 |
Mohammed Ali JALLAL
Postdoctoral researcher in artificial intelligence and the energy sector, CEA, France |
Mohammed Ali JALLAL is a dynamic researcher with a Ph.D. in Artificial Intelligence, Renewables Energies, and Electrical Engineering from the Faculty of Sciences Semlalia, Cadi Ayyad University, Marrakesh, Morocco (2021). His academic journey includes a research internship at the Faculty of Computer Science, University of Murcia, Spain, where he focused on consumption energy forecasting in smart buildings using hybrid machine-learning algorithms. Currently a postdoctoral researcher at the French Alternative Energies and Atomic Energy Commission (CEA), Mohammed Ali is deeply involved in cutting-edge research at the intersection of artificial intelligence and the energy sector. His expertise extends to uncertainty quantification and machine learning applications in multi-vector energy systems, as demonstrated through his participation in the TRILOGY project. Mohammed Ali is also contributing his expertise in digitalization and energy systems to drive impactful research outcomes within the International Energy Agency (IEA) District Heating and Cooling (DHC) project. With a passion for innovation, Mohammed Ali’s research interests span artificial intelligence, renewable energies, and electrical engineering, focusing on developing hybrid algorithms to solve complex real-world problems across various domains. He has a proven track record of collaboration with industrial partners and mentoring Ph.D. students, underscoring his commitment to advancing research and nurturing future talent.
Title: Trusted AI for Energy Systems: Navigating Uncertainty with Causal, Incremental, and Embedded Tiny Machine Learning
In this keynote, we will explore the transformative role of artificial intelligence (AI) in energy systems, with a particular focus on developing trusted AI solutions that are not only effective but also reliable and transparent. As AI increasingly powers critical energy infrastructures, managing uncertainty in AI predictions becomes essential for ensuring the robustness and safety of these systems. I will present innovative techniques for uncertainty quantification, providing practical solutions to reduce uncertainty and enhance the confidence in AI-driven decisions.
We will delve into feature engineering, highlighting how careful selection and transformation of data features can improve model accuracy and interpretability. Additionally, the integration of causal machine learning will be discussed as a method to uncover underlying cause-effect relationships, moving beyond correlation-based approaches to support more robust decision-making in energy management.
As AI systems deployed in energy environments must adapt over time, I will discuss incremental learning, which enables models to evolve and improve as new data becomes available, ensuring long-term adaptability. The rise of embedded AI and tiny machine learning algorithms is another key focus, where we will examine how these lightweight, efficient algorithms can be deployed in resource-constrained environments, such as sensors and edge devices within energy systems, to enable real-time decision-making with minimal computational power.
Through real-world applications in energy systems, including energy demand forecasting, fault detection, and optimization of renewable energy generation, I will demonstrate how these AI advancements are being applied to address critical challenges. Supported by research insights, this keynote will illustrate how trusted, embedded AI is not only improving the efficiency, reliability, and resilience of energy technologies but also opening up new possibilities for innovation in the field.