Federated Learning of Binary Neural Networks: Enabling Low-Cost Inference

TL;DR

FedBNN introduces a rotation-aware binary neural network framework for federated learning, enabling low-cost, resource-efficient inference on edge devices without significant accuracy loss.

Contribution

The paper presents FedBNN, a novel binary neural network approach for federated learning that directly learns binary weights, reducing model size and computation during inference.

Findings

FedBNN significantly reduces model size and FLOPs compared to real-valued models.

FedBNN maintains comparable accuracy to existing federated methods on benchmark datasets.

Resource consumption during inference is substantially lowered with FedBNN.

Abstract

Federated Learning (FL) preserves privacy by distributing training across devices. However, using DNNs is computationally intensive at the low-powered edge during inference. Edge deployment demands models that simultaneously optimize memory footprint and computational efficiency, a dilemma where conventional DNNs fail by exceeding resource limits. Traditional post-training binarization reduces model size but suffers from severe accuracy loss due to quantization errors. To address these challenges, we propose FedBNN, a rotation-aware binary neural network framework that learns binary representations directly during local training. By encoding each weight as a single bit $\{+1, -1\}$ instead of a $32$-bit float, FedBNN shrinks the model footprint, significantly reducing runtime (during inference) FLOPs and memory requirements in comparison to federated methods using real models. Evaluations across multiple benchmark datasets demonstrate that FedBNN significantly reduces resource consumption while performing similarly to existing federated methods using real-valued models.