Single-shot Hyper-parameter Optimization for Federated Learning: A General Algorithm & Analysis
It addresses hyper-parameter tuning in federated learning, a critical bottleneck for distributed and heterogeneous data systems, with incremental improvements in efficiency.
The paper tackles hyper-parameter optimization for federated learning by introducing FLoRA, a framework that enables single-shot optimization with minimal communication overhead, achieving significant model accuracy improvements on seven OpenML datasets.
We address the relatively unexplored problem of hyper-parameter optimization (HPO) for federated learning (FL-HPO). We introduce Federated Loss SuRface Aggregation (FLoRA), a general FL-HPO solution framework that can address use cases of tabular data and any Machine Learning (ML) model including gradient boosting training algorithms and therefore further expands the scope of FL-HPO. FLoRA enables single-shot FL-HPO: identifying a single set of good hyper-parameters that are subsequently used in a single FL training. Thus, it enables FL-HPO solutions with minimal additional communication overhead compared to FL training without HPO. We theoretically characterize the optimality gap of FL-HPO, which explicitly accounts for the heterogeneous non-IID nature of the parties' local data distributions, a dominant characteristic of FL systems. Our empirical evaluation of FLoRA for multiple ML algorithms on seven OpenML datasets demonstrates significant model accuracy improvements over the considered baseline, and robustness to increasing number of parties involved in FL-HPO training.