检索结果-内蒙古大学图书馆

Challenges in systematic reviews and meta-analyses of mediation analyses

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Vo, Tat-Thang Vansteelandt, Stijn Department of Statistics The Wharton School University of Pennsylvania United States Department of Applied Mathematics Computer Science and Statistics Ghent University Belgium Department of Medical Statistics London School of Hygiene and Tropical Medicine United Kingdom

Systematic reviews and meta-analyses of mediation studies are increasingly being implemented in practice. Nonetheless, the methodology for conducting such review and analysis is still in a development phase, with much room for improvement. In this paper, we highlight and discuss challenges that investigators face in mediation systematic reviews and meta-analyses, then propose ways of accommodating these in practice. © 2021, CC BY-NC-ND.

关键词：

Sequential neural posterior and likelihood approximation

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Wiqvist, Samuel Frellsen, Jes Picchini, Umberto Centre for Mathematical Sciences Lund University Lund Sweden Department of Applied Mathematics and Computer Science Technical University of Denmark Denmark Department of Mathematical Sciences Chalmers University of Technology The University of Gothenburg Gothenburg Sweden

We introduce the sequential neural posterior and likelihood approximation (SNPLA) algorithm. SNPLA is a normalizing flows-based algorithm for inference in implicit models, and therefore is a simulation-based inference method that only requires simulations from a generative model. SNPLA avoids Markov chain Monte Carlo sampling and correction-steps of the parameter proposal function that are introduced in similar methods, but that can be numerically unstable or restrictive. By utilizing the reverse KL divergence, SNPLA manages to learn both the likelihood and the posterior in a sequential manner. Over four experiments, we show that SNPLA performs competitively when utilizing the same number of model simulations as used in other methods, even though the inference problem for SNPLA is more complex due to the joint learning of posterior and likelihood function. Due to utilizing normalizing flows SNPLA generates posterior draws much faster (4 orders of magnitude) than MCMC-based methods. Copyright © 2021, The Authors. All rights reserved.

关键词： Approximation algorithms

Twisted Reed-Solomon Codes

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Beelen, Peter Puchinger, Sven Rosenkilde, Johan Department of Applied Mathematics and Computer Science Technical University of Denmark Lyngby Kongens 2800 Denmark Hensoldt Sensors GmbH Ulm89077 Germany GitHub Denmark Aps Copenhagen2100 Denmark

In this article, we present a new construction of evaluation codes in the Hamming metric, which we call twisted Reed-Solomon codes. Whereas Reed-Solomon (RS) codes are MDS codes, this need not be the case for twisted RS codes. Nonetheless, we show that our construction yields several families of MDS codes. Further, for a large subclass of (MDS) twisted RS codes, we show that the new codes are not generalized RS codes. To achieve this, we use properties of Schur squares of codes as well as an explicit description of the dual of a large subclass of our codes. We conclude the paper with a description of a decoder, that performs very well in practice as shown by extensive simulation results. Copyright © 2021, The Authors. All rights reserved.

关键词： Reed-Solomon codes

Dielectric metasurfaces for complete and independent control of the optical amplitude and phase

学校读者我要写书评

暂无评论

Light(science & Applications) 2019年第1期8卷 318-329页

作者： Adam C.Overvig Sajan Shrestha Stephanie C.Malek Ming Lu Aaron Stein Changxi Zheng Nanfang Yu Department of Applied Physics and Applied Mathematics Columbia UniversityNew YorkNY 10027USA Center for Functional Nanomaterials Brookhaven National LaboratoryUptonNY 11973USA Department of Computer Science Columbia UniversityNew YorkNY 10027USA

Metasurfaces are optically thin metamaterials that promise complete control of the wavefront of light but are primarily used to control only the phase of ***,we present an approach,simple in concept and in practice,that uses meta-atoms with a varying degree of form birefringence and rotation angles to create high-efficiency dielectric metasurfaces that control both the optical amplitude and phase at one or two *** opens up applications in computer-generated holography,allowing faithful reproduction of both the phase and amplitude of a target holographic scene without the iterative algorithms required in phase-only *** demonstrate all-dielectric metasurface holograms with independent and complete control of the amplitude and phase at up to two optical frequencies simultaneously to generate two-and three-dimensional holographic *** show that phaseamplitude metasurfaces enable a few features not attainable in phase-only holography;these include creating artifactfree two-dimensional holographic images,encoding phase and amplitude profiles separately at the object plane,encoding intensity profiles at the metasurface and object planes separately,and controlling the surface textures of three-dimensional holographic objects.

