检索结果-内蒙古大学图书馆

Incorporating Prior Knowledge into Neural Networks through an Implicit Composite Kernel

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Jiang, Ziyang Zheng, Tongshu Liu, Yiling Carlson, David Department of Civil and Environmental Engineering Duke University United States Division of Natural and Applied Science Duke Kunshan University China Program in Computational Biology and Bioinformatics Duke University School of Medicine United States Department of Civil and Environmental Engineering Department of Biostatistics and Bioinformatics Department of Computer Science Duke University United States

It is challenging to guide neural network (NN) learning with prior knowledge. In contrast, many known properties, such as spatial smoothness or seasonality, are straightforward to model by choosing an appropriate kernel in a Gaussian process (GP). Many deep learning applications could be enhanced by modeling such known properties. For example, convolutional neural networks (CNNs) are frequently used in remote sensing, which is subject to strong seasonal effects. We propose to blend the strengths of NNs and the clear modeling capabilities of GPs by using a composite kernel that combines a kernel implicitly defined by a neural network with a second kernel function chosen to model known properties (e.g., seasonality). We implement this idea by combining a deep network and an efficient mapping function based on either Nyström approximation or random Fourier features, which we call Implicit Composite Kernel (ICK). We then adopt a sample-then-optimize approach to approximate the full GP posterior distribution. We demonstrate that ICK has superior performance and flexibility on both synthetic and real-world datasets including a remote sensing dataset. The ICK framework can be used to include prior information into neural networks in many applications. © 2022, CC BY.

关键词： Remote sensing

Implementation of Apriori Algorithm for Analysis of Consumer Purchase Patterns 1

学校读者我要写书评

暂无评论

Implementation of Apriori Algorithm for Analysis of Consumer...

1st International Conference on computer science and Applied Mathematic, ICCSAM 2018

作者： Panjaitan, Suprianto Sulindawaty Amin, Muhammad Lindawati, Sri Watrianthos, Ronal Sihotang, Hengki Tamando Sinaga, Bosker Teknik Informatika STMIK Pelita Nusantara Medan Sumatera Utara Indonesia Faculty of Computer Science and Information Technology Universitas Sumatera Utara Medan Indonesia Program Doctoral Vocational Technology Education Postgraduate Engineering Faculty Padang State University Padang West Sumatera Indonesia Manajemen Informatika AMIK Logika Sumatera Utara Indonesia Manajemen Informatika AMIK Labuhan Batu Sumatera Utara Indonesia

Consumer purchasing patterns are a form of purchases made by consumers, whether someone or a lot of people to get the desired item by making a purchase transaction. One characteristic of the purchase pattern is the existence of acquiring something through exchanging money. This study aims to create an application that is used in determining consumer purchasing patterns by applying a priori algorithms and using Visual Basic 2010 as a tool for determining consumer purchase patterns. This application uses a priori algorithm calculation method where the sample consumer purchase data will be sorted and calculated by providing the value of the minium support and configuration parameters and based on the results of confidence the largest number of conclusions such as: can be used as information for determining sales, the application of a priori algorithms can provide information pattern combination item set from consumer purchase data that is with support above 15% and confidence above 50% on item set. © Published under licence by IOP Publishing Ltd.

关键词： Condensed matter physics

Surface engineering Towards Boosted Dielectric Properties for High-Temperature Polymers Under Electrical Extremes

学校读者我要写书评

暂无评论

Surface Engineering Towards Boosted Dielectric Properties fo...

Conference on Electrical Insulation and Dielectric Phenomena, (CEIDP)

作者： Jierui Zhou Marina Dabaghian Anna Marie LaChance Jing Hao Wesley Zhong Luyi Sun Yang Cao Department of Electrical and Computer Engineering University of Connecticut Storrs CT USA Electrical Insulation Research Center Institute of Materials Science University of Connecticut Storrs CT USA Department of Chemical and Biomolecular Engineering University of Connecticut Storrs CT USA Polymer Program Institute of Materials Science University of Connecticut Storrs CT USA

Polymer-thin films are critical dielectric materials for capacitive energy storage in power and electronic systems. The rapid development of wide bandgap semiconductors urgently calls for polymers that can attain high energy density and high charge-discharge efficiency at elevated temperatures. However, high-temperature polymers usually suffer from compromised dielectric breakdown strength and soaring conduction loss due to the intrinsic constraint between their bandgaps and thermal stability. In this work, we propose a facile and high throughput interface engineering technique utilizing nanocoatings formed by montmorillonite (MMT) nanoclays, which can serve as an out-of-the-box solution to bypass this constraint. The highly ordered MMT-based nanocoatings with layered structure can significantly block the excessive charge injection at high electric fields and dissipate the charge carriers along the in-plane direction. The MMT-coated PI films exhibit improved breakdown strength, suppressed conduction loss, and boosted charging-discharging efficiency. This work holds the promise of improving the dielectric properties of polymers through surface engineering, enabling their application at concurrent electrical and thermal extremes.

