检索结果-内蒙古大学图书馆

Optical Coherence Tomography Classification Based on Transfer Learning and RA-Attention 11th

学校读者我要写书评

暂无评论

Optical Coherence Tomography Classification Based on Transfe...

11th International Conference on Health Information science, HIS 2022

作者： Lian, Xiaoyi Chen, Lina Ji, Xiayan Shen, Fangyao Guo, Hongjie Gao, Hong College of Physics and Electronic Information Engineering Zhejiang Normal University Zhejiang321004 China College of Mathematics and Computer Science Zhejiang Normal University Zhejiang321004 China Faculty of Computing Harbin Institute of Technology Harbin150001 China

ISBN: (纸本)9783031206269

Macular disease is one of the major causes of blindness. Optical coherence tomography (OCT) is a commonly used ophthalmic diagnostic technique to assist ophthalmologists in their analysis and treatment. However, manual analysis is a time-consuming, heavy and subjective process. In this paper, we propose an improved EfficientNet model for retinal OCT image classification, named TL-RA-EfficientNet. It introduces a Re-Attention module to enable the model to identify the lesions with small areas and fuzzy shapes. During training, a data augmentation strategy is introduced to solve the class-imbalance problem, and a transfer learning strategy is adopted to speed up the convergence of the model. The experimental results on the UCSD dataset show that the average recognition accuracy, sensitivity and specificity of our model are 99.9%, 99.9% and 99.97%, respectively, which are more effective than previous classification methods. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Optical tomography

Formation of Metastable Solid Solutions in Bi – Ge Films at Low-Temperature Treatment

学校读者我要写书评

暂无评论

SSRN

SSRN 2024年

作者： Bogatyrenko, Sergiy Kryshtal, Pavlo Gruszczyński, Adam Kryshtal, Aleksandr Department of Physics and Technology V.N. Karazin Kharkiv National University 4 Svobody sq. Kharkiv61022 Ukraine Faculty of Metals Engineering and Industrial Computer Science AGH University of Krakow Al. A. Mickiewicza 30 KrakówPL-30 059 Poland

AbstractWe investigated the mechanism and kinetics of formation of metastable BiGe solid phases during amorphous to crystalline transformation of Ge films in contact with Bi. Ge/Bi/Ge sandwich films with Bi film between amorphous Ge films, which were fabricated by sequential deposition of components in vacuum, were used in the study. The total thickness and composition of the sandwich films varied in the range from 30 to 400 nm and 22 – 48 wt% Bi, respectively. Electron diffraction, high-resolution (S)TEM imaging, EDX and EEL spectroscopy were used for in situ and ex-situ characterization of morphology, composition, and structure of Ge/Bi/Ge films in a temperature range 20-271°*** proved formation of polycrystalline Ge films containing up to 28 wt% Bi during low temperature treatment. The interaction process was activated at ≈150 °C resulting in crystallization of Ge with simultaneous formation of homogeneous supersaturated solid solution in the whole volume of the film already at ≈210 °C. We showed that the formation of crystalline Ge films with extended solid solubility of Bi depended on the overall composition of the tri-layer film and not on its thickness. The role of the metal-induced crystallization of the amorphous germanium in the formation of the supersaturated solid phases is discussed. © 2024, The Authors. All rights reserved.

关键词： Solubility

Simulating first-order phase transition with hierarchical autoregressive networks

学校读者我要写书评

暂无评论

Physical Review E 2023年第5期107卷 054127-054127页

作者： Piotr Białas Paulina Czarnota Piotr Korcyl Tomasz Stebel Institute of Applied Computer Science Jagiellonian University 30-348 Kraków Poland Faculty of Biochemistry Biophysics and Biotechnology Jagiellonian University 30-387 Kraków Poland Institute of Theoretical Physics Jagiellonian University 30-348 Kraków Poland

We apply the hierarchical autoregressive neural network sampling algorithm to the two-dimensional Q-state Potts model and perform simulations around the phase transition at Q=12. We quantify the performance of the approach in the vicinity of the first-order phase transition and compare it with that of the Wolff cluster algorithm. We find a significant improvement as far as the statistical uncertainty is concerned at a similar numerical effort. In order to efficiently train large neural networks we introduce the technique of pretraining. It allows us to train some neural networks using smaller system sizes and then employ them as starting configurations for larger system sizes. This is possible due to the recursive construction of our hierarchical approach. Our results serve as a demonstration of the performance of the hierarchical approach for systems exhibiting bimodal distributions. Additionally, we provide estimates of the free energy and entropy in the vicinity of the phase transition with statistical uncertainties of the order of 10−7 for the former and 10−3 for the latter based on a statistics of 106 configurations.

