Objective O1: Design and implement the HYPER-AI open architecture to let swarms of nodes collaboratively build collective intelligence at the edge using decentralized AI/ML.
HOW: Deliver a scalable, self-contained distributed mesh of autonomous abstractions (“Smart-Nodes”). Enable dynamic deployment across the cloud-edge continuum using real-time DLT-based findability/access, and multi-objective optimization of resources (compute, storage, connectivity cost, energy) against diverse performance indices.
Objective O2: Enable dynamic, cognitively informed, optimized decisions across the full application lifecycle (design, execution, maintenance) via continuous inference.
HOW: Let registered entities (devices, nodes, Smart-Nodes, swarms) interpret context and internal state to optimize QoS/QoE. Support continuous incremental learning and a live library of reusable analytic functions for future transfer learning. Provide state estimation for (a) self-awareness during operation/orchestration and (b) fast response, offloading bandwidth-heavy tasks to edge resources while keeping compute-intensive work near cloud. Keep resource use transparent to apps. For critical events, reusable AI enables early detection, recognition, warning, and rapid reaction.
Objective O3: Develop autonomic entities for autonomous Smart-Nodes and rapid coordination of Smart-Node swarms during design and runtime.
HOW: Provide optimization/management tools across network hierarchies and ontology instantiations. Create self-organizing Smart-Nodes that pool available resources from real-time, data-driven situational awareness. Enable dynamic swarming into Meta-Smart-Nodes via decentralized coordination that adapts to emergent behavior and supports collaborative cross-node schemes with shared, otherwise idle meta-resources.
Objective O4: Ensure a secure, private, and robust edge-to-cloud ecosystem.
HOW: Deliver a distributed security & privacy framework using end-to-end encryption, key management, and AI-based intrusion detection. Protect data in use and at rest, preserve privacy for individuals/communities, detect attacks and errors, and enforce consistent, persistent policies via distributed ledger technologies.
Objective O5: Safeguard openness to guarantee interoperability, accessibility, and international cooperation.
HOW: Establish a framework for social innovation with active co-creation and community engagement to maximize adoption. Through Eclipse Foundation Europe, adopt open-science practices and build channels to open-source communities. Run five pilot use cases (Manufacturing, Automotive, Agriculture, Energy, Healthcare) in five countries (one led by a Korean partner). Involve innovative SMEs aligned with consortium skills. Provide open-source snippets and core functions to accelerate external development. Align with EU Open Research principles to stimulate a vibrant, international HYPER-AI community and uptake.
Objective O6: Demonstrate HYPER-AI’s usability, performance, energy efficiency, interoperability, and impact.
HOW: Assess and showcase tools, modules, and the back-end optimization stack across five use cases at multiple pilot sites and scales. Monitor all scenarios under a rigorous evaluation framework. Focus on verifying interoperability in challenging operational environments (Manufacturing, Agriculture, Automotive, Energy, Healthcare).
Objective O7: Drive wide communication, efficient dissemination, and vertical exploitation of results.
HOW: Propose novel, disruptive business models arising from HYPER-AI, analyze market uptake, and offer revised “business-as-usual” models for diverse stakeholders (cloud vendors, policy-makers, telecoms/ISPs, developers, end-users) across the continuum. Explore incentivized, secure sharing of edge resources (including data) to grow Europe’s data economy, support EU single-market data spaces, and foster a trustworthy AI ecosystem. Evaluate Robot-as-a-Service (RaaS)—subscription-based access to robots with updates/maintenance within hybrid cloud-edge settings—and cloud robotics as potential models for the hybrid adaptive and cognitive computing continuum.
