A5100405359- Cryptography and Data Security
- Cloud Data Security Solutions
- Blockchain Technology Applications and Security
- Privacy-Preserving Technologies in Data
- Chaos-based Image/Signal Encryption
- IoT and Edge/Fog Computing
- Network Security and Intrusion Detection
- Advanced Malware Detection Techniques
- Sparse and Compressive Sensing Techniques
- Metaheuristic Optimization Algorithms Research
- Caching and Content Delivery
- Advanced Multi-Objective Optimization Algorithms
- Recommender Systems and Techniques
- Internet Traffic Analysis and Secure E-voting
- Cloud Computing and Resource Management
- Machine Learning and Data Classification
- Advanced Data Storage Technologies
- Peer-to-Peer Network Technologies
- Cryptographic Implementations and Security
- Advanced Database Systems and Queries
- Domain Adaptation and Few-Shot Learning
- Service-Oriented Architecture and Web Services
- User Authentication and Security Systems
- Wireless Communication Security Techniques
- Natural Language Processing Techniques
Shenzhen University
2016-2025
Sichuan University
2025
Yibin University
2025
Sanya University
2025
Hainan University
2025
Chinese University of Hong Kong
2012-2025
Idiap Research Institute
2024
Jiangxi University of Science and Technology
2024
Shanghai Jiao Tong University
2024
Chongqing Vocational Institute of Engineering
2021-2023
Few-shot learning is challenging for algorithms that learn each task in isolation and from scratch. In contrast, meta-learning learns many related tasks a meta-learner can new more accurately faster with fewer examples, where the choice of meta-learners crucial. this paper, we develop Meta-SGD, an SGD-like, easily trainable initialize adapt any differentiable learner just one step, on both supervised reinforcement learning. Compared to popular LSTM, Meta-SGD conceptually simpler, easier...
Statistical and systematic challenges in collaboratively training machine learning models across distributed networks of mobile devices have been the bottlenecks real-world application federated learning. In this work, we show that meta-learning is a natural choice to handle these issues, propose framework FedMeta, where parameterized algorithm (or meta-learner) shared, instead global model previous approaches. We conduct an extensive empirical evaluation on LEAF datasets production dataset,...
A central challenge in training classification models the real-world federated system is learning with non-IID data. To cope this, most of existing works involve enforcing regularization local optimization or improving model aggregation scheme at server. Other also share public datasets synthesized samples to supplement under-represented classes introduce a certain level personalization. Though effective, they lack deep understanding how data heterogeneity affects each layer model. In this...
This paper reveals an intrinsic relationship between secure cloud storage and network coding for the first time. Secure was proposed only recently while has been studied more than ten years. Although two areas are quite different in their nature independently, we show how to construct a protocol given any protocol. gives rise systematic way protocols. Our construction is under definition which captures real world usage of storage. Furthermore, propose specific protocols based on recent In...
Cloud storage is more and prevalent in practice, thus how to check its integrity becomes increasingly essential. A classical solution identity-based (ID-based) provable data possession (PDP), which supports certificateless cloud auditing without entire user data. However, existing ID-PDP protocols always require that users outsource blocks, authenticators a small-sized file tag the cloud, make use of heavy elliptic curve cryptography over bilinear pairing. These disadvantages would result...
Next-generation wireless technology and machine-to-machine can provide the ability to connect share data at any time among IoT smart devices. However, traditional centralized sharing/trading mechanism lacks trust guarantee cannot satisfy real-time requirement. Distributed systems, especially blockchain, us with promising solutions. In this paper, we propose a blockchain based non-repudiation scheme for trading resolve credibility limits. The proposed has two parts, i.e., an arbitration...
The architecture of two-tiered sensor networks, where storage nodes serve as an intermediate tier between sensors and a sink for storing data processing queries, has been widely adopted because the benefits power saving well efficiency query processing. However, importance also makes them attractive to attackers. In this paper, we propose SafeQ, protocol that prevents attackers from gaining information both collected issued queries. SafeQ allows detect compromised when they misbehave. To...
Over the past few years, we have been trying to build an end-to-end system at Wisconsin manage unstructured data, using extraction, integration, and user interaction. This paper describes key information extraction (IE) challenges that run into, sketches our solutions. We discuss in particular developing a declarative IE language, optimizing for this generating provenance, incorporating feedback into process, novel wiki-based interface feedback, best-effort IE, pushing RDBMSs, more. Our work...
With cloud computing and mobile becoming more popular, there are a lot potential applications for computation outsourcing to the cloud. This paper investigates linear regression problem, which is quite common engineering task employed in various applications, as case study find out possible problems that need be solved. We propose two protocols can enable secure efficient of The protect client's data privacy well at same time have good efficiency. show all subtleties techniques designing...
Empowered by today's rich tools for media generation and distribution, the convenient Internet access, crowdsourced streaming generalizes single-source paradigm including massive contributors a video channel. It calls joint optimization along path from crowdsourcers, through servers, to end-users minimize overall latency. The dynamics of sources, together with globalized request demands high computation demand each sourcer, make live challenging even powerful support modern cloud computing....
AI-driven mineral prospectivity mapping (MPM) is a valid and increasingly accepted tool for delineating the targets of exploration, but it suffers from noisy unrepresentative input features. In this study, set fractal multifractal methods, including box-counting calculation, concentration–area modeling, analyses, were employed to excavate underlying nonlinear mineralization-related information geological Based on these multiple feature selection criteria, namely prediction–area plot, K-means...