edge-learning

We present UDP-based aggregation algorithms for federated learning. We also present a scalable framework for practical federated learning. We empirically evaluate the performance by training deep convolutional neural networks on the MNIST dataset and the CIFAR10 dataset.

Firmware for AkidaTag, the ultra-low-power AIoT board pairing a Nordic nRF5340 with BrainChip's Akida AKD1500: keyword spotting, on-device edge learning, BLE model transfer and firmware updates.

Optimization of Data Allocation with Training Time Constraints in Heterogeneous Edge Learning

"DL_WC is a repository containing python code and resources related to my work on deep learning in the context of wireless communication"