Reservoir Computing in the Wild:
From Energy-Efficient Digital Twins to Foundational Dynamical Substrates
Discover how Artificial Intelligence can model the physical world with unprecedented efficiency in an upcoming CTTC seminar by Dr. Matteo Mendula.
- Speaker: Dr. Matteo Mendula
- Title: Reservoir Computing in the Wild: From Energy-Efficient Digital Twins to Foundational Dynamical Substrates
- Date & Time: Wednesday, February 18, 2026 | 10:00 AM CET
- Location: CTTC Auditorium (Building B6)
- Available online
- Duration: Approximately 45 minutes and Q&A
About the Seminar
This seminar examines how Reservoir Computing (RC) can become a sustainable alternative to energy-intensive deep learning approaches in real-world applications. Dr. Mendula will discuss methods that drastically reduce optimization time and energy consumption while improving accuracy in Digital Twin ecosystems.
He will also introduce the concept of a “Universal Dynamical Substrate”—a structural foundation that enables AI systems to generalize across dynamic environments with minimal data and computational cost. The talk will highlight experimental results in transfer learning, online adaptation, and energy-aware computation for industrial cyber-physical systems.
Dr. Matteo Mendula is a researcher at the Sustainable Artificial Intelligence Research Unit at the Centre Tecnològic de Telecomunicacions de Catalunya (CTTC). His research focuses on bridging the gap between the computational efficiency of Reservoir Computing (RC) and the performance demands of state-of-the-art Deep Learning. He specializes in the optimization of offline-online training cycles and the application of lightweight AI in Hierarchical Digital Twin ecosystems. His recent work explores the use of Lyapunov-guided reservoirs as foundational models for extracting universal dynamical laws in resource-constrained environments.