检索结果-内蒙古大学图书馆

TTST: A Top-k Token Selective Transformer for Remote Sensing Image Super-Resolution

IEEE Transactions on Image Processing 2024年 33卷 738-752页

作者： Xiao, Yi Yuan, Qiangqiang Jiang, Kui He, Jiang Lin, Chia-Wen Zhang, Liangpei Wuhan University School of Geodesy and Geomatics Wuhan430079 China Harbin Institute of Technology School of Computer Science and Technology Harbin150001 China National Tsing Hua University Department of Electrical Engineering The Institute of Communications Engineering Hsinchu300 Taiwan Wuhan University State Key Laboratory of Information Engineering in Surveying Mapping and Remote Sensing Wuhan430079 China

Transformer-based method has demonstrated promising performance in image super-resolution tasks, due to its long-range and global aggregation capability. However, the existing Transformer brings two critical challenges for applying it in large-area earth observation scenes: (1) redundant token representation due to most irrelevant tokens;(2) single-scale representation which ignores scale correlation modeling of similar ground observation targets. To this end, this paper proposes to adaptively eliminate the interference of irreverent tokens for a more compact self-attention calculation. Specifically, we devise a Residual Token Selective Group (RTSG) to grasp the most crucial token by dynamically selecting the top-k keys in terms of score ranking for each query. For better feature aggregation, a Multi-scale Feed-forward Layer (MFL) is developed to generate an enriched representation of multi-scale feature mixtures during feed-forward process. Moreover, we also proposed a Global Context Attention (GCA) to fully explore the most informative components, thus introducing more inductive bias to the RTSG for an accurate reconstruction. In particular, multiple cascaded RTSGs form our final Top-k Token Selective Transformer (TTST) to achieve progressive representation. Extensive experiments on simulated and real-world remote sensing datasets demonstrate our TTST could perform favorably against state-of-the-art CNN-based and Transformer-based methods, both qualitatively and quantitatively. In brief, TTST outperforms the state-of-the-art approach (HAT-L) in terms of PSNR by 0.14 dB on average, but only accounts for 47.26% and 46.97% of its computational cost and parameters. The code and pre-trained TTST will be available at https://***/XY-boy/TTST for validation. © 1992-2012 IEEE.

关键词： Remote sensing

来源：评论

学校读者我要写书评

暂无评论

Near-field radiative heat transfer between nanoporous GaN films

引用

Chinese Physics B 2024年第4期33卷 109-120页

作者：韩晓政张纪红刘皓佗吴小虎冷惠文 School of Electromechanical and Automotive Engineering Yantai UniversityYantai 264005China Key Laboratory of Advanced Manufacturing and Intelligent Technology Ministry of EducationHarbin University of Science and TechnologyHarbin 150080China Thermal Science Research Center Shandong Institute of Advanced TechnologyJinan 250100China

Photon tunneling effects give rise to surface waves,amplifying radiative heat transfer in the near-field *** research has highlighted that the introduction of nanopores into materials creates additional pathways for heat transfer,leading to a substantial enhancement of near-field radiative heat transfer(NFRHT).Being a direct bandgap semiconductor,GaN has high thermal conductivity and stable resistance at high temperatures,and holds significant potential for applications in optoelectronic ***,study of NFRHT between nanoporous GaN films is currently lacking,hence the physical mechanism for adding nanopores to GaN films remains to be discussed in the field of *** this work,we delve into the NFRHT of GaN nanoporous films in terms of gap distance,GaN film thickness and the vacuum filling *** results demonstrate a 27.2%increase in heat flux for a 10 nm gap when the nanoporous filling ratio is ***,the spectral heat flux exhibits redshift with increase in the vacuum filling *** be more precise,the peak of spectral heat flux moves fromω=1.31×10^(14)rad·s^(-1)toω=1.23×10^(14)rad·s^(-1)when the vacuum filling ratio changes from f=0.1 to f=0.5;this can be attributed to the excitation of surface phonon *** introduction of graphene into these configurations can highly enhance the NFRHT,and the spectral heat flux exhibits a blueshift with increase in the vacuum filling ratio,which can be explained by the excitation of surface plasmon *** findings offer theoretical insights that can guide the extensive utilization of porous structures in thermal control,management and thermal modulation.

