Very briefly, from 1996, I have been interested in, involved, and
leading, some works in virtual laboratories. At the beginning, this
work began because I was confident that the computer was providing a
new way to study complex systems dynamics, in particular natural
systems. Computers are providing a new tool to perform virtual
experiments to help experimental scientists (e.g., biologists,
psychologists, and in many other fields...) in their experimental work
to study their models, and test hypothesis that can not be tested in
the open world.
This work has been largely based on collaborations with two fields of science, oceanologists, and experimental psychologists.
Some key features where: how do we model? how do we simulate? how do we discuss the relevance of the simulation with regards to the real world? how do we interpret the result of simulations and obtain meaningful conclusions for the real world? which tools should omputer scientist design and implement? are there technological locks to break to be able to go further in this kind of modeling, and simulations? ...
The work with oceanologists has begun with Yvan Lagadeuc, then with the University of Lille 1, and Christophe Cambier, then at the University du Littoral Côte d'Opale in Calais (like myself), around 1997. The work went on with Éric Ramat when Christophe left the university (and eventually moved to Paris 6), and some students. Basically, the goal was to model and simulate the behavior of the copepod in its environment and try to fuel the debate on whether the copepod has an active behavior (towards food, towards mates, ...), or the copepod is merely the subject of flows and eats, reproduces itself, ... by chance.
I am no longer involved in this work which continues at the University du Littoral Côte d'Opale under the heading of Éric Ramat.
The work with experimental psychologists had begun around 1995, until around 2005. The goal was to investigate models of the dynamics of living beings (human beings in particular) and simulate them. After some wanderings, our attention was drawn towards reinforcement learning.