This repository provides code for simulating emergent collective motion from groups of continuous-time and -space active inference agents. This code has been adapted and originates from the paper "Collective behavior from surprise minimization" (2024) by Conor Heins, Beren Millidge, Lancelot Da Costa, Richard Mann, Karl Friston, and Iain Couzin.
All coding modifications were created for an undergraduate research experience supported by Cal-Bridge and the University of California Riverside Mentoring Summer Research Internship Program (MSRIP). The program ran from June - August 2026 with a final symposium presentation on August 15 [See poster].
The original codebase contained both a JAX and a Julia implementation of a multi-agent active inference algorithm for generating collective motion. **All coding modifications for this project have been applied to the JAX implementation, focusing on demo_nolearning (see the official instructions).
- JAX backend installation/run instructions: jax_backend/README_JAX.md