Luiz Chamon, Junior Research Group Leader at ELLIS/SimTech, is giving an invited tutorial on "Constrained supervised and reinforcement learning with constraints" at the Learning for Dynamics and Control (L4DC) conference in Oxford, UK. The tutorial is part of a lineup designed to explore innovative approaches to machine learning with a focus on dynamics and control. This year, the L4DC conference takes place from July 15-17, gathering leading researchers and practitioners in the field.
Chamon, along with his collaborators Miguel Calvo-Fullana (Universitat Pompeu Fabra, Spain), Santiago Paternain (Rensselaer Polytechnic Institute, USA), and Alejandro Ribeiro (University of Pennsylvania, USA), cover the recent theoretical advancements in constrained learning, offering practical advice and discussing a range of example applications. This tutorial is geared towards researchers and professionals interested in machine learning systems' robustness, safety, and fairness.
A unique aspect of this tutorial is its focus on constrained learning, where requirements are incorporated as statistical constraints rather than through traditional modification of training objectives. This approach has the potential to streamline computational processes and reduce the need for extensive hyperparameter tuning, a challenge in conventional methods.
If you couldn't make it to Vancouver to catch this tutorial presentation at the Conference on Artificial Intelligence (AAAI-24) in Feburary 2024, Oxford presents a great opportunity to engage with this innovative content. A version of this tutorial will also be presented at the 32nd European Signal Processing Conference (EUSIPCO), the flagship conference of the European Association for Signal Processing (EURASIP), in Lyon at the end of August.