BAFFLE: A Baseline of Backpropagation-Free Federated Learning

2023/01/28 by Haozhe Feng, Feng, Haozhe, Tianyu Pang +9 · 4 citations
Computer Science · #Privacy-Preserving Technologies in Data #Adversarial Robustness in Machine Learning #Cryptography and Data Security

paper · pdf · doi:10.48550/arxiv.2301.12195

Abstract

Federated learning (FL) is a general principle for decentralized clients to train a server model collectively without sharing local data. FL is a promising framework with practical applications, but its standard training paradigm requires the clients to backpropagate through the model to compute gradients. Since these clients are typically edge devices and not fully trusted, executing backpropagation on them incurs computational and storage overhead as well as white-box vulnerability. In light of this, we develop backpropagation-free federated learning, dubbed BAFFLE, in which backpropagation is replaced by multiple forward processes to estimate gradients. BAFFLE is 1) memory-efficient and easily fits uploading bandwidth; 2) compatible with inference-only hardware optimization and model quantization or pruning; and 3) well-suited to trusted execution environments, because the clients in BAFFLE only execute forward propagation and return a set of scalars to the server. Empirically we use BAFFLE to train deep models from scratch or to finetune pretrained models, achieving acceptable results. Code is available in https://github.com/FengHZ/BAFFLE.

Cited by

• Learning without Global Backpropagation via Synergistic Information Distillation
• Federated Fine-Tuning of Sparsely-Activated Large Language Models on Resource-Constrained Devices
• Communication-Efficient Zero-Order and First-Order Federated Learning Methods over Wireless Networks
• The Cost of Avoiding Backpropagation

Related

• FedFwd: Federated Learning without Backpropagation
• FedPrune: Towards Inclusive Federated Learning
• FedLP: Layer-wise Pruning Mechanism for Communication-Computation Efficient Federated Learning
• FedBiF: Communication-Efficient Federated Learning via Bits Freezing
• FedMap: Iterative Magnitude-Based Pruning for Communication-Efficient Federated Learning