Multi-Agent Behavior:
Representation, Modeling, Measurement, and Applications
CVPR 2022



The behavior of intelligent agents is often shaped by their interactions with other agents in the environment. In multi-agent scenarios, complex interactions occur between agents, such that the behavior of an individual cannot be understood in isolation. Models of multi-agent behavior are of interest to researchers and scientists across disciplines spanning from ecology and neuroscience to sports analytics, video games, and autonomous vehicles. It is difficult to understand behavior in any of these contexts without understanding how agents sense or interact with other agents.

The purpose of this workshop is to provide a forum for exchanging perspectives on defining, interpreting, measuring, and modeling the behavior of multiple agents given predominantly visual data. A panel of speakers from a variety of disciplines will present their work, and discuss the key goals of multi-agent behavior research as it applies to their own field. By identifying common challenges and themes across fields, we aim to foster new cross-disciplinary approaches to the modeling and analysis of multi-agent behavior.


  • Computational ethology

  • Generative modeling

  • Multi-agent reinforcement learning

  • Safe/social multi-agent modeling

  • Timescale analysis: ways to study behavior at different timescales

  • Different ways to define behavior: discrete and continuous

  • Supervised, semi-supervised, or unsupervised methods to quantify behavior