Abstract Autonomous systems are increasingly pervasive, capable, and complex—but behind every such system lie one or more feedback loops that shape how it senses, decides, and acts. Control theory has guided feedback design for decades, yet modern systems now expose the limits of its model-based foundations. The models it relies on are frequently inaccessible, inaccurate, or too rigid to fully exploit today’s data, storage, and computational resources. This talk explores a modeling approach grounded in behavioral systems theory, where systems are defined not by equations but by the set of trajectories they exhibit. The structure—or “shape”—of this set determines not only its mathematical properties but, crucially, also its tractability, governing the analytical and algorithmic tools suitable for analysis and design. I will show how this framework offers a principled bridge between parametric and non-parametric representations, enabling scalable, adaptive, and certifiable control in real-world settings—from fault detection in building automation systems, to certification in algorithmic pipelines, to large-scale control in urban traffic networks. I will conclude by outlining emerging directions and open challenges, including nonlinear and distributed dynamics, uncertainty quantification, and architectural considerations. Co-sponsored by: nagamune@mech.ubc.ca Speaker(s): Alberto , Agenda: Agenda: Gathering: 3:45pm – 4:00pm Talk followed by Q&A: 4pm – 5:30pm Room: 1203, Bldg: CEME , UBC Point Grey Campus, 6250 Applied Science Lane, Vancouver , British Columbia, Canada, Virtual: https://events.vtools.ieee.org/m/479975