Research Projects Neural States Dynamics

Here is a potential project for future students. Two more projects in this research directions are ongoing.

Energy-based neural network models for capturing neural dynamics 

This project aims to model the critical, yet largely unknown, role of neuromodulators in regulating neural state transitions. At the heart of this exploration is a cutting-edge model of attractor networks, utilizing energy-based models to represent each neural state as a point in an energy landscape. Energy-based models are emerging as one of the most theoretically tractable methods for understanding dynamics in machine learning models. By leveraging advanced machine learning tools, this project aims to model these energy-based systems, offering a computational framework to decode how neural networks navigate the energy landscape and how neuromodulators orchestrate this navigation. The ultimate goal is to unlock insights into the brain’s remarkable adaptability and the mechanisms that enable it to respond to diverse conditions.