We currently have an open postdoctoral position. For more information, see here.

The Learn-Net project is a collaboration between Nokia and Inria coordinated by Alberto Conte (Nokia Bell Labs) and Malcolm Egan (Inria). Learn-Net is at the interface between networking and machine learning, aiming to support distributed machine learning (e.g., federated learning) and AI-native networks. The project involves 7 Inria and 2 Nokia teams with expertise in networking, privacy, and machine learning. The project hosts 4 PhD candidates and 2 postdoctoral researchers.

The Learn-Net project is organized in three axes:

Axis 1: Inference Delivery Networks. In this axis, solutions to integrate inference delivery throughout the infrastructure continuum (access, edge, regional data center, cloud) are explored.

Axis 2: Heterogeneous Learning. In this axis, deep learning techniques are tailored to heterogeneous architectures, heterogeneous data, and heterogeneous hidden models.

Axis 3: Dataflow Optimization. In this axis, new joint communication and learning strategies are investigated to satisfy constraints on privacy, energy, bandwidth, latency, and reliability.

The partners in the project are listed here.

Publications from this project are listed here.