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

2026-03-16

AI Summary

This work addresses the challenge of deploying conventional deep neural networks in resource-constrained edge devices within federated learning settings, where post-training binarization often incurs significant accuracy degradation. To overcome this limitation, the authors propose FedBNN, a novel framework that enables end-to-end training of binary neural networks directly during local client updates. FedBNN introduces, for the first time in federated learning, a rotation-aware weight update mechanism and encodes weights as single-bit ±1 values, substantially reducing memory footprint and computational overhead during inference. Experimental results demonstrate that FedBNN achieves considerable reductions in FLOPs and storage requirements across multiple benchmark datasets while maintaining accuracy comparable to full-precision models, thereby effectively balancing privacy preservation with deployment efficiency on edge devices.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Computer Vision: Adversarial Attacks & Robustness
• Natural Language Processing: Learning & Optimization for NLP

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

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.

Problem

• Federated Learning
• Binary Neural Networks
• Edge Inference
• Model Compression
• Resource Efficiency

Innovation

• Federated Learning
• Binary Neural Networks
• Model Compression
• Edge Inference
• Rotation-aware Quantization