关键词：

Spacetime meshing with adaptive refinement and coarsening

学校读者我要写书评

暂无评论

Spacetime meshing with adaptive refinement and coarsening

Proceedings of the Twentieth Annual Symposium on Computational Geometry (SCG'04)

作者： Abedi, Reza Chung, Shuo-Heng Erickson, Jeff Fan, Yong Garland, Michael Guoy, Damrong Haber, Robert Sullivan, John M. Thite, Shripad Zhou, Yuan Ctr. for Proc. Simulation and Design University of Illinois Urbana-Champaign Department of Theoretical Mechanics Univ. Illinois at Urbana-Champaign Urbana IL 61801 Department of Computer Science UIUC Ctr. for Simulat. of Adv. Rockets Compl. Sci. and Engineering Program UIUC Department of Mathematics UIUC Department of Mathematics Technische Universität Berlin

We propose a new algorithm for constructing finite-element meshes suitable for spacetime discontinuous Galerkin solutions of linear hyperbolic PDEs. Given a triangular mesh of some planar domain Ω and a target time value T, our method constructs a tetrahedral mesh of the spacetime domain Ω × [0, T] in constant running time per tetrahedron in 3 using an advancing front method. Elements are added to the evolving mesh in small patches by moving a vertex of the front forward in time. Spacetime discontinuous Galerkin methods allow the numerical solution within each patch to be computed as soon as the patch is created. Our algorithm employs new mechanisms for adaptively coarsening and refining the front in response to a posteriori error estimates returned by the numerical code. A change in the front induces a corresponding refinement or coarsening of future elements in the spacetime mesh. Our algorithm adapts the duration of each element to the local quality, feature size, and degree of refinement of the underlying space mesh. We directly exploit the ability of discontinuous Galerkin methods to accommodate discontinuities in the solution fields across element boundaries.

关键词： computer graphics

Under-confidence Backdoors Are Resilient and Stealthy Backdoors

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Peng, Minlong Xiong, Zidi Nguyen, Quang H. Sun, Mingming Doan, Khoa D. Li, Ping Cognitive Computing Lab Baidu Research No.10 Xibeiwang East Road Beijing100193 China 10900 NE 8th St. Bellevue Washington98004 United States College of Engineering and Computer Science VinUniversity Viet Nam

Backdoor attacks aim to manipulate the victim model into producing specific outputs on any input injected with pre-designed triggers. Existing dirty-label backdoor attacks, despite showing high attack efficiency, suffer from the overconfidence problem: the victim models will behave peculiarly when the backdoor is implanted. Built upon this characteristic, existing defense algorithms are highly effective in mitigating the risks of these attacks. This work proposes a novel method to overcome this over-confidence problem in the existing attacks and accordingly, increase their stealthiness against the existing defense algorithms relying on this over-confidence characteristic. The rationale of our method is to reduce the backdoor effect in such a way that the victim model will only predict the poisoned input as the target class with a probability slightly greater than those of the other classes. This is achieved by re-designing the label-changing strategy of dirty-label backdoor attacks: the label of a poisoned input x will be changed to the target class with an instance-specific probability of pt(x), instead of 100% as in previous methods. We empirically show that our method considerably improves the stealthiness of several representative backdoor attacks against the defense algorithms on four benchmark datasets in the backdoor domain. The results encourage backdoor researchers to develop defensive countermeasures to mitigate this type of attack. © 2022, CC BY.

关键词：

Development of Automated Detection Methods for Calculating Blood Vessel Width and Stenosis Rate in Cardiac Catheterization Imaging 24

学校读者我要写书评

暂无评论

Development of Automated Detection Methods for Calculating B...

8th International Conference on Medical and Health Informatics, ICMHI 2024

作者： Huang, Hsien-Cheng Chang, Hsiao-Huang Ke, Xiang-Wei Chien, Ting-Ying Department of Computer Science and Engineering Yuan Ze University Taoyuan City Taiwan Cardiovascular Surgery Department Taipei Veterans General Hospital Taipei City Taiwan Taipei Medical University Taipei City Taiwan Graduate Program in Biomedical Informatics Yuan Ze University Taoyuan City Taiwan Innovation Center for Big Data and Digital Convergence Yuan Ze University Taoyuan City Taiwan

ISBN: (纸本)9798400716874

Coronary artery disease is the most prevalent type of heart disease and has a considerable mortality rate. In diagnosing and assessing coronary artery disease, physicians must integrate a variety of clinical information and test results to evaluate the degree of coronary artery obstruction and continually learn and improve their skills in clinical practice to ensure the most accurate assessment of patients. Currently, the use of information technology to assist medical staff allows more effective and accurate analysis of symptoms. This approach not only serves as an auxiliary standard for physician diagnosis, thereby reducing the possibility of misdiagnosis, but also significantly improves patient treatment outcomes and medical quality while reducing the time burden on physicians for diagnosing patients. This study utilized image processing techniques and thinning algorithms to automate the detection of coronary artery width and, following the recommended methods of physicians in related fields, used specific vessel widths to calculate the degree of coronary artery narrowing. This provides a more accurate and efficient assessment method for evaluating the degree of vascular constriction in patients with coronary artery disease. © 2024 Copyright ACM.

