- Submission Deadline: 15 September 2023
Publication: Late 2024
The global big data market is experiencing tremendous growth, driven by advancements in the Internet, mobile networks, IoT devices, and AI technologies. While digital transformation and social media applications are accelerating data generation, AI-driven big data analytics are enabling industries and businesses to expand their market, enhance customer relationships, reduce operational costs, improve efficiency, and discover new business opportunities, establishing data as the new oil of the 21st century.
However, the deployment of AI technologies in big data applications is hindered by growing security and privacy concerns. Recent data privacy regulations, such as GDPR in Europe and HIPAA in the USA, limit data sharing and cause data silos. The European Commission proposed the Artificial Intelligence Act in 2021 to regulate the usage of AI in the EU.
To address data privacy issues, federated learning has been proposed to enable collaborative training of AI models between different users or organizations without sharing data. It has garnered significant attention from both academia and industry. However, federated learning still faces several crucial challenges that impede its practical deployment into real-world big data applications. For instance, statistical data heterogeneity among the participants, such as non-IID datasets and various data qualities, may slow down the training process and depress the performance of the learned models. Furthermore, system heterogeneity arising from differences in hardware capacities, network connections, reliability, and availability of participants poses further challenges for the distributed optimization of the learning task. Although current federated learning schemes only require the transmission of gradients information, they may still potentially reveal private and sensitive information. Many privacy-preserving schemes for federated learning impose significant computing and communication overheads, which must be further addressed. Finally, effectively incentivizing more users to actively participate in federated learning by providing high-quality data and more computing resources is of paramount importance.
We believe this special issue will shed more light on the recent research updates in federated learning that benefit the big data applications. Topics of interest include, but are not limited to:
- Vertical federated learning
- Data-centric federated learning
- Data augmentation in federated learning
- Data quality in federated learning
- Data versioning in federated learning
- Federated learning with Non-IID data
- Federated learning and blockchain
- Privacy-preserving techniques in federated learning
- Incentive mechanism design for federated learning
- Attacks and defenses in federated learning
- Deep reinforcement learning and federated learning
- Federated learning for content recommendation
- Federated learning for large models
- Simulation systems and platforms for large-scale federated learning
For author information and guidelines on submission criteria, please visit the TBD’s Author Information page. Please submit papers through the ScholarOne system, and be sure to select the special-issue name. Manuscripts should not be published or currently submitted for publication elsewhere. Please submit only full papers intended for review, not abstracts, to the ScholarOne portal.
Contact the guest editors:
- Xiaowen Chu, The Hong Kong University of Science and Technology (Guangzhou)
- Wei Wang, The Hong Kong University of Science and Technology, Hong Kong, China
- Cong Wang, The City University of Hong Kong, Hong Kong, China
- Yang Liu, Tsinghua University, China
- Rongfei Zeng, Northeastern University, China
- Christopher G. Brinton, Purdue University, USA