Main features: Pipeline with data and model versioning as well and automatic training and management
Objectives: Aligned with the AIMLFW specification of O-RAN, realized with state of the art cloud-negative technologies
Advantages: Zero-touch like model training and management.
What is new: Such system for cellular networks
Why is it important: Applicable for the core network and also in different verticals:IoT, vehicular communications, smart cities, etc.
Addressing target audiences and expressing needs
- To raise awareness and possibly influence policy
The platform targets research teams, ML engineers, and organizations seeking robust, interoperable, and production-ready MLOps solutions.
- EU and Member State Policy-makers
- Research and Technology Organisations
- Academia/ Universities
R&D, Technology and Innovation aspects
- Result refinement
- Partnerships and Collaborations
- Continuous monitoring and improvement
Replicability is ensured through a peer-reviewed publication in the Journal of Network and Computer Applications, titled “An overview and solution for democratizing AI workflows at the network edge” which fully documents the system design and methodology, and a maintained open GitHub repository containing the complete implementation, configuration files, and deployment instructions. A demonstration video further supports correct setup and execution, making the system straightforward to reproduce and extend.
