Privacy Preserving Machine Learning over Partitioned Data
Bidragets beskrivning
This project will be concerned with using machine learning techniques privately. In many cases, the datasets used for machine learning contain sensitive or confidential information. It is observed that the commonly used machine learning methods leak this sensitive information, which limits the usage of those methods. This gave rise to the field of privacy preserving machine learning. In this work, we will be concerned with private machine learning on data distributed to multiple parties, such that each party possesses a subset of the data. This project will result in software which will be publicly available and will give third parties the ability to use the data without having to access the confidential information. This builds trust in machine learning methods since no confidential data is publicly released, and consequently, allows for bigger datasets and better training models.
Visa merStartår
2021
Slutår
2024
Beviljade finansiering
Finansiär
Finlands Akademi
Typ av finansiering
Forskardoktorer
Utlysning
Övriga uppgifter
Finansieringsbeslutets nummer
343555
Vetenskapsområden
Data- och informationsvetenskap
Forskningsområden
Tietojenkäsittelytieteet
Identifierade teman
security, privacy, cybersecurity