关键词： near-field radiative heat transfer nanoporous GaN film surface phonon polaritons surface plasmon polaritons

来源：评论

学校读者我要写书评

暂无评论

Biologically inspired anthraquinone redox centers and biomass graphene for renewable colloidal gels toward ultrahigh-performance flexible micro-supercapacitors

引用

Journal of Materials science & technology 2023年第20期151卷 178-189页

作者： Tiansheng Wang Shunyou Hu Yuanyuan Hu Dong Wu Hao Wu Jinxu Huang Hao Wang Weiwei Zhao Wen Yu Mi Wang Jie Xu Jiaheng Zhang School of Materials Science and Engineering Harbin Institute of TechnologyHarbin 150001China Key Laboratory of Micro-systems and Micro-structures Manufacturing of Ministry of Education Harbin Institute of TechnologyHarbin 150001China National Innovation Center for Advanced Medical Devices Harbin Institute of TechnologyShenzhen 457001China Research Centre of Printed Flexible Electronics Harbin Institute of TechnologyShenzhen 518055China Sauvage Laboratory for Smart Materials School of Materials Science and EngineeringHarbin Institute of Technology(Shenzhen)ShenzhenGuangdong 518055China Department of Chemistry University of IdahoMoscowIdaho 83844-2343USA

Biomass carbon and small redox biomolecules are attractive materials for green,sustainable energy storage devices owing to their environmentally friendly,low-cost,scalable,and novel ***,most devices manufactured using these materials have low specific capacitance,poor cycle stability,short lifetime,complexity,and low precision of device ***,we report the directed self-assembly of mononuclear anthraquinone(MAQ)derivatives and porous lignin-based graphene oxide(PLGO)into a renewable colloidal gel through noncovalent *** self-assembled gel electrode materials exhibited high capacitance(484.8 F g^(−1) at a current density of 1 A g^(−1))and could be further printed as flexible micro-supercapacitors(FMSCs)with arbitrary patterns and a relatively high resolution on specific *** FMSCs exhibited excellent areal capacitance(43.6 mF cm^(−2)),energy and power densities(6.1μWh cm^(−2) and 50μW cm^(−2),respectively),and cycle stability(>10,000 cycles).Furthermore,the printed FMSCs and integrated FMSC arrays exhibited remarkable flexibility while maintaining a stable *** proposed approach can be applied to other quinone biomolecules and biomass-based carbon *** study provides a basis for fabricating green and sustainable energy storage device architectures with high capacitance,long-term cycling,high scalability,and high precision.

关键词： Biomass carbon Mononuclear anthraquinone Noncovalent interactions Renewable colloidal gel Flexible micro-supercapacitors

来源：评论

学校读者我要写书评

暂无评论

Prediction of Time Series Data based on PCA-HMM-BN 1

Prediction of Time Series Data based on PCA-HMM-BN

引用

1st International Conference on Software, Systems and Information technology, SSITCON 2024

作者： Liu, Yutong Heilongiang University of Science and Technology School of Computer and Information Engineering Harbin China

ISBN: (纸本)9798350352931

This present work describes a framework for the analysis of score dynamics and energy fluctuations in competitive environments, aimed at forecasting match outcomes and providing strategic decisions useful to coaching and players' performance. Potential energy was defined, seven key measures were identified related to Kendall's tau, shown as heat maps to outline the underlying patterns. Using the CV method and the AHP composite weighting, overall scores can determine the overall possible energy level of players. The line graph should be drawn based on the latter in order to assess volatility throughout the course of the match. Then, it carries out a derivation method for potential energy series, followed by extracting the most critical turning points in energy dynamics through the PCA-HMM-BN algorithm. Feature engineering with lag features and effect coding improves predictability. This graph presents a comparison among different machine learning models, namely BP Neural Network, XGBoost, and SVM, by showing the high performance obtained by an SVM model tuned with the ASAA algorithm, with an R-squared score of 0.99967. © 2024 IEEE.

关键词： Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

Multi-channel mantis search spatial attention convolutional neural network based hyperspectral image change detection

引用

Multimedia Tools and Applications 2025年 1-22页

作者： Divya, V.M. Department of Computer Science and Engineering East West Institute of Technology Visvesvaraya Technological University Belagavi560018 India

Hyperspectral imaging has transformed remote sensing by offering detailed spectral data for monitoring environmental changes, like urban development and terrestrial cover variations. However, existing change detection techniques often face challenges like underfitting and overfitting. In this research, a novel approach is introduced for hyperspectral image change detection using a Multi-Channel Mantis Search (MCMS). This innovative methodology leverages the power of a Multi-Channel Convolutional Neural Network (MCNN) to process each spectral band independently, facilitating the capture of unique features within each band. The input hyperspectral images are gathered from the Pavia university Dataset and the Indian Pines Dataset. This approach leads to significantly improved feature extraction, enhancing the discrimination of different materials in hyperspectral images. To further enhance the performance of the change detection model, the Mantis Search Algorithm (MSA) is integrated. The MSA is employed to optimize the model's hyperparameters, fine-tuning its architecture and training parameters to achieve superior results. The combination of MCNN and MSA improves hyperspectral image change detection accuracy, achieving 98.21% and 99.31% with the Indian Pines and Pavia university datasets correspondingly. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2025.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

