Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Towards One-shot Federated Learning: Advances, Challenges, and Future Directions

Flora Amato, Lingyu Qiu, Mohammad Tanveer, Salvatore Cuomo, Fabio Giampaolo, Francesco Piccialli

TL;DR

This survey addresses One-shot Federated Learning, a single-round, communication-efficient paradigm designed for privacy-sensitive, resource-constrained environments. It organizes existing work into theory-based optimization, knowledge-distillation ensembles, data-heterogeneity handling, and adversarial robustness, and surveys mechanisms for ensemble aggregation and synthetic data generation. The authors review datasets, benchmarks, and open-source implementations, and identify key open challenges—data access limitations, trust and privacy, scalability, and privacy-accuracy trade-offs—while proposing future directions such as IoT and satellite integration, SciML fusion, and standardized benchmarks. By providing a structured framework and practical guidance, the paper aims to accelerate the design and deployment of effective One-shot FL systems in real-world, privacy-conscious settings.

Abstract

One-shot FL enables collaborative training in a single round, eliminating the need for iterative communication, making it particularly suitable for use in resource-constrained and privacy-sensitive applications. This survey offers a thorough examination of One-shot FL, highlighting its distinct operational framework compared to traditional federated approaches. One-shot FL supports resource-limited devices by enabling single-round model aggregation while maintaining data locality. The survey systematically categorizes existing methodologies, emphasizing advancements in client model initialization, aggregation techniques, and strategies for managing heterogeneous data distributions. Furthermore, we analyze the limitations of current approaches, particularly in terms of scalability and generalization in non-IID settings. By analyzing cutting-edge techniques and outlining open challenges, this survey aspires to provide a comprehensive reference for researchers and practitioners aiming to design and implement One-shot FL systems, advancing the development and adoption of One-shot FL solutions in a real-world, resource-constrained scenario.

Towards One-shot Federated Learning: Advances, Challenges, and Future Directions

TL;DR

This survey addresses One-shot Federated Learning, a single-round, communication-efficient paradigm designed for privacy-sensitive, resource-constrained environments. It organizes existing work into theory-based optimization, knowledge-distillation ensembles, data-heterogeneity handling, and adversarial robustness, and surveys mechanisms for ensemble aggregation and synthetic data generation. The authors review datasets, benchmarks, and open-source implementations, and identify key open challenges—data access limitations, trust and privacy, scalability, and privacy-accuracy trade-offs—while proposing future directions such as IoT and satellite integration, SciML fusion, and standardized benchmarks. By providing a structured framework and practical guidance, the paper aims to accelerate the design and deployment of effective One-shot FL systems in real-world, privacy-conscious settings.

Abstract

One-shot FL enables collaborative training in a single round, eliminating the need for iterative communication, making it particularly suitable for use in resource-constrained and privacy-sensitive applications. This survey offers a thorough examination of One-shot FL, highlighting its distinct operational framework compared to traditional federated approaches. One-shot FL supports resource-limited devices by enabling single-round model aggregation while maintaining data locality. The survey systematically categorizes existing methodologies, emphasizing advancements in client model initialization, aggregation techniques, and strategies for managing heterogeneous data distributions. Furthermore, we analyze the limitations of current approaches, particularly in terms of scalability and generalization in non-IID settings. By analyzing cutting-edge techniques and outlining open challenges, this survey aspires to provide a comprehensive reference for researchers and practitioners aiming to design and implement One-shot FL systems, advancing the development and adoption of One-shot FL solutions in a real-world, resource-constrained scenario.
Paper Structure (57 sections, 4 equations, 8 figures, 3 tables)

This paper contains 57 sections, 4 equations, 8 figures, 3 tables.

Table of Contents

  1. 1 Introduction
  2. 1.1 Background
  3. 1.1.1 Federated Learning Overview
  4. 1.1.2 Communication Costs in Federated Learning
  5. 1.1.3 One-shot Federated Learning Concept
  6. 1.2 Motivation
  7. 1.3 Contribution
  8. 1.4 Applicability
  9. 1.4.1 Web3.0
  10. 1.4.2 Autonomous Vehicles
  11. 1.4.3 IoT and Edge Computing
  12. 1.5 Survey structure
  13. 2 Related literature survey
  14. 2.1 Retrieval strategy
  15. 2.1.1 Articles search
  16. ...and 42 more sections

Figures (8)

  • Figure 1: The training and deployment process of OSFL.
  • Figure 2: The structure of the survey
  • Figure 3: The methodology for retrieving research articles in this study follows a three-step process: identification, screening and eligibility, and inclusion. In the identification step, relevant keywords were selected to gather research articles related to the topic under investigation. During the screening and eligibility step, specific criteria were defined to filter out literature that did not align with the study's goals and objectives. Finally, in the inclusion step, the articles that satisfied the predefined criteria and matched the research requirements were incorporated into the study.
  • Figure 4: The number of included papers per year.
  • Figure 5: The pipeline of One-shot FL training.
  • ...and 3 more figures