REDEEM

Preview

Resilient, decentralized and privacy-friendly machine learning

Cédric Gouy-Pailler , Research Engineer CEA
Sonia Ben Mokhtar, Research director CNRS

This project aims to explore new distributed learning approaches that are resilient, robust to noise and adversarial attacks, and respectful of privacy. These distributed approaches should make it possible to go beyond current federated learning. From a theoretical point of view, REDEEM aims to provide a solid foundation for the proposed approaches, particularly in the case of malicious protagonists participating in the learning phase, and with the overriding objective of ensuring data confidentiality as far as possible. In addition to new approaches to distributing learning, REDEEM also aims for efficient implementations, by offering the community open-source code and tools.

Keywords : Decentralized Machine learning, Robustness; Privacy; Byzantine-resilience; distributed optimization; consensus algorithms in machine learning; Foundation models/extremely large models

Project website: https://redeem-pepria.github.io

Missions

Our researches

Specifications and guideline for decentralized system design with identification of associated threats

Formalize the foundational framework of the project with the identification of the primary functions to be fulfilled by the learnt system (detection, classification, recommendation), mathematically defining potential existing constraints (communication, computation resources), and expliciting a set of targeted properties related to the robustness, privacy, resilience and personalization abilities of the systems.

Algorithmic aspects of decentralized learning in an adversary-free environment

Investigate decentralized learning by focusing on algorithmic aspects while assuming that participants are honest. These investigations will take into account specifications like a dynamic and heterogeneous environment, extremely large models and personalization. 

Decentralized learning under attack

Investigate novel privacy and Byzantine attacks as well as mitigation algorithms in a decentralized setting

Advanced trade-offs management

Consider advanced learning algorithms with new optimisation strategies over large and decentralized models in dynamic networks and in a hostile environment experiencing attacks