A5079949925- Privacy-Preserving Technologies in Data
- Text and Document Classification Technologies
- Software Engineering Research
- Parallel Computing and Optimization Techniques
- Embedded Systems Design Techniques
- Advanced Graph Neural Networks
- Recommender Systems and Techniques
- Image Retrieval and Classification Techniques
- Stochastic Gradient Optimization Techniques
- Data Stream Mining Techniques
- Machine Learning and Data Classification
- Geoscience and Mining Technology
- Rough Sets and Fuzzy Logic
- Face and Expression Recognition
- Wireless Networks and Protocols
- Data Quality and Management
- Cryptography and Data Security
Southeast University
2024
Zhangzhou Normal University
2020-2021
Federated learning is a new framework that protects data privacy and allows multiple devices to cooperate in training machine models. Previous studies have proposed approaches eliminate the challenges posed by non-iid inter-domain heterogeneity issues. However, they ignore \textbf{spatio-temporal} formed different distributions of increasing task intra-domain. Moreover, global generally long-tailed distribution rather than assuming balanced practical applications. To tackle dilemma, we...
Download This Paper Open PDF in Browser Add to My Library Share: Permalink Using these links will ensure access this page indefinitely Copy URL DOI
Federated learning (FL) is an efficient framework designed to facilitate collaborative model training across multiple distributed devices while preserving user data privacy. A significant challenge of FL data-level heterogeneity, i.e., skewed or long-tailed distribution private data. Although various methods have been proposed address this challenge, most them assume that the underlying global are uniformly all clients. This article investigates heterogeneity with a brief review and...
Federated learning is an efficient framework designed to facilitate collaborative model training across multiple distributed devices while preserving user data privacy. A significant challenge of federated data-level heterogeneity, i.e., skewed or long-tailed distribution private data. Although various methods have been proposed address this challenge, most them assume that the underlying global uniformly all clients. This paper investigates heterogeneity with a brief review and redefines...
Federated learning shows promise as a privacy-preserving collaborative technique. Existing heterogeneous federated mainly focuses on skewing the label distribution across clients. However, most approaches suffer from catastrophic forgetting and concept drift, when global of all classes is extremely unbalanced data client dynamically evolves over time. In this paper, we study new task, i.e., Dynamic Heterogeneous Learning (DHFL), which addresses practical scenario where distributions exist...