Xuewen Ding

Hohai University

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Trends

Author Order

Document Type

Co-Authors

P.J. Griffin

Doble (Norway)

Zhentao Liu

China University of Geosciences

Sibo Wang

Chinese University of Hong Kong

Yuxin Zhang

Virginia Tech

Hongmin Gao

Hohai University

Shufang Xu

Hohai University

Y. Liu

Danyun Li

Shanghai University of Traditional Chinese Medicine

Yufan Liu

Shaanxi University of Science and Technology

Yang Liu

Virginia Tech

Cooperative Institutions

Shandong Institute of Automation

Chinese Academy of Sciences

Hohai University

University of Chinese Academy of Sciences

China University of Geosciences

North China Electric Power University

Ministry of Education of the People's Republic of China

Ministry of Water Resources of the People's Republic of China

Northeastern University

Tsinghua University

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Field

Electroencephalogram Emotion Recognition Using Variational Modal Decomposition Based Dispersion Entropy Feature Extraction

Si-Jun Hu Zhentao Liu Xuewen Ding

Electroencephalogram (EEG) emotion recognition has gained considerable attention due to its ability to reflect people's inner emotional states objectively and naturally. Feature extraction is a critical step in EEG emotion recognition because of non-stationarity and irregularity of EEG signals. A feature extraction method using Variational Modal Decomposition (VMD) to extract Dispersion Entropy (DispEn) is proposed in this paper. Raw EEG signal is decomposed into several components, and DispEn of each component is extracted in eight emotion-related channels. Our method was tested on DEAP dataset in which the EEG emotional states are accessed in Valence-Arousal emotional space. Four emotional states (i.e., HVHA, HVLA, LVHA, LVLA) are classified by Support Vector Machine (SVM). The experimental results show that the accuracy of emotion recognition is 77.87%, which demonstrates its effectiveness.

10.23919/ccc52363.2021.9549884

Cite

Citations (2)

Electroencephalogram Emotion Recognition Based on A Stacking Classification Model

Oiao Xie Zhentao Liu Xuewen Ding

To improve the accuracy of Electroencephalogram (EEG) emotion recognition, a stacking emotion classification model is proposed, in which different classification models such as XGBoost, LightGBM and Random Forest are integrated to learn the features. In addition, the Renyi entropy of 32 channels' EEG signals are extracted as the feature and Linear discriminant analysis (LDA) is employed to reduce the dimension of the feature set. The proposal is tested on the DEAP dataset, and the EEG emotional states are accessed in Arousal-Valence emotion space, in which HA/LA and HV/LV are classified, respectively. The result shows that the average recognition accuracies of 77.19% for HA/LA and 79.06% for HV/LV are obtained, which demonstrates that the proposal is feasible in EEG emotion recognition.

Feature vector

Feature (linguistics)

10.23919/chicc.2018.8483496

Cite

Citations (15)

Multi-objective optimization design of passive filter based on particle swarm optimization

Journal of Physics Conference Series (2020)

Sibo Wang Xuewen Ding Jin Wang

Abstract Using particle swarm optimization algorithm, aiming at the investment cost, filtering effect and reactive power compensation ability of passive filter, the weight of each objective function is determined according to the dispersion order method, and the passive filter is designed by multi-objective optimization. In addition, the actual calculation is carried out by taking the power grid parameters of the self-elevating and self-propelled petroleum engineering ship of “Feng Long 01” as an example. Further, a ship power grid simulation model is built under the Matlab/Simulink environment to verify the practicality of the optimized passive filter design.

10.1088/1742-6596/1549/3/032017

Cite

Citations (5)

Research on overflow diffusion characteristics of dredger

Journal of Physics Conference Series (2020)

Mingqi Shao Jin Wang Xuewen Ding

Abstract Dredging engineering is widely used in port and waterway construction, water conservancy facilities construction, environmental protection dredging, land reclamation, subsea tunnel excavation and deep-sea mining, among which drag suction dredger is an important tool for dredging operation. During the process of its construction, the mud overflow process is an important link to realize the mud treatment of excavation. According to the actual engineering conditions, the loading and overflow process is an important part of the treatment of excavated slurry. With the increasing requirements of environmental protection, the related ecological pollution caused by dredging has become a hot spot today, and the diffusion of overflow slurry is one of the sources of pollution generated by dredging suction dredgers. Therefore, combining with actual engineering conditions, this paper adopts the theory of computational fluid dynamics (CFD) to study the diffusion characteristics, diffusion range, and influencing factors and proposes construction strategies to reduce its pollution impact.

