Efficient AI for Networked Energy Systems

Kurzfassung

What if you could train a powerful AI model in seconds instead of hours? Emerging unconventional computing methods like reservoir computing (RC) make this possible, and the energy sector is an ideal playground for it with plenty of possible applications for efficient AI. We start our presentation with a tour of AI research at the German Aerospace Center (DLR) and an accessible introduction to reservoir computing. We then dive into real applications at the DLR Institute of Networked Energy Systems: what the energy industry actually wants from AI, how RC performs on demand and electricity price forecasting, and why its tiny computational footprint makes it a hot candidate for edge AI and anomaly detection. We look forward to an open discussion on efficient AI and its role in future energy systems.