Abstract : The ANR-JCJC MIMICO (2024-2027) project, led by Dr. Nicolas Verstaevel, enhances agent-based modeling (ABM) using machine learning. It introduces surrogate modeling and active learning to reduce computational costs in model calibration and scenario exploration. By allowing the model to autonomously select the most informative data, MIMICO improves efficiency with minimal training. The project delivers an interactive ABM exploration framework, an open-source surrogate model, and advances in explainable AI. Applications in urban mobility and crowd simulation will demonstrate its real-world impact.

Please consult the program and registration page here : https://www.irit.fr/mimico/invitation-kick-off/.