A5012385922- Cryptography and Data Security
- Privacy-Preserving Technologies in Data
- Network Security and Intrusion Detection
- Caching and Content Delivery
- Internet Traffic Analysis and Secure E-voting
- Peer-to-Peer Network Technologies
- RFID technology advancements
- Adversarial Robustness in Machine Learning
- Advanced Authentication Protocols Security
- Advanced Malware Detection Techniques
- Cloud Data Security Solutions
- User Authentication and Security Systems
- Anomaly Detection Techniques and Applications
- Advanced Data Storage Technologies
- Mobile Crowdsensing and Crowdsourcing
- Vehicular Ad Hoc Networks (VANETs)
- Opportunistic and Delay-Tolerant Networks
- Indoor and Outdoor Localization Technologies
- Traffic Prediction and Management Techniques
- IoT and Edge/Fog Computing
- Complex Network Analysis Techniques
- Stochastic Gradient Optimization Techniques
- Privacy, Security, and Data Protection
- Spam and Phishing Detection
- Blockchain Technology Applications and Security
University of Electronic Science and Technology of China
2016-2025
China Southern Power Grid (China)
2024-2025
Yangtze University
2025
University of Sheffield
2024
Energy Foundation
2024
Xi'an Jiaotong University
2024
Ghana Institute of Management and Public Administration
2024
First Research Institute of the Ministry of Public Security
2023
Chongqing University of Science and Technology
2023
Jiangsu University
2021
In recent years, artificial intelligence technologies have been widely used in computer vision, natural language processing, automatic driving, and other fields. However, systems are vulnerable to adversarial attacks, which limit the applications of (AI) key security Therefore, improving robustness AI against attacks has played an increasingly important role further development AI. This paper aims comprehensively summarize latest research progress on attack defense deep learning. According...
Heart failure is considered one of the leading cause death around world. The diagnosis heart a challenging task especially in under-developed and developing countries where there paucity human experts equipments. Hence, different researchers have developed intelligent systems for automated detection failure. However, most these methods are facing problem overfitting i.e. recently proposed improved accuracy on testing data while compromising training data. Consequently, constructed models...
Network slicing in a 6G environment is an important research area the current years. However, satisfying demands of network slice requests challenging task. Energy-efficient, secure, and Quality Service (QoS) aware since slices must share fewer amounts resources. Further, implementing secure for software-defined networks (SDNs)/network function virtualization (NFV) crucial. In this article, we tackle issues, such as security, QoS, resource consumption issues through load balancing mechanisms...
Real-world industrial systems contain a large number of interconnected sensors that generate significant amount time series data during system operation. Performing anomaly detection on these multivariate can timely find faults, prevent malicious attacks, and ensure safe reliable However, the rarity abnormal instances leads to lack labeled data, so supervised machine learning methods are not applicable. Furthermore, most current techniques do take full advantage spatial temporal dependencies...
Diagnosis of heart disease is a difficult job, and researchers have designed various intelligent diagnostic systems for improved diagnosis. However, low prediction accuracy still problem in these systems. For better risk accuracy, we propose feature selection method that uses floating window with adaptive size elimination (FWAFE). After the elimination, two kinds classification frameworks are utilized, i.e., artificial neural network (ANN) deep (DNN). Thus, types hybrid proposed this paper,...
Congestion road condition is an important factor that must be considered in urban traffic path planning, while most planning algorithms only consider the distance factor, which not suitable for current complex congestion condition. In order to solve above problems, this article proposes a dynamic method based on improved ant colony algorithm congested traffic. The quantifies main attributes of length, number lanes, incoming and outgoing flow, introduces used replacing parameters particle...
Space information networks is network systems that can receive, transmit, and process spatial lively. It uses satellites, stratosphere airships, Unmanned Aerial Vehicles, other platforms as the carrier. supports high-dynamic, real-time broadband transmission of earth observations ultra-long-distance, long-delay reliable deep space exploration. The deeper integration, higher system's security concerns more likely SINs will be controlled destroyed in terms cybersecurity. How to integrate new...
