检索结果-南通市图书馆

Advanced Algorithm for Parameters Estimation of Negative Binomial Distribution with High Dimensional Sparse Group Structure

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Journal of Systems science & Complexity 2024年第5期37卷 2173-2195页

作者： LI Meiqi JIN Baisuo school of Management University of Science and Technology of China

Negative binomial regression is a powerful technique for modeling count data, particularly when dealing with overdispersion. However, estimating the parameters for large-dimensional sparse models is challenging due to the complexity of optimizing the mean and dispersion parameter of the negative binomial distribution. To address this issue, the authors propose a novel approach that employs two iterations of the majorize-minimize(MM) algorithm, one for estimating the dispersion parameter and the other for estimating the mean parameters. These approaches improve the convergence speed and stability of the algorithm. The authors also use group penalty for variable selection, which enhances the accuracy and efficiency of the algorithm. The proposed method provides an explicit solution,simplifies the iteration process, and maintains good stability while ensuring algorithm ***, the authors apply the proposed algorithm to the zero-inflated model and demonstrate its promising predictive performance on specific data sets. The research has important implications for count data modeling and analysis in various fields, such as data mining, machine learning, and bioinformatics.

关键词： Group structure majorize-minimize (MM) negative binomial

Output feedback stabilization of stochastic high-order planar nonlinear systems with stochastic inverse dynamics and output-constraint

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

science China(Information sciences) 2024年第3期67卷 182-196页

作者： Ruiming XIE Shengyuan XU Nanjing University of science and technology school of Automation

In this paper, we solve the output feedback control problem of stochastic high-order planar nonlinear systems with output constraint and stochastic integral input-to-state stability(SiISS) inverse dynamics. By employing a key coordinate transformation, a stochastic nonlinear system with output constraint and SiISS inverse dynamics is converted into an unconstrained system. By skillfully constructing an observer and adopting SiISS small-gain conditions, we develop a new output feedback control design and analysis method, and prove that all the closed-system signals are bounded almost surely, the output constraint is not violated almost surely, and the equilibrium point of the closed-loop system is stochastically asymptotically stable.

关键词： stochastic high-order planar nonlinear system outputconstraint stochastic inverse dynamics output feedbackstabilization stochastic integral input-to-state stability(SiISS)

Deep Guided Attention Network for Joint Denoising and Demosaicing in Real Image

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Chinese Journal of Electronics 2024年第1期33卷 303-312页

作者： Tao ZHANG Ying FU Jun ZHANG school of Computer science and technology Beijing Institute of Technology Advanced Reasearch Institute of Multidisciplinary science Beijing Institute of Technology

Denoising(DN) and demosaicing(DM) are the first crucial stages in the image signal processing pipeline. Recently, researches pay more attention to solve DN and DM in a joint manner, which is an extremely undetermined inverse problem. Existing deep learning methods learn the desired prior on synthetic dataset, which limits the generalization of learned network to the real world data. Moreover, existing methods mainly focus on the raw data property of high green information sampling rate for DM, but occasionally exploit the high intensity and signalto-noise(SNR) of green channel. In this work, a deep guided attention network(DGAN) is presented for real image joint DN and DM(JDD), which considers both high SNR and high sampling rate of green information for DN and DM, respectively. To ease the training and fully exploit the data property of green channel, we first train DN and DM sub-networks sequentially and then learn them jointly, which can alleviate the error accumulation. Besides, in order to support the real image JDD, we collect paired raw clean RGB and noisy mosaic images to conduct a realistic dataset. The experimental results on real JDD dataset show the presented approach performs better than the state-of-the-art methods, in terms of both quantitative metrics and qualitative visualization.

关键词： Training Visualization Inverse problems Green products Noise reduction Pipelines Signal processing

Enhanced conductivity and weakened magnetism in Pb-doped Sr2IrO4

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Chinese Physics B 2024年第1期33卷 696-701页

作者：岳智来甄伟立牛瑞焦珂珂朱文卡皮雳张昌锦 High Magnetic Field Laboratory Hefei Institutes of Physical ScienceChinese Academy of Sciences science Island Branch of graduate school University of Science and Technology of China Collaborative Innovation Center of Advanced Microstructures Nanjing University

Group IV element Pb has been selected as the dopant to dope at the Sr site of Sr2IrO4. It is exciting to find that the single-phase crystal structure could be maintained with a high Pb doping level of up to x=0.3 in Sr2-xPbxIrO4. The mapping data obtained from energy-dispersive x-ray spectroscopy analyses give solid evidence that the Pb ions are uniformly distributed in the Sr2IrO4matrix. The incorporation of Pb leads to a moderate depression of the canted antiferromagnetic ordering state. The electrical conductivity could be greatly enhanced when the Pb doping content is higher than x=*** present results give a fresh material base to explore new physics in doped Sr2IrO4systems.

