Gianluca Ceruti

Dr. Gianluca Ceruti is a researcher affiliated with the Center for Intelligent Systems (CIS) at École Polytechnique Fédérale de Lausanne (EPFL). His work focuses on computational efficiency in neural networks through dynamical low-rank approximation methods. He has collaborated with institutions including the Karlsruhe Institute of Technology (KIT), Gran Sasso Science Institute (GSSI), and University of Innsbruck.

Research Interests:

• Low-rank optimization for neural networks
• Reducing computational/memory footprint of AI models
• Numerical integration techniques for machine learning

His 2023 presentation at EPFL CIS-RIKEN AIP series introduced the DLRT algorithm, which adaptively modifies neural network weight matrix ranks during training to preserve accuracy while lowering resource requirements. This aligns with CIS research pillars like AI for medicine and decentralized edge AI infrastructure.