UPCAST Plugins
The high-level objective of UPCAST is to design and deploy a set of universal plugins for data sharing, monetisation and trading platforms that enable actors in common European data spaces to collaboratively negotiate, improve and enforce data sharing contracts automatically, providing dynamic fair pricing mechanisms while implementing energy-efficient data exchange, ensuring privacy, confidentiality and legislation compliance and adhering to ethical and responsibility guidelines.
Privacy Plugin
Responsible partner: University of Southampton
For a Data Provider, it is essential that the privacy and usage constraints they wish to apply to their data are respected across the Data Sharing journey, as automatically as possible. The UPCAST Privacy Plugin helps towards that end by providing the following:
-
-
- A Graphical User Interface to enable Data Providers the definition of data usage and privacy policies in ODRL, removing the complexity of manually writing the policies
- An engine to detect and highlight conflicts between ODRL usage and data policies. This is useful to either to compare updates in policies, or to compare requests made by potential Data Consumers
- A Consent Manager to enable Data Providers to pre-define ODRL policies and collect consent from data subjects in a user-friendly way
-
Discovery Plugin
Responsible partner: University of Southampton
Enables the publishing and indexing of dataset metadata in a catalog powered by the CKAN engine, enabling the Discovery of datasets through:
-
-
- Keyword and faceted search
- Semantic Similarity of descriptions
- Finding replacement datasets for those used in a given workflow
-
Negotiation and Contracting Plugin
Responsible partner: University of Southampton
Allows users to create and negotiate data sharing contracts. UPCAST contracts extend the usage control specification defined by International-Data-Spaces-Association (IDSA).
The Negotiation process builds upon IDSA’s contract negotiation protocol to support the exchange of offers and counter-offers of contracts describing the specification of the dataset, price and usage policies up to the generation and signature of a data sharing contract.
Environmental Impact Optimiser Plugin
Responsible partner: CeADAR Ireland
This plugin aims to reduce carfbon emissions by optimising data processing workflows. It estimates energy costs for data storage and processing, identifies and explains the factors affecting energy consumption using explainable AI techniques (XAI), and helps comply with EU regulation. It also provides real-time monitoring of energy usage during data processing.
Pricing Plugin
Responsible partner(s): LSTech & CeADAR Ireland
This plugin aids data providers in setting fair prices for their data products and helps data consumers estimate costs for data acquisition. It can also be used to assess the pricing of customised data products, making it a valuable tool for informed decision-making in data trading platforms. Using explainable AI techniques (XAI), the plugin also provides transparency by identifying the key factors influencing the price.
Federate Machine Learning (FML) Plugin
Responsible partner: Nokia
This plugin aims to enable data providers to participate in machine learning projects without the need to share their data. It provides different entities for the server(orchestrator) and client side to enable this functionality.
This plugin would provide the following:
-
-
- Server side application that orchestrate the whole FML process and aggregate the locally trained models.
- Client agent that would engage with the server and enable the local training
- A seamless flow together with APIs that enable it to work and being integrated with different applications.
-
This project has received funding from the European Union´s Horizon Research and Innovation Actions under Grant Agreement nº 101093216.