关键词： iridates doping conductivity magnetism

Novel Rifle Number Recognition Based on Improved YOLO in Military Environment

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Computers, Materials & Continua 2024年第1期78卷 249-263页

作者： Hyun Kwon Sanghyun Lee Department of Artificial Intelligence and Data science Korea Military AcademySeoulKorea graduate school of Information Security Korea Advanced Institute of Science and TechnologyDaejeonKorea

Deep neural networks perform well in image recognition,object recognition,pattern analysis,and speech *** military applications,deep neural networks can detect equipment and recognize *** military equipment,it is necessary to detect and recognize rifle management,which is an important piece of equipment,using deep neural *** have been no previous studies on the detection of real rifle numbers using real rifle image *** this study,we propose a method for detecting and recognizing rifle numbers when rifle image data are *** proposed method was designed to improve the recognition rate of a specific dataset using data fusion and transfer *** the proposed method,real rifle images and existing digit images are fusedas trainingdata,andthe final layer is transferredto theYolov5 *** detectionand recognition performance of rifle numbers was improved and analyzed using rifle image and numerical *** used actual rifle image data(K-2 rifle)and numeric image datasets,as an experimental *** was used as the machine learning *** results show that the proposed method maintains 84.42% accuracy,73.54% precision,81.81% recall,and 77.46% F1-score in detecting and recognizing rifle *** proposed method is effective in detecting rifle numbers.

关键词： Machine learning deep neural network rifle number recognition detection

N-doped graphene quantum dot-decorated N-TiO2/P-doped porous hollow g-C_(3)N_(4) nanotube composite photocatalysts for antibiotic photodegradation and H2 production

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

International Journal of Minerals,Metallurgy and Materials 2024年第1期31卷 165-178页

作者： Jingshu Yuan Yao Zhang Xiaoyan Zhang Junjie Zhang Shen’gen Zhang Institute for Advanced Materials and technology University of Science and Technology BeijingBeijing 100083China Xi’an Fanyi University Xi’an 710105China Shunde graduate school of University of science and technology Beijing Foshan 528399China State Key Laboratory of Lunar and Planetary sciences Macao University of Science and TechnologyMacao SAR 820000China

Exclusive responsiveness to ultraviolet light (~3.2 eV) and high photogenerated charge recombination rate are the two primary drawbacks of pure TiO_(2). We combined N-doped graphene quantum dots (N-GQDs), morphology regulation, and heterojunction construction strategies to synthesize N-GQD/N-doped TiO_(2)/P-doped porous hollow g-C_(3)N_(4) nanotube (PCN) composite photocatalysts (denoted as G-TPCN). The optimal sample (G-TPCN doped with 0.1wt% N-GQD, denoted as 0.1% G-TPCN) exhibits significantly enhanced photoabsorption, which is attributed to the change in bandgap caused by elemental doping (P and N), the improved light-harvesting resulting from the tube structure, and the upconversion effect of N-GQDs. In addition, the internal charge separation and transfer capability of0.1% G-TPCN are dramatically boosted, and its carrier concentration is 3.7, 2.3, and 1.9 times that of N-TiO_(2), PCN, and N-TiO_(2)/PCN(TPCN-1), respectively. This phenomenon is attributed to the formation of Z-scheme heterojunction between N-TiO_(2) and PCNs, the excellent electron conduction ability of N-GQDs, and the short transfer distance caused by the porous nanotube structure. Compared with those of N-TiO_(2), PCNs, and TPCN-1, the H2 production activity of 0.1%G-TPCN under visible light is enhanced by 12.4, 2.3, and 1.4times, respectively, and its ciprofloxacin (CIP) degradation rate is increased by 7.9, 5.7, and 2.9 times, respectively. The optimized performance benefits from excellent photoresponsiveness and improved carrier separation and migration efficiencies. Finally, the photocatalytic mechanism of 0.1% G-TPCN and five possible degradation pathways of CIP are proposed. This study clarifies the mechanism of multiple modification strategies to synergistically improve the photocatalytic performance of 0.1% G-TPCN and provides a potential strategy for rationally designing novel photocatalysts for environmental remediation and solar energy conversion.