A honeypot is a supplemented active defense system for network security. It traps attacks, records intrusion information about tools and activities of the hacking process, prevents attacks outbound compromised system. Integrated with other security solutions, can solve many traditional dilemmas. We expatiate key components data capture control in honeypot, give classification honeypots according to goals application goals. review technical progress contribution production research honeypots....
The Vehicular Ad Hoc Network (VANET) plays a vital role in the development of smart cities, especially ensuring vehicles' safety on roads. However, VANET wireless-based networks face some challenges such as security, stability, communication, and reliability. To resolve these issues, we propose fuzzy cluster head selection scheme Cognitive Radio (CR) VANET, which uses CR technology for spectrum sensing algorithm. In this technology, free spectrums primary user are utilized by secondary users...
In recent years, due to the frequent occurrence of network intrusions, more and researchers have begun focus on intrusion detection. However, it is still a challenge detect unknown attacks. Currently, there are two main methods attack detection: clustering honeypot. But they unsolved problems such as difficulty in collecting samples failure time. Zero-Shot learning proposed deal with problem this article, which can recognize attacks by mapping relations between feature space semantic (such...
VANET is the spontaneous evolving creation of a wireless network, and clustering in these networks challenging task due to rapidly changing topology frequent disconnection networks. The cluster head (CH) stability plays prominent role robustness scalability network. stable CH ensures minimum intra- intercluster communication, thereby reducing overhead. These challenges lead authors search for selection method based on weighted amalgamation four metrics: befit factor, community neighborhood,...
Due to the different load demands in areas of city, effect trunk line planning is difficult ensure effective utilization electric energy. Therefore, a study on urban distribution network based high-resolution remote sensing images proposed. Considering objective demand load, divided into three-layer grid model according three voltage levels network: high voltage, medium and low combined with image information. In specific division stage, IBIWO algorithm introduced. After initial architecture...
We introduce X-Dyna, a novel zero-shot, diffusion-based pipeline for animating single human image using facial expressions and body movements derived from driving video, that generates realistic, context-aware dynamics both the subject surrounding environment. Building on prior approaches centered pose control, X-Dyna addresses key shortcomings causing loss of dynamic details, enhancing lifelike qualities video animations. At core our approach is Dynamics-Adapter, lightweight module...
Abstract Dye-sensitised solar cells (DSSCs) are of great potential for science development to solve the problem energy crisis. The sensitizers (dyes) DSSCs explored and discussed in this project. Computational chemistry investigations based on Gaussian 09 carried out determine computational data dye molecules deduce their properties. Light absorptions, molecular orbitals, levels dyes specifically discussed. Molecular structures modified diversely achieve better features. new molecule with...
Current black-box adversarial attacks have demonstrated significant efficacy in creating texts against natural language processing models, exposing potential robustness vulnerabilities of these models. However, present attack techniques exhibit inefficiency due to their failure account for the query counts needed text generation process, causing a disparity between existing methodology and practical scenario. To this end, work proposes query-efficient hard-label method called QEAttack, which...
Abstract Migration-based microseismic event location methods based on signal coherency, such as semblance and crosscorrelation, have been extensively used for hydro-fracturing monitoring of unconventional resources because they do not depend first pick accuracy. However, to a great degree, these that are solely either imaging function or crosscorrelation can be significantly influenced by low signal-to-noise ratio (SNR) circumstances. Thus, in order enhance the precision, we presented...
Recent advancements in 2D and multimodal models have achieved remarkable success by leveraging large-scale training on extensive datasets. However, extending these achievements to enable free-form interactions high-level semantic operations with complex 3D/4D scenes remains challenging. This difficulty stems from the limited availability of large-scale, annotated or multi-view datasets, which are crucial for generalizable vision language tasks such as open-vocabulary prompt-based...