Dredging

Subsea

10.1088/1742-6596/1600/1/012003

Cite

Citations (0)

An Artificial Neural Network Approach to Transformer Fault Diagnosis

IEEE Power Engineering Review (1996)

Yuxin Zhang Xuewen Ding Yuanwei Liu P.J. Griffin

This paper presents an artificial neural network (ANN) approach to the diagnosis and detection of faults in oil-filled power transformers based on dissolved gas-in-oil analysis. A two-step ANN method is used to detect faults with or without cellulose involved. Good diagnosis accuracy is obtained with the proposed approach.

10.1109/mper.1996.4311026

Cite

Citations (259)

ANN based transformer fault diagnosis using gas-in-oil analysis

Xuewen Ding Erlin Yao Y. Liu P.J. Griffin

This paper presents an artificial neural network (ANN) approach to detect faults in oil-filled power transformers. The relationship between transformer faults of power transformers and gases dissolved in insulating oil is reviewed. ANN based data processing and diagnostic techniques are described. Preliminary simulation results show a 95% correct diagnosis rate for the ANN based method with modest amount of training data.

Dissolved Gas Analysis

Oil analysis

Source

Cite

Citations (14)

Numerical simulation of coarse particle pipeline transportation based on Eulerian Lagrangian model

Journal of Physics Conference Series (2020)

Sibo Wang Xuewen Ding Jin Wang

Abstract In this paper, based on the Eulerian Lagrangian model, the non-phase change flow process of the coarse particle slurry in the horizontal pipeline is studied. The particle motion state is tracked, the dynamic characteristics of the coarse particle slurry in the pipeline transportation are explored, and the resistance characteristics of the pipeline and the operation stability under different conditions are analyzed. The comparison of the results of velocity distribution, concentration distribution and pressure curve with the empirical formula and experimental data shows that the simulation results are in good agreement with the calculation results of wasp formula and the experimental values of Roco&Balakrishnam, and the simulation can better describe the flow characteristics in the basin. In addition, the flow characteristics of particles are further analyzed according to the key factors such as the flow stability, the transition process and the critical deposition velocity of particles in different positions of the pipeline.

Particle (ecology)

Multiphase flow

10.1088/1742-6596/1600/1/012005

Cite

Citations (2)

Dual-feature attention-based contrastive prototypical clustering for multimodal remote sensing data

IEEE Transactions on Geoscience and Remote Sensing (2024)

Shufang Xu Xuewen Ding Yiyan Zhang Zhen Zhang Hongmin Gao

Feature (linguistics)

10.1109/tgrs.2024.3492500

Cite

Citations (0)

Token Caching for Diffusion Transformer Acceleration

arXiv (Cornell University) (2024)

Jing Lou Wenbo Luo Yufan Liu Bing Li Xuewen Ding

Diffusion transformers have gained substantial interest in diffusion generative modeling due to their outstanding performance. However, their high computational cost, arising from the quadratic computational complexity of attention mechanisms and multi-step inference, presents a significant bottleneck. To address this challenge, we propose TokenCache, a novel post-training acceleration method that leverages the token-based multi-block architecture of transformers to reduce redundant computations among tokens across inference steps. TokenCache specifically addresses three critical questions in the context of diffusion transformers: (1) which tokens should be pruned to eliminate redundancy, (2) which blocks should be targeted for efficient pruning, and (3) at which time steps caching should be applied to balance speed and quality. In response to these challenges, TokenCache introduces a Cache Predictor that assigns importance scores to tokens, enabling selective pruning without compromising model performance. Furthermore, we propose an adaptive block selection strategy to focus on blocks with minimal impact on the network's output, along with a Two-Phase Round-Robin (TPRR) scheduling policy to optimize caching intervals throughout the denoising process. Experimental results across various models demonstrate that TokenCache achieves an effective trade-off between generation quality and inference speed for diffusion transformers. Our code will be publicly available.

10.48550/arxiv.2409.18523

Cite

Citations (0)

An Intelligent Fault Diagnosis Scheme Based On PCA-BP Neural Network for the Marine Diesel Engine

IOP Conference Series Materials Science and Engineering (2020)

Sibo Wang Jin Wang Xuewen Ding

Abstract Using AVL-BOOST to simulate the thermal fault of diesel engine, the principal component analysis method is used to analyze the thermal fault of diesel engine, and the three principal components are selected that can reflect the original variable 99.589% information as the input of BP neural network. The failure mode of the diesel engine is used as an output to construct a three-layer neural network prediction model. The results show that the PCA-BP neural network model can well diagnose the failure mode of diesel engines.

10.1088/1757-899x/782/3/032079

Cite

Citations (5)