Periodic Reporting for period 2 - TRUSTS (Trusted Secure Data Sharing Space)
Période du rapport: 2021-07-01 au 2022-12-31
TRUSTS contributed to the European Commission’s Digital Single Market Strategy by securing trust in the concept of data markets. This was achieved by building a platform that enables all key players in the Data Services Ecosystem to exchange and process data in a secure way while making the platform accessible for existing and emerging data markets through various interoperability solutions. Hence, the platform possesses the necessary technical capabilities to operate autonomously and serve as a platform aggregator. Through integration of state of the art privacy-preserving technologies, TRUSTS maintained a strong focus on the secure sharing of data. During the project, an in-depth analysis and framework on legal and ethical aspects, covering the entire data value chain, from data providers to consumers was developed. The development of the platform and the legal framework was guided by the practical experiences of 3 use cases (UCs) performed with companies and corporate business data in the financial and operator industries. The development of the platform drew on practical experiences of 3 UCs in order to co-create a viable, compliant and impactful governance, legal and business model for the TRUSTS platform.
WP1 continued the coordination of joint tasks across WPs. This included quality assurance, risk monitoring, regular organisation of project management meetings to ensure alignment between WPs as well as regular monitoring of work performed and resources used to ensure that high-quality results were delivered in time and budget.
WP2 finalised the analysis of the worldwide data marketplace ecosystem in terms of status, markets, trends, success and failure stories. Furthermore, a list with updated Functional Requirements was finalised. A detailed analysis of TRUSTS platform’s requirements as well as the final UC definitions were generated. Additionally, the trials evaluation testing methodology and the respective business evaluation methodology were finalised. Another achievement was the finalisation of the TRUSTS platform architecture.
WP3 technically implemented the innovative TRUSTS platform taking into account the requirements identified in WP2. To achieve this, the TRUSTS platform was evolved in an agile development and by leveraging existing data market technologies. WP3 focused on continuously refining the TRUSTS platform by incorporating feedback from WP5 and using the semantic layer to support the metadata ingestion and the interoperability with external data markets/EOSC. A brokerage and recommendation service was provided. The sell/buy actions were recorded in the smart contracts blockchain.
WP4 focused on
1) Cryptographically secure techniques that enabled us to perform analysis on privacy-sensitive data;
2) Privacy Preserving Transfer Learning and Classification, providing a novel encrypted transfer learning method and competitive results in terms of runtime and performance;
3) Performing de-anonymization attacks to the provided datasets, in order to find out the extent to which they complied with privacy regulations and inquired about the necessary anonymization measures;
4) Federated Deep Learning methodologies and new algorithms, providing an innovative encrypted federated learning prototype;
5) Strived to convert risky algorithms that compromise privacy into safe and privacy-preserving algorithms without harming their functionality.
WP5 defined and set up the test environment for the 3 UCs for each demonstration cycle while performing the relevant planning and pilot operational management. WP5 focused on conducting advanced field trials within the sectors of UC1: Financial Institutions, UC2: Telecom Operators, UC3: Corporate data providers, etc. while using the test results and data, delivered impact analysis and impact assessment reports and systematically addressed the pilots’ stakeholder perspectives for further recommendations and suggestions.
WP6 provided a legally and ethically sound background to the project by focusing on:
1) Laying down and monitoring compliance with research ethical principles to be followed by all TRUSTS partners;
2) Analysing the legal frameworks at EU (and, when applicable, national) level affecting the delivery of the TRUSTS platform and its use after the project.
3) Carrying out a legal and ethical assessment of the TRUSTS project focusing on its technology, underlying architecture and data workflows.
4) Providing legal and policy recommendations for stakeholders and policy makers based on the open legal issues and lessons learned.
WP7 conducted market research to identify existing business models for data markets, specifically those that combined scientific and non-scientific founders, in line with TRUSTS' mixed private and public ownership structure. WP7 also produced deliverables related to ecosystem needs, innovation aspects, intellectual property, data management, and best practices for successful business models. Notably, the WP developed an unified business model taxonomy for data marketplaces and selected a business model for TRUSTS as a data market and future federator of data marketplaces. It created standardisation recommendations that were well-received and developed further in high-profile W3C & EBDVF workshops.
WP8 continued producing and publishing regular online content (via Website, Social Media, YouTube, Podcast Tool) to inform stakeholders at a wider European and international level about the project objectives, processes and results.
In terms of WP9, all partners continuously monitored potential legal and ethical risks in close collaboration with KUL as the central contact point for ethics related questions.
