A5065641592- Network Security and Intrusion Detection
- Mineral Processing and Grinding
- Video Coding and Compression Technologies
- Advanced Malware Detection Techniques
- Chaos-based Image/Signal Encryption
- Security and Verification in Computing
- Information and Cyber Security
- Adversarial Robustness in Machine Learning
- Tunneling and Rock Mechanics
- Internet Traffic Analysis and Secure E-voting
- Domain Adaptation and Few-Shot Learning
- Thermography and Photoacoustic Techniques
- International Business and FDI
- Energy Efficiency and Management
- Cloud Data Security Solutions
- Smart Grid Security and Resilience
- Blind Source Separation Techniques
- Optimization and Search Problems
- Reliability and Maintenance Optimization
- Occupational Health and Safety Research
- Guidance and Control Systems
- Sparse and Compressive Sensing Techniques
- Quality and Supply Management
- Ga2O3 and related materials
- Elevator Systems and Control
Shanghai University
2025
Xidian University
2022-2025
Purdue University West Lafayette
2024-2025
Guilin University of Electronic Technology
2014-2024
Shanghai Jiao Tong University
2018-2022
Yancheng Institute of Technology
2020
Communication University of China
2018
Huaiyin Institute of Technology
2016
Network intrusion detection systems based on deep learning are gaining significant traction in cyber security due to their high prediction accuracy and strong adaptability evolving threats. However, a serious drawback is vulnerability evasion attacks that rely adversarial examples. To provide robustness guarantees for neural networks against any possible perturbations, certified defenses perturbations within l p -bounded region around the input being increasingly explored. Unfortunately,...
Internet of Things (IoT) cyber threats, exemplified by jackware and crypto mining, underscore the vulnerability IoT devices. Due to multi-step nature many attacks, early detection is vital for a swift response preventing malware propagation. However, accurately detecting early-stage attacks challenging, as attackers employ stealthy, zero-day, or adversarial machine learning evade detection. To enhance security, we propose ARIoTEDef, an Adversarially Robust Early Defense system, which...
Abstract Existing perception and identification methods for coal–rock interfaces are generally based on various cutting signals during shearer mining, but they cannot achieve pre-perception. Further, some pre-identification significantly affected by the mining environment accurate identification. In this study, a universal method is proposed to pre-perception recognition of interfaces. Using interface test-bed, active excitation infrared images were tested with different time, distance,...
Deep neural networks excel at solving intuitive tasks that are hard to describe formally, such as classification, but easily deceived by maliciously crafted samples, leading misclassification. Recently, it has been observed the attack-specific robustness of models obtained through adversarial training does not generalize well novel or unseen attacks. While data augmentation mixup in input space shown improve generalization and models, there limited research progress on latent space....
The rise of 4K and 8K techniques has led to the growth video data streaming. Consequently, greater challenges encryption efficiency information leakage facing selective (SE) makes it necessary reduce ratio as much possible. In this paper, we design a SE scheme for H.264/AVC which achieves trade-off between low high safety, both cryptographic attack sketch point view. As starting point, propose novel calculation model, called block weight model that can take advantage reference structure...
In the rapidly developing product–service paradigm, a multi-unit manufacturing system will be leased many times during life cycle. Besides maintenance programming for single leasing period, an upgrade policy is essential to achieve life-cycle profit optimization sequential periods. This article presents framework integrating opportunistic scheduling and decision throughout its cycle based on real-time prognostic method. With machine-level prognosis updating, system-level predictive...
Abstract To achieve efficient and low energy consumption mining under different cutting depths of a shearer, multiparameter coupling optimization method for the shearer based on three‐dimensional (3D) characterization coal wall was proposed. First, seven‐axis absolute articulated arm measuring machine used to obtain 3D point cloud data wall, then surface reconstructed by using segmentation, filtering, stitching processing, thereby obtaining average thickness areas. Second, through quadratic...
This paper presents a new dynamic recognition model to identify the coal-rock interface, which overcame problems of low identification accuracy and hysteresis quality. The proposed designed on basis cutting acoustic emission fuzzy D-S theory is able confirm ratio between coal rock. For this purpose, seven specimens were poured with different ratios. Thus, feature signals tested in process given specimens. Furthermore, by using wavelet packet decomposition energy reconstruction, distributions...
Abstract Using online information resources to build knowledge bases provide answering services would help auto companies or third-party platforms gain competitive advantages. Therefore, a construction plan of automobile maintenance expert system based on graph was proposed by integrated modelling. In terms the entity recognition algorithm, BM LSTM (Boyer-Moore Long Short-Term Memory) algorithm integrating hidden Markov model, Conditional Random Field (CRF), Bi-directional Memory (BiLSTM),...
At present, almost all the provable data possession schemes require a large number of exponent calculations.In this paper new scheme based on homomorphy NTRU algorithm is proposed.The includes modules Key Generation, Tag Challenge, Proof and Verification.This requires only plus, minus, multiplication modular computation.The correctness security are analyzed result shows that reliable.Compared with other existing schemes, higher security, less computation, storage communication cost.
As the number of interconnected IoT devices continues to skyrocket, so does attack surface for malicious actors. loT are often resource-constrained and deployed in diverse challenging environments (e.g., critical infras-tructures), making them susceptible various vulnerabilities attacks. Additionally, attacks multi-step, which complicates detection mitigation efforts. Without proactive measures, these can be exploited devastating effects. In this paper, we propose a defense framework systems...
The following topics are dealt with: finite element analysis; design engineering; aircraft; vibrations; optimisation; fasteners; beams (structures); structural aircraft control; cutting.