A novel pure data-selection framework for day-ahead wind power forecasting

引用

Fundamental Research 2023年第3期3卷 392-402页

作者： Ying Chen Jingjing Zhao Jiancheng Qin Hua Li Zili Zhang School of Economics and Management Harbin Institute of TechnologyHarbin 150001China School of Computer Science and Technology Harbin Institute of TechnologyHarbin 150001China Department of Mechanical and Industrial Engineering Texas A&M University-KingsvilleTexas 78363USA

Numerical weather prediction(NWP)data possess internal inaccuracies,such as low NWP wind speed corresponding to high actual wind power *** study is intended to reduce the negative effects of such inaccuracies by proposing a pure data-selection framework(PDF)to choose useful data prior to modeling,thus improving the accuracy of day-ahead wind power ***,we convert an entire NWP training dataset into many small subsets and then select the best subset combination via a validation set to build a forecasting *** a small subset can increase selection flexibility,it can also produce billions of subset combinations,resulting in computational *** address this problem,we incorporated metamodeling and optimization steps into *** then proposed a design and analysis of the computer experiments-based metamodeling algorithm and heuristic-exhaustive search optimization algorithm,*** results demonstrate that(1)it is necessary to select data before constructing a forecasting model;(2)using a smaller subset will likely increase selection flexibility,leading to a more accurate forecasting model;(3)PDF can generate a better training dataset than similarity-based data selection methods(e.g.,K-means and support vector classification);and(4)choosing data before building a forecasting model produces a more accurate forecasting model compared with using a machine learning method to construct a model directly.

关键词： Day-ahead wind power forecasting Data selection Design and analysis of computer experiments Heuristic optimization Numerical weather prediction data

来源：评论

学校读者我要写书评

暂无评论

First-principles study on the interfacial bonding strength and segregation at Mg/MgZn_(2)matrix interface

引用

Journal of Magnesium and Alloys 2024年第10期12卷 4053-4062页

作者： Longke Bao Peng Du Shengkun Xi Cuiping Wang Kaihong Zheng Rongpei Shi Guoqiang Xie Xingjun Liu State Key Laboratory of Advanced Welding and Joining Harbin Institute of TechnologyShenzhen 518055PR China College of Materials Science and Engineering and Institute of Materials Genome and Big DataHarbin Institute of TechnologyShenzhen 518055PR China College of Materials Science and Engineering Xiamen UniversityXiamen 361005PR China Institute of New Materials Guangdong Academy of SciencesGuangzhou 510650PR China

To understand the interface characteristics between the precipitateβ2'and the Mg matrix,and thus guide the development of new Mg-Zn alloys,we investigated the atomic interface structure,work of adhesion(Wad),and interfacial energy(γ)of Mg(0001)/β2’(MgZn_(2))(0001)interface,as well as the effect of segregation behavior of the introduced transition metal atoms(3d,4d and 5d)on interfacial bonding *** calculated works of adhesion and interfacial energies dementated that the Zn2-terminated MT+HCP configuration is the most stable structure for all considered *** the Zn2-MT+HCP interface as the research object,estimated segregated energies(Eseg)reveal that added transition metal atoms prefer to segregate at Mg-I and Mg-II *** predicted Wad and charge density difference results reveal that the segregation of alloying additives employed may all strengthen Mg(0001)/MgZn_(2)(0001)interface,with the enhancement effect of Os,Re,Tc,W,and Ru at the Mg-II site being the most pronounced.

