Title: Enhancing security for data federation distributed in silos through Polystore system
Abstract: Training the recent machine learning model is often scattered across disparate collections of datasets, which can be referred to as data silos. Due to confidentiality and legislative constraints, the data often cannot leave the premises of these data silos. This fragmentation poses a significant challenge for data-intensive learning applications. Advanced research work focuses on how to integrate distributed data from silos for data analysis while ensuring data security.
In this talk, he will present an exhaustive state-of-the-art on data privacy and the surrounding security measures by tracing the evolution of data privacy and security concepts. He will present an approach that enables remote access to heterogeneous data from multiple partners in different countries while adhering to a set of security constraints. Polystore paradigm is considered to be the system for federating multi-sources data. He will present a resilient Access Control Mechanism for data sharing through Polystore.
Français 
English (UK)