关键词： holographic phase complete

Characterizing the hyperuniformity of ordered and disordered two-phase media

学校读者我要写书评

暂无评论

Physical Review E 2021年第1期103卷 012123-012123页

作者： Jaeuk Kim Salvatore Torquato Department of Chemistry Princeton University Princeton New Jersey 08544 USA Department of Physics Princeton University Princeton New Jersey 08544 USA Princeton Institute for the Science and Technology of Materials Princeton University Princeton New Jersey 08544 USA and Program in Applied and Computational Mathematics Princeton University Princeton New Jersey 08544 USA

The hyperuniformity concept provides a unified means to classify all perfect crystals, perfect quasicrystals, and exotic amorphous states of matter according to their capacity to suppress large-scale density fluctuations. While the classification of hyperuniform point configurations has received considerable attention, much less is known about the classification of hyperuniform two-phase heterogeneous media, which include composites, porous media, foams, cellular solids, colloidal suspensions, and polymer blends. The purpose of this article is to begin such a program for certain two-dimensional models of hyperuniform two-phase media by ascertaining their local volume-fraction variances σV2(R) and the associated hyperuniformity order metrics B¯V. This is a highly challenging task because the geometries and topologies of the phases are generally much richer and more complex than point-configuration arrangements, and one must ascertain a broadly applicable length scale to make key quantities dimensionless. Therefore, we purposely restrict ourselves to a certain class of two-dimensional periodic cellular networks as well as periodic and disordered or irregular packings of circular disks, some of which maximize their effective transport and elastic properties. Among the cellular networks considered, the honeycomb networks have minimal values of the hyperuniformity order metrics B¯V across all volume fractions. On the other hand, for all packings of circular disks examined, the triangular-lattice packings have the smallest values of B¯V for the possible range of volume fractions. Among all structures studied here, the triangular-lattice packing of circular disks have the minimal values of the order metric for almost all volume fractions. Our study provides a theoretical foundation for the establishment of hyperuniformity order metrics for general two-phase media and a basis to discover new hyperuniform two-phase systems with desirable bulk physical properties by inverse d

关键词： Patterns in complex systems

DRG-Net: Interactive Joint Learning of Multi-lesion Segmentation and Classification for Diabetic Retinopathy Grading

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Tusfiqur, Hasan Md Nguyen, Duy M.H. Truong, Mai T.N. Nguyen, Triet A. Nguyen, Binh T. Barz, Michael Profitlich, Hans-Jürgen Than, Ngoc T.T. Le, Ngan Xie, Pengtao Sonntag, Daniel Germany Machine Learning and Simulation Lab University of Stuttgart Germany Department of Multimedia Engineering Dongguk University Korea Republic of Department of Mathematics University of Architecture Ho Chi Minh City Viet Nam AISIA Lab University of Science VNUHCM Viet Nam Byers Eye Institute Stanford University United States Department of Computer Science and Computer Engineering University of Arkansas United States Department of Electrical and Computer Engineering University of California San Diego United States Applied Artificial Intelligence Oldenburg University Germany

Diabetic Retinopathy (DR) is a leading cause of vision loss in the world, and early DR detection is necessary to prevent vision loss and support an appropriate treatment. In this work, we leverage interactive machine learning and introduce a joint learning framework, termed DRG-Net, to effectively learn both disease grading and multi-lesion segmentation. Our DRG-Net consists of two modules: (i) DRG-AI-System to classify DR Grading, localize lesion areas and provide visual explanation and (ii) DRG-Expert-Interaction to receive feedback from user-expert and improve the DRG-AI-System. To deal with sparse data, we utilize transfer learning mechanisms to extract invariant feature representations by using Wasserstein distance and adversarial learning-based entropy minimization. Besides, we propose a novel attention strategy at both low- and high-level features to automatically select the most significant lesion information and provide explainable properties. In terms of human interaction, we further develop DRG-Net as a tool that enables expert users to correct the system’s predictions, which may then be used to update the system as a whole. Moreover, thanks to the attention mechanism and loss functions constraint between lesion features and classification features, our approach can be robust given a certain level of noise in the feedback of users. We have benchmarked DRG-Net on the two largest DR datasets, i.e., IDRID and FGADR, and compared it to various state-of-the-art deep learning networks. In addition to outperforming other SOTA approaches, DRG-Net is effectively updated using user feedback, even in a weakly-supervised manner. © 2022, CC BY.

