Privacy-Preserving Techniques in Deep Learning for Mobile Computing
1Guangzhou University, Guangzhou, China
2University of Illinois Springfield, Springfield, USA
3Chinese Academy of Sciences, Beijing, China
Privacy-Preserving Techniques in Deep Learning for Mobile Computing
Description
In the past few years, deep learning has achieved great breakthroughs in computer vision, speech recognition, and many other areas. To support the training of deep learning, large datasets are collected from different entities in the real world. Along with tremendous efforts devoted to mobile computing, the high communication efficiency in current technology such as 5G and the strong calculation power in various mobile terminals have drawn great attention from distributed learning.
In the big data background, how to exploit potentialities from the distributed deep learning with strong mobile computation has become an important and meaningful issue to discuss. Meanwhile, extensive usages training data has also raised great concerns about data privacy. Although many privacy-preserving techniques have offered solutions to protect our personal data, many still have practical limitations (e.g., not effective or efficient enough, less user-centric). As a result, there is an urgent need for innovative privacy-preserving techniques to adequately explore the big data vs. privacy dilemma in deep learning.
This Special Issue aims to advance privacy technologies and methodologies in deep learning and further promote research activities in large-scale data-based service. The Special Issue seeks original theory- and application-driven studies to address some emerging issues and challenges from the perspective of privacy-preserving deep learning and its applications in areas such as natural language processing, computer vision, speech recognition, and so on. Original research and review articles are welcome.
Potential topics include but are not limited to the following:
- Deep learning methods and architecture based on mobile computation
- Privacy-preserving crowdsensing systems for mobile computing
- Private information storage, aggregation, and retrieval in mobile computing
- Privacy-enhanced deep learning for mobile computing, such as federated learning, etc.
- Blockchain-based deep learning techniques for privacy-preserving of mobile computing
- Privacy of deep learning in different big data contexts such as online social networks, healthcare, IoT, and e-government
- Individual privacy-preserving methods in distributed deep learning for mobile computing
- Evaluation mechanism of models in distributed deep learning with multiple terminals
- Privacy computing theory and language for mobile computation
- Auditable guarantee in cryptography-based privacy-preserving methods for mobile computing
- Incentives in game theory for the trade-off between privacy and accuracy, utility, reliability, and fault-tolerance in deep learning for mobile computing