A5100701473- Topic Modeling
- Advanced Measurement and Detection Methods
- Infrared Target Detection Methodologies
- Oil Spill Detection and Mitigation
- Bone fractures and treatments
- Explainable Artificial Intelligence (XAI)
- Remote-Sensing Image Classification
- Data Management and Algorithms
- Color Science and Applications
- Advanced Control and Stabilization in Aerospace Systems
- Advanced Chemical Sensor Technologies
- Dental Radiography and Imaging
- Advanced Steganography and Watermarking Techniques
- Stock Market Forecasting Methods
- Anatomy and Medical Technology
- Guidance and Control Systems
- Artificial Intelligence in Healthcare and Education
- Advanced Image Fusion Techniques
- Prosthetics and Rehabilitation Robotics
- Medical Imaging and Analysis
- Soft Robotics and Applications
- Simulation and Modeling Applications
- Sustainable Supply Chain Management
- Food composition and properties
- Biomedical and Engineering Education
Jiangnan University
2024
Southern University of Science and Technology
2024
National Research Center for Rehabilitation Technical Aids
2015-2023
Ministry of Civil Affairs
2020-2023
Hong Kong University of Science and Technology
2023
University of Hong Kong
2023
Nanjing University of Aeronautics and Astronautics
2022
Beijing Institute of Technology
2015-2021
China University of Mining and Technology
2021
Beijing Institute of Optoelectronic Technology
2019
The emergence of Large Language Models (LLMs), such as ChatGPT, has revolutionized general natural language preprocessing (NLP) tasks. However, their expertise in the financial domain lacks a comprehensive evaluation. To assess ability LLMs to solve NLP tasks, we present FinLMEval, framework for Financial Model Evaluation, comprising nine datasets designed evaluate performance models. This study compares fine-tuned auto-encoding models (BERT, RoBERTa, FinBERT) and LLM ChatGPT. Our findings...
Reduced polyoxomolybdate has been immobilized on reduced graphene oxide exhibiting remarkable performance for sulfur mustard simulant detoxification.
Indoor positioning data reflects human mobility in indoor spaces. Revealing patterns of trajectories may help us understand mobility. Clustering methods, which are based on the measurement similarity between trajectories, important tools for identifying those patterns. However, due to specific characteristics trajectory data, it is difficult clustering methods measure trajectories. These manifested two aspects. The first that nodes have clear semantic attributes; example, a shopping mall,...
Abstract Background The introduction of fracture reduction robot can solve the problem large forces during surgeries and need to collect multiple medical images. However, because its safety has not been certified, there are few academic achievements on this type robot. To calculate factor operation, a musculoskeletal model needs be established study constraints muscles existing modelling mainly for application such as rehabilitation treatment collision in car accidents. Methods A applied is...
Abstract Background Traditional fracture reduction surgery cannot ensure the accuracy of while consuming physical strength surgeon. Although monitoring process through radiography can improve reduction, it will bring radiation harm to both patients and surgeons. Methods We proposed a novel solution that parallel robot is used for surgery. The binocular camera indirectly obtains position posture fragment wrapped by tissue measuring external markers. According clinical experience path designed...
During traditional multi-resolution infrared and visible image fusion processing, the low contrast ratio target may be weakened become inconspicuous because of opposite DN values in source images. So a novel pseudo-color enhanced algorithm based on modified attention model fast discrete curvelet transformation is proposed. The interesting regions are extracted from images by introducing motion features gained model, performed gray via rules physical characteristics sensors domain. final...
With the increased global attention to environmental protection, social responsibility and corporate governance (ESG) standards, integrating ESG into supply chain management (SCM) is becoming a strategic requirement for companies achieve competitiveness sustainable development. This study reviews impact factors of practices on competitiveness, empirically examines mechanism performance chains through questionnaire survey with sample Chinese SMEs. It was found that normative pressure,...
ABSTRACT In the clinical application of parallel external fixator, medical practitioners are required to quantify deformity parameters develop corrective strategies. However, manual measurement angles is a complex and time‐consuming process that susceptible subjective factors, resulting in nonreproducible results. Accordingly, this study proposes an automatic method based on deep learning, comprising three stages: tibial segment localization, contour point detection, angle calculation....
This paper puts forward a design which is presented to estimate relatively accurate HAGC control system and then predict the rolling gap. Considering many factors that influence precision of gap, we can obtain final formula gap according theoretical calculation. Besides, A SVM (support vector machine) regression model based on machine learning proposed applied According data collected in working field, train Regression predicted achieved light model. Compared with RBF neural network,...
Intelligent agents such as robots are increasingly deployed in real-world, safety-critical settings. It is vital that these able to explain the reasoning behind their decisions human counterparts, however, behavior often produced by uninterpretable models deep neural networks. We propose an approach generate natural language explanations for agent's based only on observations of states and actions, agnostic underlying model representation. show how a compact representation can be learned...
The emergence of Large Language Models (LLMs), such as ChatGPT, has revolutionized general natural language preprocessing (NLP) tasks. However, their expertise in the financial domain lacks a comprehensive evaluation. To assess ability LLMs to solve NLP tasks, we present FinLMEval, framework for Financial Model Evaluation, comprising nine datasets designed evaluate performance models. This study compares encoder-only models and decoder-only Our findings reveal that while some demonstrate...
A six-degree-of-freedom simulation of kinetic interceptors and modeling a strapdown antenna platform have been worked out in Matlab Simulink to analyze the implication nonlinear factors cross-coupling effect for platform. From results simulation, effect, which is between inner gimbal outer platform, has less on angular speed output. When factor was introduced, both that missile body were increased. The provide theoretical foundation model further research
The article develops the model of bone deformities, making 6-DOF Parallel Mechanism have widely applied to correction deformities. platform’s positional direct solution is posture motion platform. Malformation can be measured by X-ray, based on space coordinate transformation find final Regarding reverse platform kinematics paper gives a quick arithmetic program, six actuators realize requirements. For computer-assisted fracture reduction, we produced an application interface.