A5091829195- Machine Fault Diagnosis Techniques
- Fault Detection and Control Systems
- Solid State Laser Technologies
- Advanced Image and Video Retrieval Techniques
- Advanced Neural Network Applications
- Glass properties and applications
- Luminescence Properties of Advanced Materials
- Robotics and Sensor-Based Localization
- Privacy-Preserving Technologies in Data
- Robot Manipulation and Learning
- Web Data Mining and Analysis
- Engineering Diagnostics and Reliability
- Laser Applications in Dentistry and Medicine
- Domain Adaptation and Few-Shot Learning
- Internet Traffic Analysis and Secure E-voting
- Thermography and Photoacoustic Techniques
- Coal Properties and Utilization
- Laser Material Processing Techniques
- Usability and User Interface Design
- Interactive and Immersive Displays
- Human Pose and Action Recognition
- Apelin-related biomedical research
- Urban Transport and Accessibility
- Metaheuristic Optimization Algorithms Research
- Rock Mechanics and Modeling
Nanjing Hydraulic Research Institute
2025
PLA Army Engineering University
2022-2025
Shangqiu First People's Hospital
2022-2025
Wuhan University of Technology
2024
Shijiazhuang University
2022-2024
Xi'an Jiaotong University
2024
Ningbo University
2024
Shanghai Jiao Tong University
2023-2024
Southwest Forestry University
2024
Shanghai Ninth People's Hospital
2023
Given a single scene image, this paper proposes method of Category-level 6D Object Pose and Size Estimation (COPSE) from the point cloud target object, without external real pose-annotated training data. Specifically, beyond visual cues in RGB images, we rely on shape information predominately depth (D) channel. The key idea is to explore alignment each instance against its corresponding category-level template shape, symmetric correspondence object category for estimating coarse 3D shape....
Abstract Transformer is extensively employed in natural language processing, and computer vision (CV), with the self-attention structure. Due to its outstanding long-range dependency modeling parallel computing capability, some leading researchers have recently attempted apply intelligent fault diagnosis tasks for mechanical equipment, achieved remarkable results. Physical phenomena such as changes vibration, sound, heat play a crucial role research of equipment diagnosis, which directly...
To address the challenges of mode aliasing and endpoint effects inherent in current diesel engine vibration signal decomposition methods, to enhance reliability subsequent anomaly detection fault diagnosis, this paper proposes a novel approach by integrating successive variational (SVMD) with multivariate (MVMD), resulting Successive Multivariate Variational Mode Decomposition (SMVMD) method tailored for signals. The technical principles SMVMD are elucidated, its feasibility effectiveness...
Underground reservoirs are a key technology for storing mine-impacted water resources, and the long-term stability of their coal pillar dams in high-stress environments is critical. The safety such closely related to creep seepage phenomena. To better illustrate this phenomenon, internal expansion coefficients porosity blocking proposed study characterize how affects evolution permeability water-bearing samples. A novel model developed capture interaction between matrix fractures influence...
Regarding the difficulty of extracting fault information in faulty status UAV (unmanned aerial vehicle) engines and high time cost large data requirement existing deep learning diagnosis algorithms with many training parameters, this paper, a small-sample transfer algorithm is proposed. First, vibration signals under engine are converted into two-dimensional time-frequency map by multiple simultaneous squeezing S-transform (MSSST), which reduces randomness manually extracted features....
In response to the problem that nonlinear and non-stationary rolling bearing fault signals are easily disturbed by noise, which leads difficulty of feature extraction, take full advantage superiority variational mode decomposition (VMD) in noise reduction, maximum correlation kurtosis deconvolution (MCKD) highlighting continuous pulses masked a method based on sparrow search algorithm (SSA), VMD, MCKD is proposed, namely, SSA–VM–MCKD, for faint extraction. To improve extraction effect, uses...
The diesel engine, as the main power source of equipment, faces practical problems in maintenance process, such difficulty fault location and a lack preventive techniques. Currently, breakdown cyclical are means support after engine failure, but these methods require professional personnel to carry out manual diagnosis, which is time-consuming. Prognostics health management (PHM), new technology field equipment support, has significant advantages improving reliability safety, enhancing...
