We model the behavior of edge-cloud nodes by monitoring computational resources. We conduct research to define methodologies for predictive maintenance, artificial intelligence-based operations management (AIOps), and service offloading.
We build edge-first systems to mitigate privacy and latency issues. We conduct research to define methodologies for compression (i.e., quantization, pruning), distribution (federated learning), and deployment (over-the-air protocols) of artificial intelligence models in low-power, low-computational-capacity devices.
TEMA is a four-year project funded by the European Union. It is part of the Horizon Europe program, Research and Innovation Action, under the call HORIZON-CL4-2022-DATA-01-01.
The goal is to provide an intelligent cloud-edge technical solution that supports disaster response and management by providing situational data to relevant end users (e.g., first responders, firefighters), enabling transferability to address different types of disasters in various geographic regions, and thus providing relevant information that can help make the best possible operational decisions.
Grant Agreement: 101093003
NEUROKIT2E is a three-year project funded by the European Union. It is part of the Horizon Europe program, under the HORIZON-KDT-JU-2022-2-RIA call for proposals.
The goal is to propose a deep learning platform for hardware integrated around a consolidated European value chain (providing artificial intelligence hardware and software). The project involves edge intelligence methodologies for training and inference, such as knowledge distillation, federated learning, pruning, and quantization.
Grant Agreement: 101112268
Organization of international conferences:
