Seminar series

We are proud to introduce the 2026 ACSOS seminar series, a series of talks by high-profile researchers who contribute to the field of autonomous and self-* systems.

Check the latest schedule for our seminar series:

Edge-to-cloud systems link devices and cloud computing for fast, efficient services. In this seminar, Prof. Alessandro Vittorio Papadopoulos will explain how they work, why predictability matters in time-sensitive contexts, and the key challenges and solutions in making such systems reliable.

Date & Time: Wednesday, May 27, 2026 · 5:30–7:00 PM (CET)

Event location: Sala Armi, Palazzo Malvezzi, via Zamboni 22 – In presence and online event

Attend online: Join on Microsoft Teams

Register in person: Book your seat

Event info: https://site.unibo.it/isa/en/events/on-the-edge-of-the-cloud-making-smart-systems-predictable-when-time-matters

As AI systems become more autonomous and increasingly embedded in everyday life, social intelligence is becoming just as critical as technical capability. Today’s AI systems often exhibit superhuman performance in narrow domains, yet lack key aspects of human social reasoning—creating challenges as they interact more deeply with people and institutions.

In this seminar, Dr. Peter Lewis will explore how autonomous systems can be designed to function effectively within hybrid human–AI societies. Using the case study of institution formation for the sustainable governance of shared resources, he will demonstrate how agent-based modeling and evolutionary game theory provide complementary tools for understanding and shaping collective behavior.

Date & Time: Thursday, April 2, 2026 · 10:00–11:00 AM (ET)/3:00-4:00 PM (GMT)

Please RVSP here: https://osu.az1.qualtrics.com/jfe/form/SV_8GEUyNNNX0wIRVk

Event info: https://tdai.osu.edu/events/tdai-speaker-series-reflective-social-intelligence-autonomous-agents

Engineered intelligence (EI) systems use non-biological components to reason about their best course of action in complex environments while weighing competing short- and long-term objectives and operating under resource constraints. With massive investment driving proliferation, EI assistance will soon become pervasive in everyday life. However, EI deployed in novel environments and operating contexts struggle to understand the nuanced abstractions their deployment context demands, making them unreliable, inefficient, and too costly.

In this talk, I argue for a new approach to computer systems research — one that prioritizes convergent, purpose-specific work aimed squarely at the emerging tasks not yet supported by EI systems. These "timely systems" reduce the cost of EI, making the technology more accessible and sustainable. Timely systems help EI (1) adapt general skills to domain-specific tasks, (2) navigate non-stationary, heterogeneous contexts for physical intelligence at the edge, and (3) build sustainable, vibrant ecosystems for all forms of intelligence. Rather than competing with EI for the same optimization problems, timely systems complement it: they deliver immediate, principled cost reductions and reliability gains while the EI core continues to improve its own models. This symbiotic approach accelerates the transition from expensive, experimental prototypes to affordable, ready-to-use services.

Date & Time: Friday, August 7, 2026 · 10:00–11:00 AM (ET)/2:00–3:00 PM (GMT)

Event details and RSVP can be found here here: https://forms.gle/PzcsahAefWU1sG379