Scientific Colloquium – Andreagiovanni Reina – The power of inhibition for collective decision making in minimalistic robot swarms

SPEAKER: Andreagiovanni Reina,

Centre for the Advanced Study of Collective Behaviour,

University of Konstanz and Max Planck Institute of Animal Behavior,

Konstanz, Germany

ABSTRACT: I investigate how large groups of simple robots can reach a consensus with decentralised minimalistic algorithms. Simple robots can be useful in nanorobotics and in scenarios with low-cost requirements. I show that through decentralised voting algorithms, swarms of minimalistic robots can make best-of-n decisions. In my research, I show that using a biologically-inspired voting model based on inhibitory signals, the swarm can collectively perform better and be more resilient against a minority of misbehaving robots than in models without inhibition. Animal behaviour is the source of inspiration for robot algorithms; in turn, the resulting large-scale robotics experiments and mathematical models can help better understand the natural systems and the evolutionary advantages of certain signals. Collective behaviour can be counterintuitive, and as Anderson said in 1972, More Is Different. I will also show some counterintuitive results where the stereotypical view that noise is something to suppress is broken; on the contrary, there are relevant conditions where noise and individual errors can be beneficial components to the group behaviour, improving group accuracy and responsiveness to environmental changes.

BIO: Andreagiovanni Reina is a Research Group Leader of the Centre for the Advanced Study of Collective Behaviour of the University of Konstanz and Max Planck Institute of Animal Behavior, Germany. His research is highly interdisciplinary in both its scope and its methodology, with numerous contributions to a variety of disciplines, including computer science, robotics, theoretical biology, physics, cognitive neuroscience and psychology. Andreagiovanni's interdisciplinary approach consists of combining techniques from dynamical systems theory, statistical physics, network science, statistical optimality theory, multiagent simulation and large-scale robotics. From 2021 to 2023, Andreagiovanni was a Research Fellow in Collective Behaviour at the Interdisciplinary Institute for Artificial Intelligence (IRIDIA) of the Université Libre de Bruxelles, funded by the Belgian F.R.S.-FNRS as a Chargé de Recherches. Previous to that, from 2015 to 2020, Andreagiovanni was a Research Fellow at the University of Sheffield (UK). He holds a PhD in Applied Sciences from IRIDIA, Université Libre de Bruxelles, and an MSc in Computer Engineering from Politecnico di Milano, Italy. He has been a researcher in several European projects on distributed robotic systems since 2009.

DATE/TIME: Thursday 14 November 2024, at 2.30pm CET.

The colloquium will also be broadcast online, on Zoom.