Pradipta Deb

Research Engineer

INRIA•  October 2019 - Present

In this team, I focus on developing privacy-preserving machine learning models, unit testing, integration testing, and performance tuning specifically for deployment in low-resource devices. Project Details: Collaborative Computation Platform: The work for this project mainly involves creating a computation platform, where users can send their data for participating in a collaborative statistical study without having to think about the privacy of their data. The platform performs some U-statistics on gathered user data and securing the platform from any adversarial attack utilizing concepts and algorithms from the fields of Differential Privacy, Homomorphic Encryption, and com plete Decentralized Learning. Privacy Specification Language: For the second project, I am collaborating with a Ph.D. student on developing a new language that can mention the privacy constraint of a contract. This language supports all functionalities of a typical OOP language along with the possibility of performing probabilistic analysis of data privacy and data provenance.