WP2 finalised the analysis of the worldwide data marketplace ecosystem in terms of status, markets, trends, success and failure stories. Furthermore, a list with updated Functional Requirements was finalised. A detailed analysis of TRUSTS platform’s requirements as well as the final UC definitions were generated. Additionally, the trials evaluation testing methodology and the respective business evaluation methodology were finalised. Another achievement was the finalisation of the TRUSTS platform architecture.
WP3 technically implemented the innovative TRUSTS platform taking into account the requirements identified in WP2. To achieve this, the TRUSTS platform was evolved in an agile development and by leveraging existing data market technologies. WP3 focused on continuously refining the TRUSTS platform by incorporating feedback from WP5 and using the semantic layer to support the metadata ingestion and the interoperability with external data markets/EOSC. A brokerage and recommendation service was provided. The sell/buy actions were recorded in the smart contracts blockchain.
WP4 focused on
1) Cryptographically secure techniques that enabled us to perform analysis on privacy-sensitive data;
2) Privacy Preserving Transfer Learning and Classification, providing a novel encrypted transfer learning method and competitive results in terms of runtime and performance;
3) Performing de-anonymization attacks to the provided datasets, in order to find out the extent to which they complied with privacy regulations and inquired about the necessary anonymization measures;
4) Federated Deep Learning methodologies and new algorithms, providing an innovative encrypted federated learning prototype;
5) Strived to convert risky algorithms that compromise privacy into safe and privacy-preserving algorithms without harming their functionality.
WP5 defined and set up the test environment for the 3 UCs for each demonstration cycle while performing the relevant planning and pilot operational management. WP5 focused on conducting advanced field trials within the sectors of UC1: Financial Institutions, UC2: Telecom Operators, UC3: Corporate data providers, etc. while using the test results and data, delivered impact analysis and impact assessment reports and systematically addressed the pilots’ stakeholder perspectives for further recommendations and suggestions.
WP6 provided a legally and ethically sound background to the project by focusing on:
1) Laying down and monitoring compliance with research ethical principles to be followed by all TRUSTS partners;
2) Analysing the legal frameworks at EU (and, when applicable, national) level affecting the delivery of the TRUSTS platform and its use after the project.
3) Carrying out a legal and ethical assessment of the TRUSTS project focusing on its technology, underlying architecture and data workflows.
4) Providing legal and policy recommendations for stakeholders and policy makers based on the open legal issues and lessons learned.
WP7 conducted market research to identify existing business models for data markets, specifically those that combined scientific and non-scientific founders, in line with TRUSTS' mixed private and public ownership structure. WP7 also produced deliverables related to ecosystem needs, innovation aspects, intellectual property, data management, and best practices for successful business models. Notably, the WP developed an unified business model taxonomy for data marketplaces and selected a business model for TRUSTS as a data market and future federator of data marketplaces. It created standardisation recommendations that were well-received and developed further in high-profile W3C & EBDVF workshops.
WP8 continued producing and publishing regular online content (via Website, Social Media, YouTube, Podcast Tool) to inform stakeholders at a wider European and international level about the project objectives, processes and results.
In terms of WP9, all partners continuously monitored potential legal and ethical risks in close collaboration with KUL as the central contact point for ethics related questions.
In the second reporting period, to release TRUSTS project as a pioneer open, semantic-aware trustworthy, federated and interoperable data marketplace, the following tasks were achieved: (1) Planning of publishing the project code as an open-source on IDSA GitHub repositories to get more visibility for the project while removing other components with commercial and incompatible licence with ensuring the functionality of the core and important project features (2) Integration of semantic layer “Vocabulary Hub component” which includes the TRUSTS-IM to allow searching and discovering of assets; defining their usage policies and the access to them; notifying consumers of changes of their subscribed assets (3) Integration of IDS connector as a core component in each of TRUSTS nodes where each one; then can be considered as a separate functional and independent node regarding its internal work, data storage and usage rules and policies (4) development of interoperability component to enable harvesting of metadata from third parties and EOSC e.g. OpenAIRE and Europeana (5) Provision of brokerage and profiles services by integrating the recommender systems (6) Enabling smart contracting between the business partners using smart contracts as well as providing payment methods, e.g. PayPal, Google Pay, and Revolut. (7) Provision of sovereign datasets access by three methods (i) direct access to the datasets (ii) access using REST services (iii) access using data applications. Furthermore, a unified business model taxonomy for data marketplaces was created and a business model for TRUSTS as a data market and future federator of data marketplaces was developed, including a core business model, value-added services, and potential additional business models.