The field of federated learning and distributed systems is moving towards more efficient, privacy-preserving, and adaptable solutions. Researchers are focusing on developing novel techniques to reduce communication costs, improve model accuracy, and enhance system scalability. Recent developments have introduced innovative methods for split learning, federated graph learning, and decentralized AI-driven architectures. These advancements have significant implications for various applications, including healthcare, smart grids, and edge computing. Notably, papers such as Checkmate, FedSkipTwin, and ACME have made substantial contributions to the field by introducing new approaches to model checkpointing, client skipping, and adaptive customization of large models. Overall, the field is progressing towards more efficient, secure, and scalable solutions for distributed machine learning.
Advancements in Federated Learning and Distributed Systems
Sources
Checkmate: Zero-Overhead Model Checkpointing via Network Gradient Replication
FedSkipTwin: Digital-Twin-Guided Client Skipping for Communication-Efficient Federated Learning
Semi-Supervised Federated Learning via Dual Contrastive Learning and Soft Labeling for Intelligent Fault Diagnosis
On Splitting Lightweight Semantic Image Segmentation for Wireless Communications
ACME: Adaptive Customization of Large Models via Distributed Systems
Optimal Batch-Size Control for Low-Latency Federated Learning with Device Heterogeneity
Federated Split Learning with Improved Communication and Storage Efficiency
Decentralized AI-driven IoT Architecture for Privacy-Preserving and Latency-Optimized Healthcare in Pandemic and Critical Care Scenarios
Multi-Agent Reinforcement Learning for Sample-Efficient Deep Neural Network Mapping
A Comprehensive Data-centric Overview of Federated Graph Learning
An Experimental Study of Split-Learning TinyML on Ultra-Low-Power Edge/IoT Nodes
FOGNITE: Federated Learning-Enhanced Fog-Cloud Architecture
P3SL: Personalized Privacy-Preserving Split Learning on Heterogeneous Edge Devices
Eco-Friendly AI: Unleashing Data Power for Green Federated Learning
Caching Techniques for Reducing the Communication Cost of Federated Learning in IoT Environments
PowerTrip: Exploiting Federated Heterogeneous Datacenter Power for Distributed ML Training
A Novel Coded Computing Approach for Distributed Multi-Task Learning
FedSA-GCL: A Semi-Asynchronous Federated Graph Learning Framework with Personalized Aggregation and Cluster-Aware Broadcasting