关键词： Grading

A Quantum Computing View on Unitary Coupled Cluster Theory

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Anand, Abhinav Schleich, Philipp Alperin-Lea, Sumner Jensen, Phillip W. K. Sim, Sukin Díaz-Tinoco, Manuel Kottmann, Jakob S. Degroote, Matthias Izmaylov, Artur F. Aspuru-Guzik, Alán Chemical Physics Theory Group Department of Chemistry University of Toronto Canada Department of Computer Science University of Toronto Canada Applied and Computational Mathematics Department of Mathematics RWTH Aachen University Aachen Germany Vector Institute for Artificial Intelligence Toronto Canada Department of Chemistry and Chemical Biology Harvard University United States Department of Physical and Environmental Sciences University of Toronto Scarborough Canada Lebovic Fellow Toronto Canada

We present a review of the Unitary Coupled Cluster (UCC) ansatz and related ansätze which are used to variationally solve the electronic structure problem on quantum computers. A brief history of coupled cluster (CC) methods is provided, followed by a broad discussion of the formulation of CC theory. This includes touching on the merits and difficulties of the method and several variants, UCC among them, in the classical context, to motivate their applications on quantum computers. In the core of the text, the UCC ansatz and its implementation on a quantum computer are discussed at length, in addition to a discussion on several derived and related ansätze specific to quantum computing. The review concludes with a unified perspective on the discussed ansätze, attempting to bring them under a common framework, as well as with a reflection upon open problems within the field. © 2021, CC BY.

关键词： Cluster analysis

How we can use text classification in the Back-Office environment of a bank as ‘business as usual’ solution

学校读者我要写书评

暂无评论

How we can use text classification in the Back-Office enviro...

International Symposium on applied computational Intelligence and Informatics ( SACI)

作者： Zsolt Krutilla Attila Kovari Doctoral School of Applied Informatics and Applied Mathematics Óbuda University Budapest Hungary Informatics Institution University of Dunaújváros Dunaújváros Hungary Institute of Engineering Alba Regia Technical Faculty Óbuda University Budapest Hungary Institute of Digital Technology Faculty of Informatics Eszterházy Károly Catholic University Eger Hungary Department of Natural Science Institute of Engineering University of Dunaújváros Dunaújváros Hungary GAMF Faculty of Engineering and Computer Science John von Neumann University Kecskemét Hungary

Natural Language Processing nowadays provides scientists with many research areas and opportunities, but as with most applied sciences, our goal in Natural Language Processing is to refine the underlying science and technology until it can be used reliably for business purposes. Transformer models, such as GPT or BERT, are currently showing outstanding results in the field of natural-language processing, but they require huge computational power and data to teach, but these conditions can only be met by larger research centers and large companies, and their inaccuracy makes them unsuitable for use as a ‘business as usual’ (BAU) solution. In this paper, we present a solution that is able to overcome these problems by focusing on accuracy and usability, and that can also bring a new perspective to the process of teaching deep learning models.

关键词：

Effect of the Stark shift on the low-energy interference structure in strong-field ionization

学校读者我要写书评

暂无评论

Physical Review A 2021年第5期103卷 053105-053105页

作者： Weifeng Yang Jie Li Wenbin Jia Hongdan Zhang Xiwang Liu Ming Zhu Xiaohong Song Jing Chen Research Center for Advanced Optics and Photoelectronics Department of Physics College of Science Shantou University Shantou Guangdong 515063 China Institute of Mathematics Shantou University Shantou Guangdong 515063 China MOE Key Laboratory of Intelligent Manufacturing Technology Shantou University Shantou Guangdong 515063 China HEDPS Center for Applied Physics and Technology Peking University Beijing 100084 China Institute of Applied Physics and Computational Mathematics P.O. Box 8009 Beijing 100088 China

An improved quantum trajectory Monte Carlo method including the Stark shift of the initial state, Coulomb potential, and multielectron polarization-induced dipole potential is adopted to revisit the origin of the low-energy interference structure in the photoelectron momentum distribution of the xenon atom subjected to an intense laser field, and resolve the different contributions of these three effects. We found that the Stark shift plays an essential role on the low-energy interference structure, which moves the ringlike constructive interference structure to the lower momentum region. The formation of the low-energy interference structure is a result of the combined effects of Stark shift, laser, and Coulomb fields, while the multielectron polarization mainly enhance the probability of the low energy photoelectron spectrum. Our finding provides insight into the electron dynamics of atoms and molecules when driven by the intense laser fields.

关键词： Atomic & molecular processes in external fields Multiphoton or tunneling ionization & excitation