Data-driven decision-making with machine learning (ML) algorithms is changing the way society and the economy work, has a big positive impact on our daily life. To be effective, these solutions often need to process data close to the end user, and such data can be private and sensitive.
Distributed learning, especially federated learning (FL), has emerged as a dominant paradigm within the ML branch that meets these two properties. Federated learning has grown in tandem with the extension of the cloud to the edge (CloudEdge), but interestingly both have evolved almost independently despite their natural parallelism. In MLEDGE (Cloud and Edge Machine Learning), IMDEA Networks collaborated with Dr. Nikolaos Lautalis as Principal Investigator to implement FL as an independent but optimized cross-sector layer on the cloud his edge and the real world work with us to reverse this trend. Applications and data to demonstrate that this synergy can greatly benefit everyone.
The data economy is estimated to impact 27 EU countries at €827 billion by 2025. (1)*. So the goals are: Enabling a thriving ecosystem of secure and efficient FL services at the edgecan facilitate training machine learning models using highly sensitive personal and B2B data (collaborating under various trust assumptions from full to zero and any level in between). individual end-users or administratively independent organizations).
Efficiency, Sustainability and Security
As highlighted in Elisa CabanaIMDEA Networks Postdoctoral Fellow: “This project will contribute to research specifically on federated learning as a service (FLaaS), cloud edge processing, efficient use of FL in hybrid clouds, and protection from attacks., protecting sensitive or sensitive data exchanged, managing data portability challenges at the edge, etc. In this context, the team has developed a development framework and components that will help popularize this type of service, as well as poisoning and inference attacks initiated from unruly edge servers and “honest but curious” aggregation nodes. design a solution for This includes creating “watermarks” to prevent redistribution of data or metadata exchanged between edge servers under FLaaS.
Other highlights, as Cabana summarizes: “Creating an economic and business logic layer that distributes costs and revenues fairly among stakeholders when collaborating on training ML models, supporting DevOps (a set of practices that integrate software development and IT) “The goal is to shorten the software development lifecycle, provide high-quality continuous delivery, and provide continuous development of machine learning services in the cloud to monitor, predict, and It is about optimizing costs by allocating them intelligently and energy-efficiently.” This research will be used to process sensitive data of individuals and in areas of traditional and digital economy such as fintech, identity, health, transportation and access control. Contribute further to the design, implementation, and creation of public demonstrators that provide useful models.
Transfer of technology to society
The innovations in this project will develop favorable market conditions for the use of integrated learning in cloud and integrated data architectures (such as those defined by institutions such as IDSA and Gaia-X) in an international context. It is of great interest in addressing important economic, business and social issues related to the existence of silos in data storage and utilization in the economy. “MLEDGE Makes Advanced Federated Learning Technology Accessible to More Organizations and IndividualsSustainable business for all actors in the value chain (machine learning experts/suppliers, cloud and data service providers, traditional and digital industries, public sector and academia, etc.), including SMEs and government agencies. Promote creation. said Nikolaos Laoutaris, Research Professor at IMDEA Networks and Principal Investigator of the Institute’s project.
This project lays the foundation for the development of cloud and ML/FL infrastructure in Spain and the promotion of national R&D&I. Contribute to the 2030 Sustainable Development Goals set by the United Nations. It also promotes the sustainable development of efficient networks and FL solutions through practical work that can have a real and positive impact on the environment. From a technical solution perspective, these include:
- traditional economics (construction, finance, health, etc.). This use case is developed by an enterprise to improve processes and decision-making (e.g. real-time) based on data and models from Florida.
- digital economy. One example is in the area of digital health, such as leveraging information from mobile devices and wearable technology. Another is training a digital advertising model.
- Optimizing your CloudEdge infrastructure. A key feature of MLEDGE, it uses federated-trained machine learning algorithms.
MLEDGE (Cloud and Edge Machine Learning) is funded by the Ministry of Economy and Digital Transformation of the European Union NextGeneration-EU (January 2023 to June 2025).
*(1) European Data Strategy 2025-2030.
Disclaimer: AAAS and EurekAlert! We are not responsible for the accuracy of news releases posted on EurekAlert!. Use of information by contributing institutions or via the EurekAlert system.
