Attend LIONS seminar with speaker Ted Pavlic, Sept. 18
Join the Learning, Information, Optimization, Networks and Statistics, or LIONS, group for a seminar with Ted Pavlic, associate professor in the School of Computing and Augmented Intelligence, part of the Ira A. Fulton Schools of Engineering with a joint appointment in the School of Life Sciences. He was also the founding associate director for research at ASU’s Biomimicry Center. His interdisciplinary Bringing Ecology and Engineering Together laboratory, or BEET Lab, includes students from computer science, industrial engineering, animal behavior, biology and applied mathematics for the life and social sciences. Research in the lab ranges from empirical studies of social-insect colonies in the field and laboratory to artificial intelligence, multi-robot systems and ethical and moral dimensions of the design of cyber-physical systems.
Many living systems have adaptations to solve problems familiar to control engineers: Maintaining safety constraints despite disturbances, regulating important quantities around desired operating points and sensing global system states using only local information. Social-insect colonies are particularly interesting examples because these functions emerge from the interactions of many individuals without centralized control. Not only can engineering tools and perspectives help to better understand these natural systems, but discovery of novel information-processing and control patterns in these systems can enhance bio-inspired design of artificially intelligent collectives.
In this survey of some of the work on biological and bio-inspired collectives, Pavlic describes several examples in which social insects rapidly reorganize their collective behavior to regulate performance quantities or maintain safety invariants in uncertain environments. One example involves army ants that form temporary behavioral barriers between neighboring traffic flows, effectively enforcing a safety constraint against aggressive encounters. These barriers can be reproduced by directly implementing the observed behavioral rules, but strikingly similar structures also emerge in multi-agent reinforcement learning when agents are given higher-level objectives that make no explicit reference to walling. He then highlights examples of decentralized regulation in social insects and how local interactions can plausibly close the regulatory feedback loop. Finally, Pavlik describe a physically embodied form of multi-agent quorum sensing in which a collective detects a group-level state threshold even though no individual can directly measure the relevant collective state.
Together, these examples illustrate how ideas from feedback control, safety, distributed sensing and multi-agent learning can help explain collective animal behavior. In the other direction, the natural history of social insects provides a rich catalog of decentralized mechanisms that may suggest new approaches to sensing, control and coordination in engineered multi-agent systems.
Visit the LIONS Seminar website for more information.
LIONS Seminar: Safety, Setpoints, and Sensing on Six Legs: Decision and Control in Social Insects
Friday, Sept. 18, 2026
1:30–2:30 p.m.
Goldwater Center (GWC) 487, Tempe campus [map]
Attend online via Zoom