Posts

Cortical networks, made up of interconnected neurons, are key to brain function, with connectivity between fully structured and random. While various models of cortical connectivity have been proposed since the 1980s, a unified framework is missing. This article highlights a common feature: low-rank connectivity matrices. It links neural dynamics to connectivity structure, showing how low-rank designs can support specific computations, illustrated through four tasks.

This project, tackles the challenge of predicting tidal surges, proposed by the FLUMINANCE team at Inria. Using past data and pressure fields, the goal is to improve responses to extreme tidal events. My approach to addressing this challenge involves detailed data analysis followed by the implementation of an Encoder-Decoder model.