Enhancing Fairness and Efficiency in Federated Learning

Advances in Federated Learning: Addressing Heterogeneity and Privacy Concerns

Recent developments in federated learning (FL) have focused on addressing the inherent challenges of data heterogeneity and privacy preservation. Innovations in this field are moving towards more equitable and efficient learning frameworks that mitigate biases and reduce communication costs. Key advancements include novel clustering and weighting mechanisms to ensure fairness across diverse client datasets, as well as the integration of advanced cryptographic techniques to enhance privacy without compromising model performance.

One significant trend is the use of spectral knowledge and personalized preferences in graph learning to handle structural heterogeneity across domains. This approach allows for more adaptive and efficient model training in cross-domain scenarios. Additionally, the incorporation of self-supervised learning and opportunistic inference in continuous monitoring applications demonstrates a shift towards more energy-efficient and practical solutions for real-world deployment.

Notable papers in this area include:

Equitable Federated Learning with Activation Clustering: Introduces a clustering-based framework to mitigate bias and achieve fair convergence rates.
FedMABA: Towards Fair Federated Learning through Multi-Armed Bandits Allocation: Proposes a multi-armed bandit-based algorithm to enhance fairness in non-I.I.D. scenarios.
Anatomical 3D Style Transfer Enabling Efficient Federated Learning with Extremely Low Communication Costs: Utilizes 3D style transfer to align models with minimal communication costs.

These developments collectively push the boundaries of FL, making it a more robust and privacy-conscious approach to distributed machine learning.

Sources

Equitable Federated Learning with Activation Clustering

Free-Rider and Conflict Aware Collaboration Formation for Cross-Silo Federated Learning

Impact of Leakage on Data Harmonization in Machine Learning Pipelines in Class Imbalance Across Sites

A New Perspective to Boost Performance Fairness for Medical Federated Learning

Federated Anomaly Detection for Early-Stage Diagnosis of Autism Spectrum Disorders using Serious Game Data

FedMABA: Towards Fair Federated Learning through Multi-Armed Bandits Allocation

Anatomical 3D Style Transfer Enabling Efficient Federated Learning with Extremely Low Communication Costs

FedSSP: Federated Graph Learning with Spectral Knowledge and Personalized Preference

FuseFL: One-Shot Federated Learning through the Lens of Causality with Progressive Model Fusion

When Less is More: Achieving Faster Convergence in Distributed Edge Machine Learning

Federated Time Series Generation on Feature and Temporally Misaligned Data

A Unified Solution to Diverse Heterogeneities in One-shot Federated Learning

On Homomorphic Encryption Based Strategies for Class Imbalance in Federated Learning

Self-Supervised Learning and Opportunistic Inference for Continuous Monitoring of Freezing of Gait in Parkinson's Disease

Personalized Federated Learning with Mixture of Models for Adaptive Prediction and Model Fine-Tuning

Online Mirror Descent for Tchebycheff Scalarization in Multi-Objective Optimization

$r$Age-$k$: Communication-Efficient Federated Learning Using Age Factor

$\mathsf{OPA}$: One-shot Private Aggregation with Single Client Interaction and its Applications to Federated Learning

Vertical Federated Learning with Missing Features During Training and Inference

FISC: Federated Domain Generalization via Interpolative Style Transfer and Contrastive Learning

Federated UCBVI: Communication-Efficient Federated Regret Minimization with Heterogeneous Agents

Federated Learning under Periodic Client Participation and Heterogeneous Data: A New Communication-Efficient Algorithm and Analysis

(FL)$^2$: Overcoming Few Labels in Federated Semi-Supervised Learning

Communication-Efficient Federated Learning over Wireless Channels via Gradient Sketching

Local Superior Soups: A Catalyst for Model Merging in Cross-Silo Federated Learning

Generative AI-Powered Plugin for Robust Federated Learning in Heterogeneous IoT Networks

On Sampling Strategies for Spectral Model Sharding

Federated Black-Box Adaptation for Semantic Segmentation