关键词： N-doped TiO_(2) N-doped graphene quantum dots P-doped g-C_(3)N_(4) porous hollow nanotube heterojunction photocatalysis

维普期刊数据库

Enhanced Differentiable Architecture Search Based on Asymptotic Regularization

在线全文

学校读者我要写书评

暂无评论

Computers, Materials & Continua 2024年第2期78卷 1547-1568页

作者： Cong Jin Jinjie Huang Yuanjian Chen Yuqing Gong school of Computer science and technology Harbin University of Science and TechnologyHarbin150006China school of Automation Harbin University of Science and TechnologyHarbin150006China

In differentiable search architecture search methods,a more efficient search space design can significantly improve the performance of the searched architecture,thus requiring people to carefully define the search space with different complexity according to various *** rationalizing the search strategies to explore the well-defined search space will further improve the speed and efficiency of architecture *** this in mind,we propose a faster and more efficient differentiable architecture search method,***,we introduce a more efficient search space enriched by the introduction of two redefined convolution ***,we utilize a more efficient architectural parameter regularization method,mitigating the overfitting problem during the search process and reducing the error brought about by gradient ***,we introduce a natural exponential cosine annealing method to make the learning rate of the neural network training process more suitable for the search ***,group convolution and data augmentation are employed to reduce the computational ***,through extensive experiments on several public datasets,we demonstrate that our method can more swiftly search for better-performing neural network architectures in a more efficient search space,thus validating the effectiveness of our approach.

关键词： Differentiable architecture search allegro search space asymptotic regularization natural exponential cosine annealing

Federated Meta Reinforcement Learning for Personalized Tasks

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Tsinghua science and technology 2024年第3期29卷 911-926页

作者： Wentao Liu Xiaolong Xu Jintao Wu Jielin Jiang school of Computer science Nanjing University of Information Science and TechnologyNanjing 210044China school of Software Nanjing University of Information Science and TechnologyNanjing 210044China

As an emerging privacy-preservation machine learning framework,Federated Learning(FL)facilitates different clients to train a shared model collaboratively through exchanging and aggregating model parameters while raw data are kept local and *** this learning framework is applied to Deep Reinforcement Learning(DRL),the resultant Federated Reinforcement Learning(FRL)can circumvent the heavy data sampling required in conventional DRL and benefit from diversified training data,besides privacy preservation offered by *** FRL implementations presuppose that clients have compatible tasks which a single global model can *** practice,however,clients usually have incompatible(different but still similar)personalized tasks,which we called task *** may severely hinder the implementation of FRL for practical *** this paper,we propose a Federated Meta Reinforcement Learning(FMRL)framework by integrating Model-Agnostic Meta-Learning(MAML)and ***,we innovatively utilize Proximal Policy Optimization(PPO)to fulfil multi-step local training with a single round of ***,considering the sensitivity of learning rate selection in FRL,we reconstruct the aggregation optimizer with the Federated version of Adam(Fed-Adam)on the server *** experiments demonstrate that,in different environments,FMRL outperforms other FL methods with high training efficiency brought by Fed-Adam.

关键词： federated learning reinforcement learning meta-learning personalization

An in-situ polymerized interphase engineering for high-voltage allsolid- state lithium-metal batteries

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Nano Research 2024年第4期17卷 2687-2692页

作者： Lu Nie Shaojie Chen Mengtian Zhang Tianyi Gao Yuyao Zhang Ran Wei Yining Zhang Wei Liu school of Physical science and technology ShanghaiTech UniversityShanghai 201210China Tsinghua-Berkeley Shenzhen Institute&Tsinghua Shenzhen International graduate school Tsinghua UniversityShenzhen 518055China

All-solid-state lithium batteries(ASSLBs)have attracted great interest due to their promising energy density and strong ***,the interface issues,including large interfacial resistance between electrode and electrolyte and low electrochemical stability of solid-state electrolytes against high-voltage cathodes,have restricted the development of high-voltage ***,we report an ASSLB with stable cycling by adopting a conformal polymer interlayer in-situ formed at the Li_(6.4)La_(3)Zr_(1.4)Ta_(0.6)O_(12)(LLZTO)–cathode *** polymer can perfectlyﬁll the voids and create a stable interface contact between LLZTO and *** addition,the electric field across the polymer interlayer is reduced compared with pure solid polymer electrolyte(SPE),which facilitates the electrochemical stability with high-voltage *** all-solid-state Li|LLZTO-SPE|LiFe_(0.4)Mn_(0.6)PO_(4)(LMFP)cells achieve a low interface impedance,high specific capacity,and excellent cycling *** work presents an effective and practical strategy to rationally design the electrode–electrolyte interface for the application of high-voltage ASSLBs.

关键词： all-solid-state lithium batteries interface contact electric field electrochemical stability