Pre-Deployment Complexity Estimation for Federated Perception Systems

TL;DR

This paper introduces a practical, classifier-agnostic framework to estimate the complexity of federated perception tasks, aiding resource planning and feasibility assessment before training begins.

Contribution

It proposes a novel complexity metric that combines data properties and client characteristics to predict federated learning difficulty and communication costs.

Findings

The metric correlates strongly with federated learning performance.

It effectively predicts communication effort to reach accuracy targets.

The framework aids in resource planning and dataset assessment.

Abstract

Edge AI systems increasingly rely on federated learning to train perception models in distributed, privacy-preserving, and resource-constrained environments. Yet, before training begins, practitioners often lack practical tools to estimate how difficult a federated learning task will be in terms of achievable accuracy and communication cost. This paper presents a classifier-agnostic, pre-deployment framework for estimating learning complexity in federated perception systems by jointly modeling intrinsic properties of the data and characteristics of the distributed environment. The proposed complexity metric integrates dataset attributes such as dimensionality, sparsity, and heterogeneity with factors related to the composition of participating clients. Using federated learning as a representative distributed training setting, we examine how learning difficulty varies across different federated configurations. Experiments on multiple variants of the MNIST dataset and CIFAR dataset show that the proposed metric strongly correlates with federated learning performance and the communication effort required to reach fixed accuracy targets. These findings suggest that complexity estimation can serve as a practical diagnostic tool for resource planning, dataset assessment, and feasibility evaluation in edge-deployed perception systems.