Neuromorphic control design using hybrid systems tools, with applications to event-based control and nuclear fusion challenges.
We develop machine learning methods to identify mechanistic model parameters from complex, stochastic data, and apply them to uncover principles of neural computation from single neurons to circuits
We have the pleasure of welcoming members of RTE and Gérédis, French electricity network operators. We will be welcoming Emeline Georges (Haulogy) and Olivier Bronckart (Elia).
From TSO to DSO: Discover how grid operators are reinventing voltage control to tackle massive RES integration and new power flow dynamics. The presentation will be held in french
EV-GNN is a graph-based solution for large-scale EV charging optimization. It improves RL scalability and efficiency using GNNs and branch pruning, and adapts to diverse EV problems and action spaces.
We show that deep learning can learn the analysis step of data assimilation for chaotic systems. A simple CNN matches optimally tuned EnKF accuracy, even with a single forecast state, by identifying unstable dynamical perturbations without ensembles.
We design a linear state estimation technique for nonlinear systems, based on the Koopman operator framework, which allows to recover a transient state with an arbitrarily fast rate of convergence.
This talk explores key trends in micro/nanoelectronics design, covering power optimization, variability, 3D integration, and new devices. It aims to motivate the audience to tackle the field's upcoming challenges and innovations.
Automated conversions between formulations are an essential feature of optimization software. This presentation surveys current challenges and advances, with examples from a new solver interface design.
Dr. Ioannis Boukas (ENGIE, Belgium) explores AI applications in short-term power trading, optimizing batteries and wind parks. Head of Algo Trading at ENGIE, he specializes in machine learning and optimization for energy markets, with a PhD from the University of Liège.
Prof. Jochen Cremer (TUDelft) explores Deep Learning for power system reliability, focusing on N-k security-constrained power flow and real-time security assessment. An expert in AI for power systems, he leads research at Delft AI Energy Lab and the Austrian Institute of Technology.
Dr. Balthazar Donon (RTE, France) explores AI for tertiary voltage control using graph neural networks to handle real-world power system complexities. A researcher at RTE R&D, he specializes in energy and deep learning, aiming to develop AI assistants for grid operators.
