Research in the Computational Intelligence Laboratory addresses different aspects of networks in the human brain and in artificial systems.
Our lab developed an interest in Computational Neuroscience in 2015. At that time, we initiated a collaboration with the Max Planck Institute in Jupiter, FL, working on the computational analysis of two-photon imaging of calcium dynamics in the cortical astrocytes of ferrets. The objective of this work is to identify and classify flow characteristics in both the processes and the soma. To this end, we are investigating variational autoencoders. We are also developing models of astrocyte interactions using multi-compartmental models of processes and soma with a view of disentangling the complex interactions visible in the experiments. We are currently working with functional models to explore quantitative behavior.
We have begun the study of artificial networks, concentrating on recurrent and unsupervised networks. One research project involves the development of an attention model to better understand eye motion intended to find a hidden object in an image. Neural networks are also being considered as a driver in robotic systems. Finally, we are investigating the development of procedural methods for game development linked to the data-driven model of a player, that guides the generation of game components.