Search for demonstrators - HORIZON-CL4-2024-DATA-01-03 (Piloting emerging Smart IoT Platforms and decentralized intelligence)

Data is at the epicentre of the digital transformation, affecting all sectors of activity. Developments related to big data and AI have already spurred disruption over business domains, from platform economies to the roll-out of innovative and intelligent services. Centralized AI systems though, have several key issues that stem from their dependence on large data sets. One of the pivotal challenges of the next decade of AI is to determine whether data and intelligence will be democratized or will remain in control of a few large organisations. Centralized AI puts the power of smart technology in the hands of those with access to huge amounts of data and novel tools. This leaves SMEs and mid-caps behind, creating a digital divide that will gradually deepen. This can be hindered by adopting decentralized AI systems.

On the other hand, the energy demand of AI systems is a growing concern, especially with the increasing use of edge computing, deep learning and other computationally intensive algorithms. This is becoming increasingly evident with the wide adoption of large language models that use terabytes of data for training, thus a large amount of computing power and capital investment, which translates into significant carbon dioxide emissions. With the increased use of AI across several domains, from manufacturing to farming, to intelligent transport systems, Europe urgently needs to develop new data management solutions that will harness the transformative potential of AI, whilst meeting the European Green Deal objectives.

To address these challenges, MEDAILLON invests in Decentralized Artificial Intelligence (DAI). In a DAI system, decision-making processes are decentralized and based on consensus among multiple nodes instead of being controlled by a single central authority. This approach utilises Blockchain technology to store and process data and provides a more secure, transparent, and trustworthy alternative to traditional AI systems. The combination of the two technologies has enormous potential for novel data applications.

MEDAILLON takes a system-level approach across the cloud/edge/fog continuum to foster innovation, developing a meta-framework that governs hardware of smart devices to operating systems at device and at system level, to middleware and to application software of industrial mass-market applications.

Industrial mass-market applications are products, technologies, or systems that are produced on a large scale and are designed for widespread use across various industries. These applications are geared towards enhancing efficiency and productivity in industrial settings (see examples in PILOTS section).

The meta-framework will provide a comprehensive set of tools, libraries, and standardized approaches to develop, manage, and deploy applications rooted in DAI principles, accounting for algorithm’s energy-efficiency. It will act as a unifying layer that abstracts lower-level complexities, allowing developers to focus on building applications rather than dealing with the intricacies of hardware and communication protocols. It will offer standardised edge architecture and tools including:

• Device Abstraction and Compatibility: common interfaces for interacting with a wide range of devices and sensors, irrespective of their underlying hardware or communication protocols.
• Data Management and Processing tools for efficiently collecting, processing, and managing data generated by devices.
• Security and Authentication thus security features for ensuring secure communication, data encryption, access control, and device authentication within IoT networks.
• Interoperability and Standardization that promotes adherence to industry standards.
• Edge Intelligence and Analytics: Decentralized AI capabilities for analytics and decision-making at the edge.
• Application Lifecycle Management tools for deploying, monitoring, and managing IoT applications.
• Cross-Domain integration of IoT applications across different domains and industries, allowing for cross-functional solutions.
• Energy Optimization to effectively utilise resources, optimizing power consumption.
• Scalability and Flexibility to allow scaling with the growing number of devices and applications, adapting to changing requirements and environments.

The meta-framework will promote interoperability and will accelerate the deployment of industrial IoT/AI mass-market applications across diverse environments.