A5100459326- Natural Language Processing Techniques
- Topic Modeling
- Metaheuristic Optimization Algorithms Research
- Evolutionary Algorithms and Applications
- Advanced Multi-Objective Optimization Algorithms
- Semantic Web and Ontologies
- Network Security and Intrusion Detection
- Parallel Computing and Optimization Techniques
- Music Technology and Sound Studies
- Advanced Image and Video Retrieval Techniques
- Video Surveillance and Tracking Methods
- Anomaly Detection Techniques and Applications
- IoT and Edge/Fog Computing
- Advanced Data Storage Technologies
- Multimodal Machine Learning Applications
- Robotics and Sensor-Based Localization
- Advanced Image Processing Techniques
- Caching and Content Delivery
- Opportunistic and Delay-Tolerant Networks
- Music and Audio Processing
- Image and Signal Denoising Methods
- Advanced Data Compression Techniques
- Visual Attention and Saliency Detection
- Human Pose and Action Recognition
- Advanced Malware Detection Techniques
PLA Information Engineering University
2023-2025
Xi'an University of Architecture and Technology
2025
Shanghai Jiao Tong University
2015-2024
Institute of Immunology
2024
Ruijin Hospital
2024
Tongren Hospital
2024
East China Normal University
2022-2024
Huazhong University of Science and Technology
2024
Tongji University
2016-2024
Alibaba Group (China)
2021-2024
Distant supervision assumes that any sentence containing the same entity pairs reflects identical relationships. Previous works of distantly supervised relation extraction (DSRE) task generally focus on sentence-level or bag-level de-noising techniques independently, neglecting explicit interaction with cross levels. In this paper, we propose a hierarchical contrastive learning Framework for Distantly Supervised (HiCLRE) to reduce noisy sentences, which integrate global structural...
Edge caching has emerged as a promising technique against latency caused by explosive growth of mobile data traffic through popular contents at the edge networks. However, dynamically changing content popularity nature and limited capacity make it challenging to design an effective scheme reduce latency. To solve this, learning-based hierarchical (LHEC) is proposed in this work. We first propose novel deep learning architecture, namely Stacked Autoencoder-Long Short Term Memory Network...
With the rapid development of artificial intelligence, application this new technology to music generation has attracted more attention and achieved gratifying results. This study proposes a method for combining transformer deep-learning model with generative adversarial networks (GANs) explore competitive algorithm. The idea text in natural language processing (NLP) was used reference, unique loss function designed model. training process solves problem nondifferentiable gradient generating...
Abstract Background normalization is one of the hot fields in underwater acoustic signal processing. Building on conventional two-dimensional background methods, this paper extends these techniques to three dimensions, tailored for high-dimensional characteristics continuous active sonar detection data. The three-dimensional algorithm sets protection windows and reference azimuth, range, Doppler directions, then uses data from estimate power test cell. This analyzes verifies performance...
Plastic deformation plays a critical role in improving the formability and mechanical performance of metals. In titanium, zirconium, hafnium, slip twinning are primary mechanisms, while phase transformations also contribute significantly. This article summarizes recent advances on hexagonal‐close‐packed (HCP) to face‐centered‐cubic (FCC) transition these metals, focusing mechanism HCP‐to‐FCC transition‐induced plastic its orientation relationships (ORs), mechanism, influencing factors. The...
In recent years, many methods have been put forward to improve the image matching for different viewpoint images. However, these are still not able achieve stable results, especially when large variation in view occurs. this paper, an method based on affine transformation of local areas is proposed. First, regions extracted from reference and test image, transformed circular according second-order moment. Then, scale invariant features detected matched regions. Finally, we use epipolar...
In recent deep image compression neural networks, the entropy model plays a critical role in estimating prior distribution of encodings. Existing methods combine hyperprior with local context estimation function. This greatly limits their performance due to absence global vision. this work, we propose novel Global Reference Model for effectively leverage both and information, leading an enhanced rate. The proposed method scans decoded latents then finds most relevant latent assist current...
Nowadays, various methods are proposed to build effective anomaly-based Network Intrusion Detection System (NIDS). However, malicious packets extremely less than normal and this class imbalance problem will result in low performance of attack detection. In study, we have a new hybrid oversampling model using GAN improve detection NIDS. It contains three main steps: feature extraction by Information Gain PCA, data clustering DBSCAN generation WGAN-DIV. For evaluation, HTTP only datasets:...
One of the mandates No Child Left Behind Act is that states show adequate yearly progress in their English language learners’ (ELLs) acquisition proficiency. States are required to assess ELLs’ proficiency annually four domains (listening, reading, writing, and speaking) measure progress; they also report on a composite comprehension measure. Often clearest way effectively monitor students’ assessment results across grades same scale. In measurement terms, scores from tests all grade levels...
In this paper, we propose a light reflection based face anti-spoofing method named Aurora Guard (AG), which is fast, simple yet effective that has already been deployed in real-world systems serving for millions of users. Specifically, our first extracts the normal cues via analysis, and then uses an end-to-end trainable multi-task Convolutional Neural Network (CNN) to not only recover subjects' depth maps assist liveness classification, but also provide CAPTCHA checking mechanism regression...
Essays are a pivotal component of conventional exams; accurately, efficiently, and effectively grading them is significant challenge for educators. Automated essay scoring (AES) complex task that utilizes computer technology to assist teachers in scoring. Traditional AES techniques only focus on shallow linguistic features based the criteria, ignoring influence deep semantic features. The model neural networks (DNN) can eliminate need feature engineering achieve better accuracy. In addition,...
Large-scale particle swarm optimization (PSO) has long been a hot topic due to the following reasons: Swarm diversity preservation is still challenging for current PSO variants large-scale problems, resulting in difficulties balancing its exploration and exploitation. Furthermore, problems often introduce additional operators improve their ability preservation, leading increased algorithm complexity. To address these issues, this paper proposes dual-competition-based update strategy (DCS),...
Insider threat detection has been a challenging task over decades; existing approaches generally employ the traditional generative unsupervised learning methods to produce normal user behavior model and detect significant deviations as anomalies. However, such are insufficient in precision computational complexity. In this paper, we propose novel insider method, Image-based Threat Detector via Geometric Transformation (IGT), which converts anomaly into supervised image classification task,...