A5101822883- Fluid Dynamics and Heat Transfer
- Surface Modification and Superhydrophobicity
- Metaheuristic Optimization Algorithms Research
- Icing and De-icing Technologies
- Robotic Path Planning Algorithms
- Advanced Multi-Objective Optimization Algorithms
- Adhesion, Friction, and Surface Interactions
- Time Series Analysis and Forecasting
- Vehicle Routing Optimization Methods
- Radiation Effects in Electronics
- Topology Optimization in Engineering
- Anomaly Detection Techniques and Applications
- Software Testing and Debugging Techniques
- Parallel Computing and Optimization Techniques
- Advanced Computing and Algorithms
- Network Security and Intrusion Detection
- Industrial Technology and Control Systems
Dalian Polytechnic University
2024
Beihang University
2024
Space Engineering University
2024
China Electronics Technology Group Corporation
2023
Zhongyuan University of Technology
2019
Existing unsupervised Re-ID methods often rely on pseudo-labels generated by clustering algorithms. However, the effectiveness of these is highly dependent quality learned features. As a result, improving extracted features will enhance performance model. In this paper, an attention-based hybrid contrastive learning method (AHCL) proposed incorporating novel attention mechanism, which combines spatial and channel to enable model learn more effective Additionally, approach introduced,...
Abstract A new form of pancake bouncing is discovered in this work when a droplet impacts onto micro‐structured superhydrophobic surfaces an environment pressure less than 2 kPa, and unprecedented reduction contact time by ≈85% obtained. The mechanisms forming unique phenomenon are examined combining experimental observation, numeical modelling improved theoretical model for the overpressure effect arising from vaporisation inside micro‐scaled structures. competition among vapor effect,...
Multiscale nano/micro-structured surfaces with superhydrophobicity are abundantly observed in nature such as lotus leaves, rose petals and butterfly wings, where microstructures typically reinforce mechanical stability, while nanostructures predominantly govern wettability. To emulate hierarchical structures nature, various methods have been widely applied the past few decades to manufacture of multiscale which can be functionalities ranging from anti-icing water–oil separation...
A modified multi-objective harmony search algorithm called Niching Multi-objective Harmony Search Algorithm (NMOHSA) is proposed to solve multimodal optimization problems. It adopts the neighborhood information build dynamic memory for maintaining population diversity. new consideration rule also applied prevent be trapped into local optimal solution. Moreover, two key parameters, rate (HMCR) and pitch adjustment (PAR), are dynamically adjusted. Empirical results show that performs much...
Automated Program Repair (APR) aims to automatically generate patches for rectifying software bugs. Recent strides in Large Language Models (LLM), such as ChatGPT, have yielded encouraging outcomes APR, especially within the conversation-driven APR framework. Nevertheless, efficacy of is contingent on quality feedback information. In this paper, we propose ContrastRepair, a novel conversation-based approach that augments by providing LLMs with contrastive test pairs. A pair consists failing...
RPA is the completion of Robotic Process Automation. By simulating human behavior, it performs mouse and keyboard operations in human-computer interaction interface software automatically completes fixed business according to predesigned workflow, which widely applicable scenarios with high repeatability, strong complexity, clear processes. The typical application were analyzed for engineering requirements. According characteristics technology, an performance evaluation method designed,...
Real-world scenarios such as Internet, industrial equipment and finance field generate a large number of multivariate time series all the which are important for describing operational state system. Therefore, anomaly detection on has become hot topic today. How to utilize regularization eliminate overfitting is an issue since it inhibits representative power existing models. In this paper, reconstruction model called Autoregressive Graph Adversarial Network (ARGAN) proposed. First, we...