关键词： Mg/MgZn_(2)interface Work of adhesion Interfacial energy Interface segregation

来源：评论

学校读者我要写书评

暂无评论

Surface flashover in 50 years:Theoretical models and competing mechanisms

引用

High Voltage 2023年第5期8卷 853-877页

作者： Zhen Li Ji Liu Yoshimichi Ohki George Chen He Gao Shengtao Li Key Laboratory of Engineering Dielectrics and Its Application Ministry of EducationHarbin University of Science and TechnologyHarbinChina State Key Laboratory of Electrical Insulation and Power Equipment Xi'an Jiaotong UniversityXi'anChina Research Institute for Materials Science and Technology Waseda UniversityShinjuku‐kuJapan School of Electronics and Computer Science University of SouthamptonSouthamptonUK

Surface flashover is a devastating electronic avalanche along the gas–solid interface when a high electric field is applied,which is a potential issue that threatens the safe operations of advanced power electronic,electrical,and spacecraft ***,the underlying physical mechanisms for surface flashover development are still under investigation owing to the complex charge transport processes through the gas phase,solid phase,and gas–solid *** this study,the history of surface flashover theory in the last 50 years is introduced,and several key questions are reviewed from the perspective of the competing mechanisms of charge transport:the role of each phase in a surface flashover,the origin of surface charging,and effects of traps in solid on surface ***,some suggestions involve charge transport processes in each phase,and their correlations are put forward,and a predictable‘charge transport competitive flashover model’is proposed by clarifying the competing mechanisms of charge transport processes through multiple *** study summarises the history and hot topics of physical mechanisms of surface flashover proposed based on classic and recent progress and offers promising routes for developing a more precise surface flashover theory and improving surface flashover performances.

关键词： flash transport charge

来源：评论

学校读者我要写书评

暂无评论

Orbital angular momentum-mediated machine learning for high-accuracy mode-feature encoding

引用

Light(science & Applications) 2024年第3期13卷 466-477页

作者： Xinyuan Fang Xiaonan Hu Baoli Li Hang Su Ke Cheng Haitao Luan Min Gu Institute of Photonic Chips University of Shanghai for Science and TechnologyShanghai 200093China Centre for Artificial-Intelligence Nanophotonics School of Optical-Electrical and Computer EngineeringUniversity of Shanghai for Science and TechnologyShanghai 200093China

Machine learning with optical neural networks has featured unique advantages of the information processing including high speed,ultrawide bandwidths and low energy consumption because the optical dimensions(time,space,wavelength,and polarization)could be utilized to increase the degree of ***,due to the lack of the capability to extract the information features in the orbital angular momentum(OAM)domain,the theoretically unlimited OAM states have never been exploited to represent the signal of the input/output nodes in the neural network ***,we demonstrate OAM-mediated machine learning with an all-optical convolutional neural network(CNN)based on Laguerre-Gaussian(LG)beam modes with diverse diffraction *** proposed CNN architecture is composed of a trainable OAM mode-dispersion impulse as a convolutional kernel for feature extraction,and deep-learning diffractive layers as a *** resultant OAM mode-dispersion selectivity can be applied in information mode-feature encoding,leading to an accuracy as high as 97.2%for MNIST database through detecting the energy weighting coefficients of the encoded OAM modes,as well as a resistance to eavesdropping in point-to-point free-space ***,through extending the target encoded modes into multiplexed OAM states,we realize all-optical dimension reduction for anomaly detection with an accuracy of 85%.Our work provides a deep insight to the mechanism of machine learning with spatial modes basis,which can be further utilized to improve the performances of various machine-vision tasks by constructing the unsupervised learning-based auto-encoder.

关键词： polarization mode momentum

来源：评论

学校读者我要写书评

暂无评论

Research on reactor power prediction of nuclear power plant based on multivariate optimization GRU model

引用

International Journal of Advanced Nuclear Reactor Design and technology 2024年第2期6卷 78-89页

作者： Tan, Canyi Wang, Bo Li, Jiangkuan Chen, Jie Liang, Biao Zheng, Shangcai Han, Rui Tian, Ruifeng Tan, Sichao College of Nuclear Science and Technology Harbin Engineering University Harbin150001 China Heilongjiang Provincial Key Laboratory of Nuclear Power System & Equipment Harbin Engineering University Harbin150001 China Nuclear Power Institute of China Chengdu610213 China

In the operation of nuclear power plants, the accurate prediction of power change trends is crucial for ensuring safety and stability. In this work, a ML-GRU-RS method, based on model-agnostic meta-learning (MAML), gate recurrent unit (GRU), and random search optimization, is proposed for the long-term prediction of key parameters of nuclear power plants. This method combines the fast adaptability of MAML, the time series data processing capability of GRU, and the optimization efficiency of random search to achieve high-precision predictions under varying power conditions. The results demonstrate that this method can effectively predict the future trends of key parameters in nuclear power plants. The ability of operators to anticipate these trends has been significantly enhanced, contributing to the overall safety of the nuclear power plants. © 2024 Xi'an Jiaotong university

关键词： Prediction models

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：