Wireless-PDGNet Implementation

This repo is related to the following paper:

Boning Li, Jake Perazzone, Ananthram Swami, and Santiago Segarra, "Learning to Transmit with Provable Guarantees in Wireless Federated Learning," submitted to IEEE TWC 2023. The preprint is available at https://arxiv.org/abs/2304.09329.

Data Generation

Please refer to https://github.com/bmatthiesen/deep-EE-opt/tree/master/data to generate channel simulations.

Power Allocation

The implementation of the proposed model and utility functions can be found under ./PDGNet/. For the training script, see ./main.py.

Federated Learning

1. MNIST Digit Classification

For the iid experiment, please see ./FL-main.ipynb.

For the non-iid experiment, please see ./FL-main-mnist-noniid.ipynb.

2. UCI Air Quality Regression

Please see ./FL-main.ipynb.

3. IMDB Sentiment Classification

Please see ./FL-main-text.ipynb.

References are provided in the citation list and also as inline notes in code files.

bl166 / WirelessFL-PDG