Research
I am working on mathematical system theory. A first level of iterative system specification is developed as a general modeling formalism. A second level of discrete-event system specification constitutes an operational formalism. A third level of abstract simulator algorithms is used to execute model trajectories. Finally, structure-based machine learning takes advantage from the correlation between simulator activity and model performances.
The systems studied are mostly: Complex, emergent, dynamic structure, learning, multi-agents,...
Tools and applications
Iterative System Specification (ITSYS) is the general modeling formalism. Discrete-event System Specification (DEVS) is the operational formalism. Pseudorandom and parallel abstract simulators execute model trajectories. Activity-based credit assignment (ACA machine learning) searches for the different solution compositions based on components activity. Graph-based and RAndom DiscrEte-event Simulator (GRADES) implements Parallel DEVS, graph based generation/visualization and ACA machine learning.
In neuronal networks, I am considering the modeling and simulation of the activity of biomimetic networks in a sensorimotor context. This work is done inside the university priority research Modélisation Théorique et Computationnelle en Neurosciences et Sciences Cognitives.