Engineering Innovation in
Digital Industry
Our research project builds a validated, federated, privacy-preserving Machine Learning platform tested on industrial data.
Approach & Solution
Approach
Machine Learning has become a disruptive technology in many industries as it provides a scalable and flexible way to incorporate data-driven decision making. Partners in a data economy can benefit from sharing datasets in order to improve the quality of their own respective machine learning models.
In this sense, Industrial Data Platforms (IDPs) serve the need of establishing trusted networks where data can be transferred, accessed, and used in a secure mode. Currently IDPs are still encountering several obstacles that limit their usage: legal barriers, data ownership, different privacy policies among actors, mistrust towards how data is managed (e.g. fear of confidentiality breaches and of personal information leakage).
In this sense, Industrial Data Platforms (IDPs) serve the need of establishing trusted networks where data can be transferred, accessed, and used in a secure mode. Currently IDPs are still encountering several obstacles that limit their usage: legal barriers, data ownership, different privacy policies among actors, mistrust towards how data is managed (e.g. fear of confidentiality breaches and of personal information leakage).
Solution
MUSKETEER leverages new paradigms in machine learning research, such as Federated Machine Learning, Privacy Preserving Machine Learning and protection against Adversarial Attacks.
MUSKETEER aims to develop an Industrial Data platform which follows actual industrial standards and offers scalable algorithms for federated and privacy-preserving machine learning techniques with detection and mitigation of adversarial attacks and a rewarding model capable to fairly monetize datasets according to the real data value.
In the project, Engineering makes use of its skill and knowledge to design and develop Federated Machine Learning client connectors for the scalable, secure, trusted and privacy aware communication with the server side. The project has received co-funding from the European Union's Horizon 2020 programme - Contract No. 824988.
Read the final Press Release.
MUSKETEER aims to develop an Industrial Data platform which follows actual industrial standards and offers scalable algorithms for federated and privacy-preserving machine learning techniques with detection and mitigation of adversarial attacks and a rewarding model capable to fairly monetize datasets according to the real data value.
In the project, Engineering makes use of its skill and knowledge to design and develop Federated Machine Learning client connectors for the scalable, secure, trusted and privacy aware communication with the server side. The project has received co-funding from the European Union's Horizon 2020 programme - Contract No. 824988.
Read the final Press Release.
Results
Machine Learning over a high variety of privacy-preserving scenarios
Robustness against external and internal threats
Support to Data Economy development
Standardized and extensible architecture
Industrial demonstration in operational environment
Project value
Innovation
Cost cutting
Enabling Technologies
Cloud
AI Advanced Analytics
Cybersecurity
Project Team
Research & Innovation