关键词： Approximation methods for many-body systems Deep learning Lattice field theory Metropolis algorithm Monte Carlo methods Spin lattice models

Establishing a leader in a pairwise comparisons method

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Szybowski, Jacek Kulakowski, Konrad Mazurek, Jiri Ernst, Sebastian AGH University of Krakow The Faculty of Applied Mathematics al. A. Mickiewicza 30 Krakow30-059 Poland AGH University of Krakow The Department of Applied Computer Science al. A. Mickiewicza 30 Krakow30-059 Poland Silesian University in Opava School of Business Administration in Karvina Univerzitní nám. 1934 Karviná733 40 Czech Republic

Like electoral systems, decision-making methods are also vulnerable to manipulation by decision-makers. The ability to effectively defend against such threats can only come from thoroughly understanding the manipulation mechanisms. In the presented article, we show two algorithms that can be used to launch a manipulation attack. They allow for equating the weights of two selected alternatives in the pairwise comparison method and, consequently, choosing a leader. The theoretical considerations are accompanied by a Monte Carlo simulation showing the relationship between the size of the PC matrix, the degree of inconsistency, and the ease of manipulation. This work is a continuation of our previous research published in the paper [30]. © 2024, CC BY.

关键词： Monte Carlo methods

Algebraic curves as a source of separable multi-Hamiltonian systems

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Blaszak, Maciej Marciniak, Krzysztof Faculty of Physics Department of of Mathematical Physics and Computer Modelling A. Mickiewicz University Poznań61-614 Poland Department of Science and Technology Campus Norrköping Linköping University Norrköping601-74 Sweden

In this paper we systematically consider various ways of generating integrable and separable Hamiltonian systems in canonical and in non-canonical representations from algebraic curves on the plane. In particular, we consider Stäckel transform between two pairs of Stäckel systems, generated by 2n-parameter algebraic curves on the plane, as well as Miura maps between Stäckel systems generated by (n+N)-parameter algebraic curves, leading to multi-Hamiltonian representation of these *** Codes 14H70, 37J35, 70H20 © 2024, CC BY.

关键词： Hamiltonians

Characterization of two electronic subsystems in cuprates through optical conductivity

学校读者我要写书评

暂无评论

Physical Review B 2023年第14期107卷 144515-144515页

作者： C. M. N. Kumar A. Akrap C. C. Homes E. Martino B. Klebel-Knobloch W. Tabis O. S. Barišić D. K. Sunko N. Barišić Institute of Solid State Physics TU Wien 1040 Vienna Austria Institute of Physics Bijenička cesta 46 HR-10000 Zagreb Croatia The Henryk Niewodniczanski Institute of Nuclear Physics Polish Academy of Sciences 152 Radzikowskiego Strasse 31-342 Krakow Poland Department of Physics University of Fribourg Chemin du Musée 3 CH-1700 Fribourg Switzerland Condensed Matter Physics and Materials Science Department Brookhaven National Laboratory Upton New York 11973 USA National Synchrotron Light Source II Brookhaven National Laboratory Upton New York 11973 USA Institute of Physics École Polytechnique Fédéral de Lausanne (EPFL) CH-1015 Lausanne Switzerland AGH University of Science and Technology Faculty of Physics and Applied Computer Science 30-059 Krakow Poland Department of Physics Faculty of Science University of Zagreb Bijenička cesta 32 HR-10000 Zagreb Croatia

Understanding the physical properties of unconventional superconductors as well as of other correlated materials presents a formidable challenge. Their unusual evolution with doping, frequency, and temperature has frequently led to non-Fermi-liquid (non-FL) interpretations. Optical conductivity is a major challenge in this context. Here, the optical spectra of two archetypal cuprates, underdoped HgBa2CuO4+δ and optimally doped Bi2Sr2CaCu2O8+δ, are interpreted based on the standard Fermi-liquid (FL) paradigm. At both dopings, perfect frequency-temperature FL scaling is found to be modified by the presence of a second, gapped electronic subsystem. This non-FL component emerges as a well-defined mid-infrared spectral feature after the FL contribution, determined independently by transport, is subtracted. Temperature, frequency, and doping evolution of the MIR feature identify a gapped rather than dissipative response. In contrast, the dissipative response is found to be relevant for pnictides and ruthenates. Such an unbiased FL/non-FL separation is extended across the cuprate phase diagram, capturing all the key features of the normal state and providing a natural explanation why the superfluid density is attenuated on the overdoped side. Thus, we obtain a unified interpretation of optical responses and transport measurements in all analyzed physical regimes and all analyzed compounds.