Malignant mesothelioma is a type of infrequent tumor that substantially related to asbestos exposure and has terrible prognosis. We tried produce fibroblast differentiation-related gene set for creating novel classification prognostic prediction model MESO. Three databases, including NCBI-GEO, TCGA, MET-500, separately provide single-cell RNA sequencing data, bulk profiles MESO, information on bone metastatic tumors. Dimensionality reduction clustering analysis were leveraged acquire...
Network traffic anomaly detection is an important research content in the field of network and security management. By analyzing traffic, health environment can be intuitively evaluated. In particular, provides practical effective guidance for identification classification anomaly. This paper proposes a method based on wavelet analysis pcap files contain two different delay injections. The effectively extract information from signal suitable Firstly, used to waveform features, then support...
To analyze the role of PD-1/PD-L1 signaling pathway in regulating T cell activation and secretion proinflammatory factors atrial fibrillation.
A small-scale delivery medium for CO2 laser energy with stable performance, flexibility, and high-strength is crucial in extreme processing environments, especially minimally invasive surgery high-humidity, twisty narrow channels. Here, flexible robust multimaterial infrared fibers made of selenium-based chalcogenide glasses thermoplastic polymer were developed a low loss 7.18 dB/m at 10.6 μm. The resulting capable stably delivering single-mode 0.42 W average power. Moreover, to achieve...
Large language models (LLMs) with hundreds of billions parameters have sparked a new wave exciting AI applications. However, they are computationally expensive at inference time. Sparsity is natural approach to reduce this cost, but existing methods either require costly retraining, forgo LLM's in-context learning ability, or do not yield wall-clock time speedup on modern hardware. We hypothesize that contextual sparsity, which small, input-dependent sets attention heads and MLP...
The process of developing a mobile application typically starts with the ideation and conceptualization its user interface. This concept is then translated into set mock-ups to help determine how well interface embodies intended features app. After creation developers translate it an app that runs in device. In this paper we propose approach, called GUIGLE, aims facilitate conceptualizing through GUI search. GUIGLE indexes images metadata extracted using automated dynamic analysis on large...
Primary squamous cell carcinoma of the thyroid (PSCCT) is a rare malignant tumor. The incidence rate PSCCT less than 1%. However, diagnosis and treatment are limited. Surgical resection considered to be one few effective intervention methods. In this article, we reported case taking tyrosine kinase inhibitors (TKIs) combined with immune checkpoint (ICIs) for PSCCT.An 80-year-old male was admitted our hospital dyspnea, cough, wheezing, hoarseness giant mass. He underwent bronchoscopy tracheal...
This paper tries to introduce a new intelligent method for the early fault diagnosis of diesel engines. Firstly, infrared thermography (IRT) is introduced into engine condition monitoring, then images engines in four health states, such as normal condition, single-cylinder misfire, multi-cylinder misfire and air filter blockage, are collected region interest (ROI) extracted. Next, conditional generative adversarial network (CGAN) deployed perform data augmentation on image datasets. Then,...
Abstract The multi-strategy improved sparrow search algorithm (MSISSA) is proposed to address the problems that (SSA) not rich in population diversity, and prone fall into local optimality poor accuracy solving multi-dimensional functions. Firstly, Cat mapping used initialize SSA population. Secondly, an elite reverse learning strategy introduced increase diversity improve global ability of SSA. Then, number discoverers aware-at-risk sparrows are dynamically adjusted by improving scaling...
Because of its efficient service and relatively low investment requirements, bus rapid transit (BRT) is growing in popularity many cities around the world. The development BRT nine China, including design, implementation, operation, management systems, summarized. Substantial data, infrastructure performance, passenger flow, have been collected from systems cities. cities’ population, size, importance, social economical backgrounds also provided. On basis five modes are summarized compared....
The ability to jointly learn from multiple modalities, such as text, audio, and visual data, is a defining feature of intelligent systems. While there have been promising advances in designing neural networks harness multimodal the enormous success data augmentation currently remains limited single-modality tasks like image classification. Indeed, it particularly difficult augment each modality while preserving overall semantic structure data; for example, caption may no longer be good...
Large language models(LLMs) have sparked a new wave of exciting AI applications. Hosting these models at scale requires significant memory resources. One crucial bottleneck for the deployment stems from context window. It is commonly recognized that model weights are hungry; however, size key-value embedding stored during generation process (KV cache) can easily surpass size. The enormous KV cache puts constraints on inference batch size, which high throughput workload. Inspired by an...