- Vibration and Dynamic Analysis
- Advanced Measurement and Detection Methods
- Structural Health Monitoring Techniques
- Peatlands and Wetlands Ecology
- Machine Learning and ELM
- Bladed Disk Vibration Dynamics
- Advanced Sensor and Energy Harvesting Materials
- Advanced Neural Network Applications
- Endometriosis Research and Treatment
- Advanced Image and Video Retrieval Techniques
- Machine Learning and Algorithms
- Optimization and Search Problems
- Structural Response to Dynamic Loads
- Advanced machining processes and optimization
- Advanced Materials and Mechanics
- Robotic Path Planning Algorithms
- Biometric Identification and Security
- Face recognition and analysis
- Transportation Safety and Impact Analysis
- Landslides and related hazards
- Advanced Algorithms and Applications
- Face and Expression Recognition
- Generative Adversarial Networks and Image Synthesis
- Constructed Wetlands for Wastewater Treatment
- Domain Adaptation and Few-Shot Learning
Liaoning Technical University
2024-2025
Aarhus University
2025
Shandong University
2010-2025
Nanjing University of Aeronautics and Astronautics
2022-2025
Peking University
2020-2025
Peking University Third Hospital
2020-2025
China University of Geosciences
2022-2024
North University of China
2023-2024
Jinzhou Medical University
2024
Ningbo University
2024
Abstract Automatic deformation forecast and warning of a catastrophic landslide can effectively avoid significant casualties economic losses. However, currently it has not come to comprehensive framework covering all the stages landslide. Moreover, prediction possesses high error false alarm rates. This work suggests novel integrated by coupling machine learning physical models. The relatively accurately predict from creeping critical sliding features 4 advantages. (a) indices are...
Deep neural network are one of the most powerful model for machine learning, which can learn underlying patterns automatically from a large amount data. So it be extensively used in more and Internet-of-Things (IoT) applications. However, training deep models is difficult, suffering overfitting gradient vanishing problem. Besides, parameters multiplication operations make impractical learning to directly execute on target hardware. In this paper, we propose method gradually pruning weakly...
Many factors influence the connection states between nodes of wireless sensor networks, such as physical distance, and network load, making network's edge length dynamic in abundant scenarios. This property makes essentially form a graph with stochastic lengths. In this paper, we study shortest path problem on directional lengths, using reinforcement learning algorithms. regard each random variable following unknown probability distribution aim to find graph. We evaluate performance...
While numerous studies have highlighted the stigma experienced by women facing infertility, there is limited research that delves into specific types of stigmas they encounter. Despite extensive discussions surrounding Stigma Management Communication (SMC) theory and its application in managing across various contexts, utilisation understanding addressing with infertility has been largely overlooked. This study seeks to bridge this gap identifying forms Chinese experiencing encounter...
In modern marine warfare, unmanned underwater vehicles (UUVs) have fast and efficient attack capabilities. However, existing research on UUV strategies is relatively limited, often ignoring the requirement for effective allocation of different strategic value areas, which restricts its performance in combat environment. To this end, paper proposes an innovative task saturation strategy. The strategy first divides area according to distribution density enemy UUVs, then reasonably allocates...
Abstract As missile-borne electronics face growing severe high overload environments, the matching protective structure also meets higher challenges. Thus, a laminated configuration with varying wave impedance is fabricated using appropriate metallic materials to facilitate energy absorption buffering. Through an examination of stress propagation within structure, model formulated. This model, in conjunction assessment filtering properties guides preparation Al-Fe heterogeneous for dynamic...
Objective: To compare the efficacy of dienogest (DNG) and levonorgestrel-releasing intrauterine system (LNG-IUS) in treatment intrinsic extrinsic subtypes adenomyosis. Methods: Totally 232 patients were enrolled study who diagnosed as adenomyosis by ultrasound or pelvic magnetic resonance imaging (MRI), classified into according to different locations lesions MRI, treated with DNG (DNG group) LNG-IUS (LNG-IUS Peking University Third Hospital from July 2019 December 2023. Clinical data...
<title>Abstract</title> Understanding phosphorus (P) transformation dynamics during peatland rewetting is crucial for developing effective management strategies, supporting ecological restoration initiatives and mitigating potential environmental risks. This incubation study explored the temporal variations in P peatlands under different land uses (cut grass, grazing, unmanaged) along with risk of leaching by simulating conditions four months at varying temperatures (10 20°C). Overall, only...
This study develops species-specific assessment models for carbon sink in marine aquaculture (CSMA) using provincial data from China’s coastal regions (2004–2023). Key findings are as follows: (1) Cumulative CSMA reached 46.3618 million tonnes, exhibiting three growth phases—initial fluctuations (2004–2008), rapid (2008–2015), and optimization maturation (2015–2023). (2) Species contributions were heterogeneous: shellfish dominated at 45%, followed by shrimp (24%), fish (15%), crab (11%),...
One kind of Deep Learning models-convolutional neural network, which can reduce the complexity network structure and number parameters to be determined through local receptive fields, weight sharing pooling operation has achieved state art results in image classification problems. But this model gradient diffusion problem, cause slow updating underlying during process training. To solve problem above make improvements, paper presents a convolutional based on principal component analysis...
Deep neural network has been one of the most powerful models in field machine learning, which acquired state-of-the-art results many tasks including image classification, object detection, text recognition, and so on. There have tricks to improve training generalization performance deep network, such as dropout, ReLU, batch normalization, etc. In this paper, we proposed a new basic element form networks, called learning automata competition unit (LCU). The LCU includes group general units...
Deep learning methods have got fantastic performance on lots of large-scale datasets for machine tasks, such as visual recognition and neural language processing. Most the progress deep in recent years lied supervised learning, which whole dataset with respect to a specific task should be well-prepared before training. However, real-world scenario, labeled data associated assigned classes are always gathered incrementally over time, since it is cumbersome work collect annotate training...
Objective To investigate diagnostic approaches for preoperative localization of secondary hyperparathyroidism, as well to give surgeons with precise parathyroid gland and imaging so that surgery can be performed safely. Methods The clinical data 710 patients hyperparathyroidism who underwent in our center from October 2009 2023 were retrospectively analyzed. changes calcium, phosphorus, hormone levels observed ascertain the anatomical location number glands. Results Among patients, 55 total...
Aero-engine casing is a kind of thin-walled rotary part for which serious deformation often occurs during its machining process. As force an important physical quantity associated with deformation, the utilization to control has been suggested. However, due complex characteristics aero-engine casing, obtaining stable and reliable can be quite difficult. To address this issue, paper proposes monitoring method via pre-support probabilistic decision model based on deep autoregressive neural...