A5078640515- Privacy-Preserving Technologies in Data
- Cryptography and Data Security
- Advanced Graph Neural Networks
- Stochastic Gradient Optimization Techniques
- Complexity and Algorithms in Graphs
- AI in cancer detection
- Distributed systems and fault tolerance
- Caching and Content Delivery
- Graph Theory and Algorithms
- Advanced Graph Theory Research
- Blockchain Technology Applications and Security
- Radiation Detection and Scintillator Technologies
- Data Quality and Management
- Tensor decomposition and applications
- Radiomics and Machine Learning in Medical Imaging
- Recommender Systems and Techniques
- Face and Expression Recognition
- Medical Imaging and Analysis
- Nuclear Physics and Applications
- Human Mobility and Location-Based Analysis
- Non-Destructive Testing Techniques
- Random Matrices and Applications
East China Normal University
2023-2024
As privacy issues are receiving increasing attention within the Natural Language Processing (NLP) community, numerous methods have been proposed to sanitize texts subject differential privacy. However, state-of-the-art text sanitization mechanisms based on a relaxed notion of metric local (MLDP) do not apply non-metric semantic similarity measures and cannot achieve good privacy-utility trade-offs. To address these limitations, we propose novel Customized Text (CusText) mechanism original...
Social recommendation has shown promising improvements over traditional systems since it leverages social correlation data as an additional input. Most existing work assumes that all are available to the platform. However, in practice, user-item interaction (e.g.,rating) and user-user usually generated by different platforms, both of which contain sensitive information. Therefore, "How perform secure efficient across where highly-sparse nature" remains important challenge. In this work, we...
Logistic regression is an algorithm widely used for binary classification in various real-world applications such as fraud detection, medical diagnosis, and recommendation systems. However, training a logistic model with data from different parties raises privacy concerns. Secure Multi-Party Computation (MPC) cryptographic tool that allows multiple to train jointly without compromising privacy. The efficiency of the online phase becomes crucial when dealing large-scale practice. In this...
Knowledge Graph (KG) has attracted more and companies' attention for its ability to connect different types of data in meaningful ways support rich services. However, the isolation problem limits performance KG prevents further development. That is, multiple parties have their own KGs but they cannot share with each other due regulation or competition reasons. Therefore, how conduct privacy preserving becomes an important research question answer. related tasks collaboratively on basis...
Recently, Niu, et. al. introduced a new variant of Federated Learning (FL), called Submodel (FSL). Different from traditional FL, each client locally trains the submodel (e.g., retrieved servers) based on its private data and uploads at choice to servers. Then all clients aggregate their submodels finish iteration. Inevitably, FSL introduces two privacy-preserving computation tasks, i.e., Private Retrieval (PSR) Secure Aggregation (SSA). Existing work fails provide loss-less scheme, or has...
As privacy issues are receiving increasing attention within the Natural Language Processing (NLP) community, numerous methods have been proposed to sanitize texts subject differential privacy. However, state-of-the-art text sanitization mechanisms based on metric local (MLDP) do not apply non-metric semantic similarity measures and cannot achieve good trade-offs between utility. To address above limitations, we propose a novel Customized Text (CusText) mechanism original...
K-means is one of the most widely used clustering models in practice. Due to problem data isolation and requirement for high model performance, how jointly build practical secure multiple parties has become an important topic many applications industry. Existing work on this mainly two types. The first type efficiency advantages, but information leakage raises potential privacy risks. second provable inefficient even helpless large-scale sparsity scenario. In paper, we propose a new...
One way to classify private set intersection (PSI) for secure 2-party computation is whether the (a) revealed both parties or (b) hidden from while only computing function of matched payload exposed. Both aim provide cryptographic security avoiding exposing unmatched elements other. They may, however, be insufficient achieve and privacy in one practical scenario: when required information leaked through function's output must considered legal, ethical, competitive reasons. Two parties, such...