A5101495054- Analog and Mixed-Signal Circuit Design
- Advancements in PLL and VCO Technologies
- Video Analysis and Summarization
- Internet Traffic Analysis and Secure E-voting
- HVDC Systems and Fault Protection
- Advanced Vision and Imaging
- Solar Radiation and Photovoltaics
- Geological formations and processes
- Human Motion and Animation
- Network Traffic and Congestion Control
- Software Engineering Research
- Human Pose and Action Recognition
- Radio Frequency Integrated Circuit Design
- Solar Thermal and Photovoltaic Systems
- Network Security and Intrusion Detection
- Computational Physics and Python Applications
- Image and Video Quality Assessment
- Wind Turbine Control Systems
- Geology and Paleoclimatology Research
- Pleistocene-Era Hominins and Archaeology
- Multimedia Communication and Technology
- Particle Detector Development and Performance
- solar cell performance optimization
- Machine Learning in Materials Science
Shandong Jiaotong University
2024
Zhengzhou University
2024
University of Chinese Academy of Sciences
2023
Nanjing Institute of Geography and Limnology
2023
Chinese Academy of Sciences
2023
NSK (United States)
2023
Southern Methodist University
2015-2016
Queen's University
2007
Existing parameter-efficient fine-tuning (PEFT) methods have achieved significant success on vision transformers (ViTs) adaptation by improving parameter efficiency. However, the exploration of enhancing inference efficiency during remains underexplored. This limits broader application pre-trained ViT models, especially when model is computationally extensive. In this paper, we propose Dynamic Tuning (DyT), a novel approach to improve both and for adaptation. Specifically, besides using...
A distributed denial of service (DDoS) attack is widely regarded as a major threat for the current Internet because its ability to create huge volume unwanted traffic. It hard detect and respond DDoS attacks due large complex network environments. In this paper, we introduce two distance-based detection techniques: average distance estimation traffic separation. They by analyzing values rates. The information packet can be inferred from time- to-live (TTL) value IP header. technique,...
Automating the modelling and simulation process benefits all related areas, such as research, design, manufacturing especially training of machine learning models. This approach significantly improves efficiency speed, providing scalability in data generation, which motivates discovery new patterns. In addition, consistency reproducibility through unified methods improve accuracy by eliminating manual mistakes repetitive processes. Machine provides an efficient for accessing large-scale...
The optimal arrangement of heliostat field is great significance to energy utilization, and the rational firstly accurately calculate heat output field. To field, first optical efficiency heliostat. After geometrical analysis, micro-element method cross-product can be used determine whether incident or reflected light on each area occluded by shadow tower other heliostats, so that occlusion calculated. Then using transformation cone coordinate system mirror symmetry with respect normal...
This paper analyzes the effect of operational amplifier (OPAMP) input offset on continuous-time (CT) delta-sigma (ΔΣ) modulators with current-mode feedback digital-to-analog converters (DAC). Our analysis shows that leads to either increased harmonics or noise floor modulator, depending whether a DAC shuffling scheme, such as data-weighted averaging (DWA), is applied not. Analysis and mathematical derivation are given verified by transistor-level simulations. An input-offset cancellation...
In response to the problems of undetectable intrusions caused by absence authentication mechanism in NTP protocol used wind power monitoring systems, an intrusion detection algorithm for farm messages based on message periodicity and autoregressive model was proposed. The periodic feature under normal intruded conditions analyzed. An current network built, parameters Yule-Walker method were estimated thresholds determined within 95% confidence intervals. proposed validated real capture from...
The goal of motion understanding is to establish a reliable mapping between and action semantics, while it challenging many-to-many problem. An abstract semantic (i.e., walk forwards) could be conveyed by perceptually diverse motions (walk with arms up or swinging), carry different semantics w.r.t. its context intention. This makes an elegant them difficult. Previous attempts adopted direct-mapping paradigms limited reliability. Also, current automatic metrics fail provide assessments the...