A5100333755- Advanced Malware Detection Techniques
- Privacy-Preserving Technologies in Data
- Security and Verification in Computing
- Blockchain Technology Applications and Security
- User Authentication and Security Systems
- Cryptography and Data Security
- Network Security and Intrusion Detection
- Adversarial Robustness in Machine Learning
- Cloud Computing and Resource Management
- Advanced Computational Techniques and Applications
- Advanced Sensor and Control Systems
- Software Testing and Debugging Techniques
- Neural Networks Stability and Synchronization
- Opportunistic and Delay-Tolerant Networks
- IoT and Edge/Fog Computing
- Bluetooth and Wireless Communication Technologies
- Smart Grid Security and Resilience
- Cloud Data Security Solutions
- Anomaly Detection Techniques and Applications
- Higher Education and Teaching Methods
- Digital and Cyber Forensics
- Air Quality Monitoring and Forecasting
- Embedded Systems and FPGA Design
- Software-Defined Networks and 5G
- Internet Traffic Analysis and Secure E-voting
China Academy of Information and Communications Technology
2025
Hubei University of Technology
2025
Chinese Academy of Medical Sciences & Peking Union Medical College
2025
Drexel University
2015-2024
University of Jinan
2015-2024
Zhengzhou Railway Vocational & Technical College
2010-2024
Chongqing Jiaotong University
2024
Shanghai University of Engineering Science
2024
First People's Hospital of Chongqing
2024
Chongqing Medical University
2024
Deep learning can achieve higher accuracy than traditional machine algorithms in a variety of tasks. Recently, privacy-preserving deep has drawn tremendous attention from information security community, which neither training data nor the model is expected to be exposed. Federated popular mechanism, where multiple parties upload local gradients server and updates parameters with collected gradients. However, there are many problems neglected federated learning, for example, participants may...
Crowdsourcing systems which utilize the human intelligence to solve complex tasks have gained considerable interest and adoption in recent years. However, majority of existing crowdsourcing rely on central servers, are subject weaknesses traditional trust-based model, such as single point failure. They also vulnerable distributed denial service (DDoS) Sybil attacks due malicious users involvement. In addition, high fees from platform may hinder development crowdsourcing. How address these...
ChatGPT is a recent chatbot service released by OpenAI and receiving increasing attention over the past few months. While evaluations of various aspects have been done, its robustness, i.e., performance to unexpected inputs, still unclear public. Robustness particular concern in responsible AI, especially for safety-critical applications. In this paper, we conduct thorough evaluation robustness from adversarial out-of-distribution (OOD) perspective. To do so, employ AdvGLUE ANLI benchmarks...
There are currently dozens of freely available tools to combat phishing and other web-based scams, many which web browser extensions that warn users when they browsing a suspected site. We developed an automated test bed for testing antiphishing tools. used 200 verified URLs from two sources 516 legitimate the effectiveness 10 popular anti-phishing Only one tool was able consistently identify more than 90% correctly; however, it also incorrectly identified 42% as phish. The performance...
Load frequency control (LFC) is widely employed to keep smart grids stable and secure. This paper proposes an adaptive resilient LFC scheme for sub-systems of under denial-of-service (DoS) attacks with energy constraint. Firstly, a triggering communication introduced, where the condition includes uncertainty item induced by DoS attacks. Secondly, event-triggering proposed further reduce burden defeat attacks, parameter can be dynamically adjusted. Third, stability criterion derived...
Medication quality and safety are crucial to the health of public. Responding urgent need for medication information provenance anti-counterfeiting, this study proposes a blockchain based method storage, inquiry, anti-counterfeiting along supply chain. Leveraging features decentralization, tamper-proof, traceability, participative node maintenance technology, proposed can assure transparency openness chains. An access control policy model on smart contract is designed prevent from being...
In recent years, considerable progress has been made on improving the interpretability of machine learning models. This is essential, as complex deep models with millions parameters produce state art performance, but it can be nearly impossible to explain their predictions. While various explainability techniques have achieved impressive results, all them assume each data instance independent and identically distributed (iid). excludes relational models, such Statistical Relational Learning...
Large language models (LLMs) are complex artificial intelligence systems capable of understanding, generating and translating human language. They learn patterns by analyzing large amounts text data, allowing them to perform writing, conversation, summarizing other tasks. When LLMs process generate there is a risk leaking sensitive information, which may threaten data privacy. This paper concentrates on elucidating the privacy concerns associated with foster comprehensive understanding....
Penetration testing is an effective way to test and evaluate cybersecurity by simulating a cyberattack. However, the traditional methods deeply rely on domain expert knowledge, which requires prohibitive labor time costs. Autonomous penetration more efficient intelligent solve this problem. In paper, we model as Markov decision process problem use reinforcement learning technology for autonomous in large scale networks. We propose improved deep Q-network (DQN) named NDSPI-DQN address sparse...
Optimizing the combustion efficiency of a thermal power generating unit (TPGU) is highly challenging and critical task in energy industry. We develop new data-driven AI system, namely DeepThermal, to optimize control strategy for TPGUs. At its core, model-based offline reinforcement learning (RL) framework, called MORE, which leverages historical operational data TGPU solve complex constrained Markov decision process problem via purely training. In we first learn simulator from dataset. The...
Mini-programs, which are programs running inside mobile super apps such as WeChat, often have access to privacy-sensitive information, location data and phone numbers, through APUs provided by the apps. This poses a risk of privacy sensitive leaks, either accidentally from carelessly programmed mini-programs or intentionally malicious ones. To address this concern, it is crucial track flow in for human analysis automated tools. Although existing taint techniques been widely studied, they...
Large Language Models (LLMs), such as ChatGPT and Bard, have revolutionized natural language understanding generation. They possess deep comprehension, human-like text generation capabilities, contextual awareness, robust problem-solving skills, making them invaluable in various domains (e.g., search engines, customer support, translation). In the meantime, LLMs also gained traction security community, revealing vulnerabilities showcasing their potential security-related tasks. This paper...
Recently, emerging SDN-based VANET (i.e., vehicular ad hoc network based on software-defined networking) enables management to be programmable and flexible. It introduces SDN controllers maintain network-wide resources applications program configurations through arbitrarily accessing via the northbound interface (NBI). However, this brings with it security issues NBI, such as resource exposure configuration manipulation. Most of existing works employed permission systems restrict access;...
A miniapp is a full-fledged app that executed inside mobile super such as WeChat or SnapChat. Being mini by nature, it often has to communicate with other miniapps accomplish complicated tasks. However, unlike web uses network domains (i.e., IP addresses) navigate between different apps, unique global appId assigned the miniapps. Unfortunately, any missing checks of sender's in receiver can lead new type attacks we name cross-miniapp request forgery (CMRF). In addition demystifying root...
Purpose Machine learning (ML) models were constructed according to non-contrast computed tomography (NCCT) images as well clinical and laboratory information assess risk stratification for the occurrence of hemorrhagic transformation (HT) in acute ischemic stroke (AIS) patients. Methods A retrospective cohort was with 180 AIS patients who diagnosed at two centers between January 2019 October 2023 followed HT outcomes. Patients analyzed factors developing HT, infarct texture features...