AdaptSFL: Adaptive Split Federated Learning in Resource-constrained Edge Networks

2024-03-19

arXiv.org

AI Summary

Resource constraints on edge devices hinder local training of deep models. Method: This paper proposes an Adaptive Split Federated Learning (SFL) framework that dynamically coordinates Model Splitting (MS) and Model Aggregation (MA) to jointly optimize communication overhead, computational load, and convergence speed. Contribution/Results: We establish the first theoretical convergence analysis framework for SFL, quantifying how split layer placement and aggregation strategies affect convergence rate. We further design the first resource-aware adaptive SFL scheduling mechanism enabling edge-cloud co-optimization. Extensive simulations across multiple datasets demonstrate that our approach significantly reduces time-to-target-accuracy compared to baseline methods—validating its advantages in low latency and high convergence efficiency.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Search and Optimization: Learning to Search
• Planning, Routing, and Scheduling: Planning/Scheduling and Learning

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

The increasing complexity of deep neural networks poses significant barriers to democratizing them to resource-limited edge devices. To address this challenge, split federated learning (SFL) has emerged as a promising solution by of floading the primary training workload to a server via model partitioning while enabling parallel training among edge devices. However, although system optimization substantially influences the performance of SFL under resource-constrained systems, the problem remains largely uncharted. In this paper, we provide a convergence analysis of SFL which quantifies the impact of model splitting (MS) and client-side model aggregation (MA) on the learning performance, serving as a theoretical foundation. Then, we propose AdaptSFL, a novel resource-adaptive SFL framework, to expedite SFL under resource-constrained edge computing systems. Specifically, AdaptSFL adaptively controls client-side MA and MS to balance communication-computing latency and training convergence. Extensive simulations across various datasets validate that our proposed AdaptSFL framework takes considerably less time to achieve a target accuracy than benchmarks, demonstrating the effectiveness of the proposed strategies.

Problem

• Optimizing split federated learning for resource-limited edge devices
• Balancing communication-computing latency and training convergence
• Adaptive control of model splitting and client-side aggregation

Innovation

• Adaptive model splitting for resource optimization
• Client-side model aggregation balancing latency
• Convergence analysis guiding adaptive SFL framework