关键词： Electrical properties Localization Microphase separation Optical conductivity Percolation Superconducting phase transition Superconductivity Superfluid density

Cartesian Genetic Programming Approach for Designing Convolutional Neural Network

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Krzywda, Maciej Lukasik, Szymon Gandomi, Amir H. Faculty of Physics and Applied Computer Science AGH University of Krakow al. Mickiewicza 30 Kraków30-059 Poland Systems Research Institute Polish Academy of Sciences ul. Newelska 6 Warsaw01-447 Poland Faculty of Engineering and IT University of Technology Sydney 5 Broadway UltimoNSW2007 Australia Óbuda University Bécsi út 96/B Budapest1034 Hungary

The present study covers an approach to neural architecture search (NAS) using Cartesian genetic programming (CGP) for the design and optimization of Convolutional Neural Networks (CNNs). In designing artificial neural networks, one crucial aspect of the innovative approach is suggesting a novel neural architecture. Currently used architectures have mostly been developed manually by human experts, which is a time-consuming and error-prone process. In this work, we use pure Genetic Programming Approach to design CNNs, which employs only one genetic operation, i.e., mutation. In the course of preliminary experiments, our methodology yields promising results. © 2024, CC BY.

关键词： Convolutional neural networks

Spin-orbit torque induced magnetization dynamics and switching in a CoFeB/Ta/CoFeB system with mixed magnetic anisotropy

学校读者我要写书评

暂无评论

Physical Review B 2021年第13期103卷 134421-134421页

作者： Stanisław Łazarski Witold Skowroński Krzysztof Grochot Wiesław Powroźnik Jarosław Kanak Marek Schmidt Tomasz Stobiecki Institute of Electronics AGH University of Science and Technology Al. Mickiewicza 30 30-059 Kraków Poland Faculty of Physics and Applied Computer Science AGH University of Science and Technology Al. Mickiewicza 30 30-059 Kraków Poland Institute of Molecular Physics Polish Academy of Sciences ul. Smoluchowskiego 17 60-179 Poznań Poland

Spin-orbit torque (SOT) induced magnetization switching in CoFeB/Ta/CoFeB trilayer with two ferromagnetic CoFeB layers exhibiting in-plane and perpendicular magnetic anisotropy (PMA), respectively, is investigated. Interlayer exchange coupling (IEC), measured using the ferromagnetic resonance technique is modified by varying thickness of the heavy metal Ta spacer (tTa). The evolution of the IEC with tTa leads to different orientations of the magnetic anisotropy axes of two CoFeB layers: for tTa>1.3nm, magnetization prefers antiferromagnetic alignment, whereas for tTa<1.3nm, ferromagnetic coupling is measured. The magnetization state of the CoFeB layer exhibiting PMA is controlled by the spin-polarized current originating from the Ta layer in micrometer sized Hall bars. The SOT induced critical current density with Ta spacer thickness is presented, showing an increase with decreasing tTa, which coincides with the CoFeB coercive field dependence. In a narrow range of tTa corresponding to the ferromagnetic IEC, the field-free SOT induced magnetization switching is achieved.

关键词： Magnetic anisotropy Magnetization switching Spin Hall effect Spin accumulation Spin torque Spin-orbit coupling Spintronics Magnetic multilayers Ferromagnetic resonance Resistivity measurements X-ray diffraction X-ray reflectivity

IDENTIFYING SUBSTITUTE AND COMPLEMENTARY PRODUCTS FOR ASSORTMENT OPTIMIZATION WITH CLEORA EMBEDDINGS

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Tkachuk, Sergiy Wróblewska, Anna Dabrowski, Jacek Lukasik, Szymon Systems Research Institute Polish Academy of Sciences ul. Newelska 6 Warsaw01-447 Poland Faculty of Mathematics and Information Science Warsaw University of Technology Warsaw Poland Faculty of Physics and Applied Computer Science AGH University of Science and Technology al. Mickiewicza 30 Kraków30-059 Poland Synerise Warsaw Poland

Recent years brought an increasing interest in the application of machine learning algorithms in e-commerce, omnichannel marketing, and the sales industry. It is not only to the algorithmic advances but also to data availability, representing transactions, users, and background product information. Finding products related in different ways, i.e., substitutes and complements is essential for users' recommendations at the vendor's site and for the vendor - to perform efficient assortment optimization. The paper introduces a novel method for finding products' substitutes and complements based on the graph embedding Cleora algorithm. We also provide its experimental evaluation with regards to the state-of-the-art Shopper algorithm, studying the relevance of recommendations with surveys from industry experts. It is concluded that the new approach presented here offers suitable choices of recommended products, requiring a minimal amount of additional information. The algorithm can be used in various enterprises, effectively identifying substitute and complementary product options. Copyright © 2022, The Authors. All rights reserved.

关键词： Electronic commerce