关键词： Patient treatment

Synthesis and properties of non-curing graphene thermal interface materials

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Naghibi, Sahar Kargar, Fariborz Wright, Dylan Huang, Chun Yu Tammy Mohammadzadeh, Amirmahdi Barani, Zahra Salgado, Ruben Balandin, Alexander Center University of California RiversideCA92521 United States Materials Science and Engineering Program Bourns College of Engineering University of California Riverside92521 United States Department of Electrical and Computer Engineering Bourns College of Engineering University of California RiversideCA92521 United States

Development of the next generation thermal interface materials with high thermal conductivity is important for thermal management and packaging of electronic devices. We report on the synthesis and thermal conductivity measurements of non-curing thermal paste, i.e. grease, based on mineral oil with the mixture of graphene and few-layer graphene flakes as the fillers. It was found that graphene thermal paste exhibits a distinctive thermal percolation threshold with the thermal conductivity revealing a sublinear dependence on the filler loading. This behavior contrasts with the thermal conductivity of curing graphene thermal interface materials, based on epoxy, where super-linear dependence on the filler loading is observed. The performance of graphene thermal paste was benchmarked against top-of-the-line commercial thermal pastes. The obtained results show that non-curing graphene thermal interface materials outperforms the best commercial pastes in terms of thermal conductivity, at substantially lower filler concentration of fg = 27 vol%. The obtained results shed light on thermal percolation mechanism in non-curing polymeric matrices laden with quasi-two-dimensional fillers. Considering recent progress in graphene production via liquid phase exfoliation and oxide reduction, we argue that our results open a pathway for large-scale industrial application of graphene in thermal management of electronics. Copyright © 2019, The Authors. All rights reserved.

关键词： Percolation (fluids)

Machine learning identification of organic compounds using visible light

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Bikku, Thulasi Fritz, Rubén A. Colón, Yamil J. Herrera, Felipe Department of Physics Universidad de Santiago de Chile Av. Victor Jara 3493 Santiago Chile Computer Science and Engineering Vignan's Nirula Institute of Technology and Science for Women Andhra Pradesh Guntur India Department of Chemical and Biomolecular Engineering University of Notre Dame IN United States ANID - Millennium Science Initiative Program Millennium Institute for Research in Optics Concepción Chile

Identifying chemical compounds is essential in several areas of science and engineering. Laser-based techniques are promising for autonomous compound detection because the optical response of materials encodes enough electronic and vibrational information for remote chemical identification. This has been exploited using the fingerprint region of infrared absorption spectra, which involves a dense set of absorption peaks that are unique to individual molecules, thus facilitating chemical identification. However, optical identification using visible light has not been realized. Using decades of experimental refractive index data in the scientific literature of pure organic compounds and polymers over a broad range of frequencies from the ultraviolet to the far-infrared, we develop a machine learning classifier that can accurately identify organic species based on a single-wavelength dispersive measurement in the visible spectral region, away from absorption resonances. The optical classifier proposed here could be applied to autonomous material identification protocols or applications. Copyright © 2022, The Authors. All rights reserved.

关键词： Refractive index

On Performance Prediction of Big Data Transfer in High-performance Networks

学校读者我要写书评

暂无评论

On Performance Prediction of Big Data Transfer in High-perfo...

IEEE International Conference on Communications (ICC)

作者： Wuji Liu Daqing Yun Chase Q. Wu Nageswara S.V. Rao Aiqin Hou Wei Shen Department of Computer Science New Jersey Institute of Technology Newark NJ USA Computer and Information Sciences Program Harrisburg University Harrisburg PA USA Computational Sciences and Engineering Division Oak Ridge National Laboratory Oak Ridge TN USA School of Information Science and Technology Northwest University Xi’an Shaanxi China School of Informatics Science and Technology Zhejiang Sci-Tech University Hangzhou Zhejiang China

ISBN: (数字)9781728150895

ISBN: (纸本)9781728150901

Big data generated by large-scale scientific and industrial applications need to be transferred between different geographical locations for remote storage, processing, and analysis. High-speed dedicated connections provisioned in High-performance Networks (HPNs) are increasingly utilized to carry out such big data transfer. HPN management highly relies on an important capability of performance (mainly throughput) prediction to reserve sufficient bandwidth and meanwhile avoid over-provisioning that may result in unnecessary resource waste. This capability is critical to improving the resource (mainly bandwidth) utilization of dedicated connections and meeting various user requests for data transfer. Conventional methods conduct performance prediction by fitting prior observed transfer history with predefined loss functions, without considering unobservable latent factors such as competing loads on end hosts. Such latent factors also have a significant impact on the application-level data transfer performance, which may result in an inaccurate prediction model. In this paper, we first investigate the impact of latent factors and propose a clustering-based method to eliminate their negative impact on performance prediction. We then develop a robust machine learning-based performance predictor by: i) incorporating the proposed latent factor elimination method into data preprocessing, and ii) adopting a customized domain guided loss function. Extensive experimental results show that our predictor achieves significantly higher prediction accuracy than several other state-of-the-art methods.

关键词： Data transfer Big Data Throughput Bandwidth Measurement Predictive models Protocols