WP3 – COSA Research & Development
The core R&D engine of COSA: infrastructure, knowledge translation and distributed data analysis.
Duration
M7–M32
Objectives
- R&D on a framework for assisted decision making.
- R&D of the platform for collaborative data mining.
- R&D of a framework for distributed data analysis.
Description of work
T3.1 – Agriculture 4.0 Infrastructure R&D (M7–M24)
Based on the requirements elicited in WP2, a novel infrastructure for Agriculture 4.0 is researched and designed. A hexagonal Edge–Fog–Cloud architecture is implemented to improve scalability, reliability and security of data flows across the agricultural ecosystem.
T3.2 – Data to Knowledge Translation (M13–M30)
Optimal knowledge preservation implies properly formalised, contextually framed knowledge. Novel ways to formalise and frame knowledge pieces are researched, together with software components that automate these traditionally manual tasks, in order to make the results of data mining reusable and trustworthy.
T3.3 – Distributed Data Analysis (M13–M30)
To support the sharing of knowledge while respecting privacy, distributed analysis of data is researched, in the form of Federated Learning methodologies. Local models are trained on Fog devices and federated within the Cloud layer; research focuses on raising the privacy and robustness of the resulting models.
Deliverables
- D3.1 — COSA Infrastructure (M24).
- D3.2 — Distributed Data Analysis and Translation to Knowledge (M30).