A5061455387- Cryptography and Data Security
- Privacy-Preserving Technologies in Data
- User Authentication and Security Systems
- Security in Wireless Sensor Networks
- Advanced Authentication Protocols Security
- Cloud Data Security Solutions
- Stochastic Gradient Optimization Techniques
- Blockchain Technology Applications and Security
- Biometric Identification and Security
- Natural Language Processing Techniques
- Advanced Graph Neural Networks
- Vehicular Ad Hoc Networks (VANETs)
- Digital Mental Health Interventions
- Traffic Prediction and Management Techniques
- Semantic Web and Ontologies
- Artificial Intelligence in Healthcare
- Topic Modeling
West Anhui University
2021-2024
Anhui Normal University
2021-2024
Federated learning has emerged as a promising paradigm by collaboratively training global model through sharing local gradients without exposing raw data. However, the shared pose threat to privacy leakage of The central server may forge aggregated results. Besides, it is common that resource-constrained devices drop out in federated learning. To solve these problems, existing solutions consider either only efficiency, or preservation. It still challenge design verifiable and lightweight...
Abstract Federated learning is a distributed machine paradigm where the goal to collaboratively train high quality global model while private training data remains local over clients. However, heterogenous distribution clients severely challenging for federated system, which damage of model. In order address this challenge, we propose prototype distillation (FedGPD) improve performance The intuition use class prototypes as knowledge instruct on client side. Eventually, objectives will be...
In edge-fog computing networks, it is very crucial to ensure that group keys are secure and reliable. recent years, there has been much interest in key updating schemes based on Chinese remainder theorem (CRT). It mostly necessary utilize the extended Euclidean algorithm compute inverse element conventional schemes. But does not always exist an element. leads computation being complicated infeasible. Thus, we propose a scheme for batch verification Euler function (EF-CRT). Group can be...
Abstract Federated learning is a distributed machine paradigm where the goal to collaboratively train high quality globalmodel while private training data remains local over clients. However, heterogenous distribution clientsis severely challenging for federated system, which damage of model. In order address thischallenge, global prototype distillation (FedGPD) proposed improve performance ofglobal The intuition use class prototypes as knowledge instruct on client side. Eventually,local...
Authentication ensures the privacy of patients by enabling access control within wireless medical sensor networks. However, many schemes do not consider resource-constrained environments, making those protocols unusable. Meanwhile, sensitive patient information continues to be stored and accessed in a central location, increasing risk “single points failure.” To solve this problem, researchers designed lightweight anonymous authentication protocol based on blockchain technology fuzzy...
Large language models (LLMs) face challenges with internal knowledge inaccuracies and outdated information. Knowledge editing has emerged as a pivotal approach to mitigate these issues. Although current techniques exhibit promising performance in single-hop reasoning tasks, they show limitations when applied multi-hop reasoning. Drawing on cognitive neuroscience the operational mechanisms of LLMs, we hypothesize that residual after causes edited revert their original answers processing...
Abstract Internet of Medical Things (IoMT) enable physicians to provide precise care over the for registered patients anywhere, bringing convenience people’s everyday life. Considering importance patient’s privacy in IoMT, data security between and medical servers should be protected. Therefore, authentication identity agreement a shared secret key are particularly important. In this work, we propose lightweight anonymous scheme IoMT. We combine blockchain technology with biometric order...
Telemedicine is one of the most rapidly de-veloping areas health care in recent years. Information Systems (TMIS) enable physicians to pro-vide remote over Internet registered patients anywhere. In this work, we propose an efficient anony-mous authentication scheme between and med-ical servers. We combine blockchain technology with biometric form a shared session secret key protect privacy through mutual au-thentication Comprehen-sive comparative measurement shows that our proposed achieves...
Ciphertext-policy attribute-based encryption (CP-ABE) is a promising primitive to achieve fine-grained access control. However, the explicit policies are prone leaking its private information. Meanwhile, most existing schemes do not effectively support malicious user revocation. To solve above issues, this paper proposes user-centric data sharing scheme that supports policy hiding, control, and The uses CP-ABE control hidden strategies. Binary tree structure utilized cloud server performs...