PhD Position F / M Goal-Oriented Communications for Federated Learning

Contexte et atouts du poste

This PhD position is supported by the Nokia-Inria Federated Learning Challenge, and will be hosted within the MARACAS team in the CITI Laboratory (a joint laboratory between Inria and INSA-Lyon). The position will be supervised by Dr Malcolm Egan and Prof. Jean-Marie Gorce (MARACAS) and Dr Alberto Conte and Marie-Line Alberia Morel (Networks Systems and Security Research, Nokia Bell Labs).

MARACAS is a research group consisting of approximately 15 people within Inria and INSA Lyon, this includes 6 PhD students, 2 postdocs, and 2 research engineers. The focus of MARACAS is in the theory, algorithms, and experimentation for communication systems, developing and applying methods in information theory, statistical signal processing and machine learning. The PhD position will complement existing projects on stochastic optimization and federated learning currently being carried out within MARACAS.

Mission confiée

The focus of the Inria-Nokia Federated Learning Challenge project is the development of federated and decentralized learning architectures and algorithms for future generation wireless communication systems. Federated learning systems support model estimation (e.g., training of deep neural networks) without providing the data stored in the clients directly to the server. Instead, the clients estimate local models and exchange the local model parameters or associated (sub)gradients with the server. This can dramatically reduce the amount of communication required and provide increased privacy. With the recent dramatic advances in the foundations of federated learning, the time is now ripe to investigate applications in emerging 6G wireless communication systems.

In the work carried out in this PhD position, we will ask two key questions :

i) What is the impact of constraints imposed by realistic wireless communication links on state-of-the-art federated learning methods?

ii) How can massive multiple access communications (e.g., coding, resource allocation, and signal processing) be optimized for state-of-the-art federated learning algorithms?

To address the first question, we will investigate the impact of wireless communications on the convergence of state-of-the-art federated learning algorithms. For the second question, new resource allocation, compression and coding schemes will be developed to optimize tradeoffs between reliability of communications (e.g., error-rates) and compression of federated learning updates.

Principales activités

The candidate will carry out research on federated learning algorithms in collaboration with members of the MARACAS Inria and Networks Systems and Security Research (Nokia Bell Labs). This includes participation in local seminars as well as in summer schools and international conferences. The candidate will also have the opportunity to do limited teaching within INSA Lyon.

Main activities :

• Analyze convergence of state-of-the-art federated learning methods in the presence of realistic limitations arising from emerging 6G wireless communication systems.
• Develop new coding, resource allocation, and compression schemes to mitigate the impact of errors arising from wireless communications.

Compétences

Technical skills :

The candidate must have a high level of spoken and written English.

Avantages

Rémunération

1st and 2nd year : 2 100 euros gross salary / month

3rd year : 2 190 euros gross salary / month