:
Recent advances in multimodal and generative AI are opening new possibilities for moving beyond the recognition of individual objects, actions, or signals towards richer representations of complex environments. But what does it mean for an AI system to actually understand a visual or multimodal situation? And how can such representations support generation, reconstruction, and meaningful interaction with humans and their physical and digital environments? In this talk, I will present selected research projects from the Information Systems Engineering group at the Institute of Information Systems, Zurich University of Applied Sciences (ZHAW). The examples span semantic understanding of visual scenes, multimodal analysis of human behaviour and social interaction, generative modelling of scientific imagery, and multimodal frameworks for human-AI interaction, while also touching on emerging applications in complex real-world environments.Through these different perspectives, the talk will explore how AI systems can move from recognising patterns towards reconstructing, understanding, and interacting with the complex worlds those data represent, while highlighting common research questions and opportunities for future collaboration.

Bio sketch:
Elena Gavagnin is Co-Head of the Information Systems Engineering group and Senior Lecturer for Data Science at the Zurich University of Applied Sciences (ZHAW), where she is also an Associate Faculty Member of the Centre for Artificial Intelligence. Her research focuses on AI and machine learning, with particular interests in multimodal AI, computer vision, and human behaviour and interaction. Originally trained as an astrophysicist, she now leads interdisciplinary research projects spanning scientific imaging, digital games, human-centred AI, and intelligent interactive systems.