A5101921959- Network Security and Intrusion Detection
- Anomaly Detection Techniques and Applications
- Catalytic Processes in Materials Science
- Internet Traffic Analysis and Secure E-voting
- Earthquake Detection and Analysis
- Plasma Applications and Diagnostics
- earthquake and tectonic studies
- Advanced Text Analysis Techniques
- Machine Learning in Bioinformatics
- Bioinformatics and Genomic Networks
- Plant Gene Expression Analysis
- Protein Structure and Dynamics
- Adversarial Robustness in Machine Learning
- Computational Drug Discovery Methods
- Advanced Malware Detection Techniques
- Microplastics and Plastic Pollution
- Sentiment Analysis and Opinion Mining
- Electrocatalysts for Energy Conversion
- Topic Modeling
- Industrial Gas Emission Control
- Recycling and Waste Management Techniques
- Web Data Mining and Analysis
- Ionosphere and magnetosphere dynamics
- Plant Molecular Biology Research
- Natural Language Processing Techniques
Tsinghua University
2019-2024
Applied Pulsed Power (United States)
2024
State Key Laboratory of Pulsed Power Laser Technology
2024
Aalto University
2024
Zhejiang University
2022-2024
Advanced Laser Technology (United Kingdom)
2024
Institute of Psychology, Chinese Academy of Sciences
2024
Sichuan University
2021-2024
Nanjing University of Posts and Telecommunications
2023
Jilin University
2023
Machine learning (ML), especially deep (DL) techniques have been increasingly used in anomaly-based network intrusion detection systems (NIDS). However, ML/DL has shown to be extremely vulnerable adversarial attacks, such security-sensitive systems. Many attacks proposed evaluate the robustness of ML-based NIDSs. Unfortunately, existing mostly focused on feature-space and/or white-box which make impractical assumptions real-world scenarios, leaving study practical gray/black-box largely...
Unsupervised Deep Learning (DL) techniques have been widely used in various security-related anomaly detection applications, owing to the great promise of being able detect unforeseen threats and superior performance provided by Neural Networks (DNN). However, lack interpretability creates key barriers adoption DL models practice. Unfortunately, existing interpretation approaches are proposed for supervised learning and/or non-security domains, which unadaptable unsupervised fail satisfy...
Pd/CuO-Ni(OH) 2 /C showed superior electrocatalytic performance for EOR in comparison with PdCuNi alloy/C and Pd/C. In-situ FTIR studies have shown that CuO Ni(OH) effectively inhibited the adsorption of CO increased breaking rate C-C.
Protein-protein interactions (PPIs) dominate intracellular molecules to perform a series of tasks such as transcriptional regulation, information transduction, and drug signalling. The traditional wet experiment method obtain PPIs is costly time-consuming.
Concept drift is one of the most frustrating challenges for learning-based security applications built on closeworld assumption identical distribution between training and deployment.Anomaly detection, important tasks in domains, instead immune to abnormal behavior due without any data (known as zero-positive), which however comes at cost more severe impacts when normality shifts.However, existing studies mainly focus concept behaviour and/or supervised learning, leaving shift zero-positive...
Recent years have seen tremendous success in the design of novel drug molecules through deep generative models. Nevertheless, existing methods only generate drug-like molecules, which require additional structural optimization to be developed into actual drugs. In this study, a learning method for generating target-specific ligands was proposed. This is useful when dataset limited. Deep can extract and learn features (representations) data-driven way with little or no human participation....
Herein, we employ a galvanic replacement approach to create atomically dispersed Au on degradable zero-valent Cu nanocubes for tumor treatments female mice. Controlling the addition of precursor HAuCl4 allows fabrication different atomic ratios AuxCuy. X-ray absorption near edge spectra indicates that and are predominant oxidation states zero valence. This suggests charges remain unchanged after replacement. Specifically, Au0.02Cu0.98 composition reveals enhanced •OH generation following O2...
At present, the research on UAV mission planning mainly focuses traditional $\text{A}^{\star}$ algorithm, artificial potential field method, $\text{D}^{\star}$ algorithm and some early popular intelligent algorithms, such as ant colony particle swarm genetic so on. To solve method in unmanned aerial vehicle (UAV) route for obstacles target distance too close without gravitational repulsion only lead to unreachable, repulsive force of net is zero, not point problems, put forward an improved...
This communication investigates possible anomalies in the lithosphere atmosphere and ionosphere on occasion of ML=3.3 earthquake that occurred 1st January 2023 close to Guidonia Montecelio (Rome, Italy). followed another event 23 December 2022 magnitude ML=3.1 with a very epicentre (distance less than 1km). Seismological investigations clearly show an acceleration seismicity last six months area. Two solutions fitting time failure power law Cumulative Benioff strain curve are more...
Deep learning (DL) performs well in many traffic analysis tasks.Nevertheless, the vulnerability of deep weakens real-world performance these analyzers (e.g., suffering from evasion attack).Many studies recent years focused on robustness certification for DL-based models.But existing methods perform far perfectly domain.In this paper, we try to match three attributes systems at same time: (1) highly heterogeneous features, (2) varied model designs, (3) adversarial operating...
Accurate land cover mapping is challenging in Southeast Asia where cloud coverage prevalent and landscape heterogenous. Object-based mapping, multi-temporal images combined use of optical microwave data, provide abundant features spectral, spatial, temporal, geometric polarimetric dimensions. And random forest usually employed due to robustness efficiency handling high-dimensional noisy data. This study assesses whether feature selection ensemble analysis, which are rarely adopted, yield...
Abstract Background Protein-protein interaction (PPI) is very important for many biochemical processes. Therefore, accurate prediction of PPI can help us better understand the role proteins in Although there are methods to predict biology, they time-consuming and lack accuracy, so it necessary build an efficiently accurately computational model field prediction. Results We present a novel sequence-based approach called DCSE (Double-Channel-Siamese-Ensemble) potential PPI. In encoding layer,...
Background: Biomedical named entity recognition is one of the important tasks biomedical literature mining. With development natural language processing technology, many deep learning models are used to extract valuable information from literature, which promotes effective BioNER models. However, for specialized domains with diverse and complex contexts a richer set semantically related types (e.g., drug molecules, targets, pathways, etc., in domain), whether dependencies these drugs,...