Beyond Uniform Compression: Budgeted Transmission Allocation for Extreme Federated Learning

2026-09-30

AI Summary

This study addresses the communication bottleneck caused by high-dimensional model updates in federated learning and the limitation of uniform compression schemes that overlook parameter heterogeneity. We propose Layer-wise Budget-Adaptive Transmission (LBAT), which, for the first time, reformulates federated communication under extreme uplink bandwidth constraints as a resource allocation problem, departing from traditional uniform compression paradigms. By dynamically evaluating the transmission value of each layer and employing an exact byte-level dynamic programming algorithm, LBAT achieves optimal layer-differentiated configurations of rank and quantization bit-width under strict budgets. Experiments demonstrate that LBAT significantly outperforms existing uniform compression baselines across heterogeneous tabular prediction and generation tasks, substantially improving the communication–utility trade-off while effectively preserving data distribution fidelity.

Abstract

Federated learning faces severe communication bottlenecks when clients upload high-dimensional model updates. Existing methods often compress these updates uniformly across all layers. This uniform approach ignores the heterogeneous value of different parameter blocks and wastes limited bandwidth on insensitive layers. To address this issue, we propose Layer-wise Budgeted Adaptive Transmission (LBAT). LBAT reframes federated communication under extreme uplink budgets as a resource allocation problem. Our framework dynamically estimates the transmission value of different layers utilising local training signals. It then employs an exact byte dynamic programming allocator to determine optimal rank and bit configurations under strict budgets. We validate LBAT on highly heterogeneous federated tabular prediction and data generation tasks. Extensive experiments demonstrate that LBAT consistently outperforms uniform rank, uniform quantisation, and fixed compression baselines across various extreme budget regimes. Furthermore, it achieves significantly better communication and utility tradeoffs while preserving essential distributional fidelity.

Problem

• Federated Learning
• Communication Bottleneck
• Budgeted Transmission
• Uniform Compression
• Resource Allocation

Innovation

• Federated Learning
• Budgeted Transmission Allocation
• Dynamic Programming
• Layer-wise Adaptive Compression
• Extreme Communication Budgets