IFLEX (Ikerlan Federated Learning EXtensible kit)
Ikerlan Federated Learning Extensible KIT provides a solution designed to collaboratively improve AI models across multiple participants in a secure and privacy-preserving manner. Service providers use the KIT to publish a specific asset containing configuration files that deploy federated learning
Purpose
The Ikerlan Federated Learning Extensible KIT enhances DSSC's objectives by enabling secure, privacy-preserving, and interoperable collaborative AI model training. It facilitates information sharing without exposing sensitive data, supports data sovereignty through controlled client-provider interactions, and utilizes interoperable standards like EDC connectors and gRPC protocol, ensuring secure federated learning workflows within data spaces.
About
Ikerlan Federated Learning Extensible KIT provides a solution designed to collaboratively improve AI models across multiple participants in a secure and privacy-preserving manner. Service providers use the KIT to publish a specific asset containing configuration files that deploy federated learning client components, which are automatically integrated with consumer’s EDC connector, enabling authorized participants to securely access federated learning service. Clients download these components, which establish secure gRPC-based data plane connecting clients to the provider's aggregation services. This allows participants to train models locally and request aggregated model updates on-demand.
Building blocks
Blueprint compliance
Details
- Technology readiness level
- TRL 5
- Domain / sector / industry
- Industrial Manufacturing, Automotive, Industry 4.0, Edge-Cloud Hybrid Environments
- Geographical application area
- Global, with a primary focus on Europe due to alignment with European data sovereignty standards and data spaces frameworks (e.g., Eclipse Tractus-X, Catena-X).
- Deployments in operation
- (Currently not applicable due to TRL below 7)
- Dependencies
- Eclipse Data Space Connector (EDC) -Flower AI Federated Learning Framework -Docker/Kubernetes infrastructure for component deploymentg -RPC protocol support for communication
- Business offerings
- (Currently not applicable; under development and research phase)
- Additional functionalities
- This tool integrates Eclipse Data Space Connector (EDC) with Flower AI federated learning framework, alongside custom-developed APIs for client-server interactions, aggregation management, and secure AI model exchange, enhancing collaborative AI capabilities across industrial data spaces.
- Similar/competing implementations
- OpenFL (Intel) -NVIDIA Clara Federated -LearningFedML
Relevant documents & links
Community rating
Ranked by people who have used it, not by self-assessment.
Not Yet Rated. This tool needs 5 ratings before it gets a score. Until then it sits mid-catalogue rather than at the bottom, so a lack of ratings is never mistaken for a poor one.
Reviews
No reviews yet — be the first to share your experience.