检索结果-内蒙古大学图书馆

Poster: Selection of Optimal Neural Model using Spiking Neural Network for Edge Computing*

学校读者我要写书评

暂无评论

Poster: Selection of Optimal Neural Model using Spiking Neur...

International Conference on Distributed Computing Systems

作者： Sanaullah Kaushik Roy Ulrich Ruckert Thorsten Jungeblut Dep. of Eng. and Math. Industrial IOT University of Applied Sciences and Arts Bielefeld Germany Dep. of Computer Science Cyber Defense and AI Lab Lab North Carolina A&T State University NC USA Dep. of Faculty of Technology Cognitive Interaction Technology Universitat Bielefeld Bielefeld Germany

ISBN: (数字)9798350386059

ISBN: (纸本)9798350386066

Spiking Neural Networks (SNNs), are inspired by the biological brain's complicated signaling mechanisms and possess unique characteristics that set them apart from traditional artificial neural networks. This research study explores the challenging domain of image classification, specifically utilizing the well-known MNIST dataset through the development and thorough evaluation of different neural models for edge computing. However, the primary contribution is the autonomous selection of the best-performing SNN model through various early stopping approaches and validation functions, allowing the models to autonomously adapt during training. In addition, this article presents the standalone AutoML-SNN model, which is the introduction of dynamic elements into selected SNN domains, enhancing their adaptability to complex patterns within the dataset. Furthermore, the early stopping methodologies are used to reduce overfitting hazards, and using the 3000-neuron set, the LIF appeared as the most proficient neural model.

关键词： Training Adaptation models Computational modeling Biological system modeling Neurons Spiking neural networks Hazards

Qudit Dynamical Decoupling on a Superconducting Quantum Processor

学校读者我要写书评

暂无评论

Physical Review Letters 2025年第5期134卷 050601-050601页

作者： Vinay Tripathi Noah Goss Arian Vezvaee Long B. Nguyen Irfan Siddiqi Daniel A. Lidar Department of Physics and Astronomy University of Southern California Los Angeles California 90089 USA Center for Quantum Information Science and Technology University of Southern California Los Angeles California 90089 USA Quantum Nanoelectronics Laboratory Department of Physics University of California at Berkeley Berkeley California 94720 USA Applied Math and Computational Research Division Lawrence Berkeley National Lab Berkeley California 94720 USA Department of Electrical and Computer Engineering University of Southern California Los Angeles California 90089 USA Department of Chemistry University of Southern California Los Angeles California 90089 USA

Multilevel qudit systems are increasingly being explored as alternatives to traditional qubit systems due to their denser information storage and processing potential. However, qudits are more susceptible to decoherence than qubits due to increased loss channels, noise sensitivity, and crosstalk. To address these challenges, we develop protocols for dynamical decoupling (DD) of qudit systems based on the Heisenberg-Weyl group. We implement and experimentally verify these DD protocols on a superconducting transmon processor that supports qudit operation based on qutrits (d=3) and ququarts (d=4). Specifically, we demonstrate single-qudit DD sequences to decouple qutrits and ququarts from system-bath-induced decoherence. We also introduce two-qudit DD sequences designed to suppress the detrimental cross-Kerr couplings between coupled qudits. This allows us to demonstrate a significant improvement in the fidelity of time-evolved qutrit Bell states. Our results highlight the utility of leveraging DD to enable scalable qudit-based quantum computing.

关键词： Qubits

Enhancing Industrial Productivity through AI-Driven Systematic Literature Reviews 19

学校读者我要写书评

暂无评论

Enhancing Industrial Productivity through AI-Driven Systemat...

19th International Conference on Web Information Systems and Technologies, WEBIST 2023

作者： Coelho, Jaqueline Gutierri Bispo, Guilherme Dantas Vergara, Guilherme Fay Saiki, Gabriela Mayumi Serrano, André Luiz Marques Weigang, Li Neumann, Clovis Martins, Patricia Helena de Oliveira, Welber Santos Albarello, Angela Brigida Casonatto, Ricardo Accorsi Missel, Patrícia de Medeiros Junior, Roberto de Oliveira Gomes, Jefferson Rosano-Peña, Carlos da Costa, Caroline Cabral F. Latitude Decision Making Technologies Lab University of Brasilia Campus Darcy Ribeiro Brasilia Brazil Projectum Research Group for Math. Methodologies Applied to Management University of Brasilia Brasilia Brazil SENAI National Service of Industrial Learning Federal District Brazil Department of Production Engineering Faculty of Technology University of Brasília Federal District Brazil Department of Computer Science University of Brasilia Federal District Brazil Department of Economics University of Brasília Federal District Brazil UNITEC Innovation and Technology Unit National Service of Industrial Learning Federal District Brazil Department of Electrical University of Brasilia Federal District Brazil Department of Administration University of Brasilia Federal District Brazil

ISBN: (纸本)9789897586729

The advent of Artificial Intelligence (AI) has opened up new possibilities for improving productivity in various industry sectors. In this paper, we propose a novel framework aimed at optimizing systematic literature reviews (SLRs) for industrial productivity. By combining traditional keyword selection methods with AI-driven classification techniques, we streamline the review process, making it more efficient. Leveraging advanced natural language processing (NLP) approaches, we identify six key sectors for optimization, thereby reducing workload in less relevant areas and enhancing the efficiency of SLRs. This approach helps conserve valuable time and resources in scientific research. Additionally, we implemented four machine learning models for category classification, achieving an impressive accuracy rate of over 75%. The results of our analyses demonstrate a promising pathway for future automation and refinements to boost productivity in the industry. Copyright © 2023 by SCITEPRESS - science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)

关键词： Natural language processing systems

A Stochastic Covariance Shrinkage Approach in Ensemble Transform Kalman Filtering

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Popov, Andrey A. Sandu, Adrian Nino-Ruiz, Elias D. Evensen, Geir Computational Science Laboratory Department of Computer Science Virginia Tech United States Applied Math and Computer Science Lab Universidad del Norte Colombia Norway

The Ensemble Kalman Filters (EnKF) employ a Monte-Carlo approach to represent covariance information, and are affected by sampling errors in operational settings where the number of model realizations is much smaller than the model state dimension. To alleviate the effects of these errors EnKF relies on model-specific heuristics such as covariance localization, which takes advantage of the spatial locality of correlations among the model variables. This work proposes an approach to alleviate sampling errors that utilizes a locally averaged-in-time dynamics of the model, described in terms of a climatological covariance of the dynamical system. We use this covariance as the target matrix in covariance shrinkage methods, and develop a stochastic covariance shrinkage approach where synthetic ensemble members are drawn to enrich both the ensemble subspace and the ensemble transformation. We additionally provide for a way in which this methodology can be localized similar to the state-of-the-art LETKF method, and that for a certain model setup, our methodology significantly outperforms it. Copyright © 2020, The Authors. All rights reserved.

关键词： Dynamical systems

HiPoNet: A Topology-Preserving Multi-View Neural Network For High Dimensional Point Cloud and Single-Cell Data

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Viswanath, Siddharth Madhu, Hiren Bhaskar, Dhananjay Kovalic, Jake Johnson, Dave Ying, Rex Tape, Christopher Adelstein, Ian Perlmutter, Michael Krishnaswamy, Smita Department of Computer Science Yale University New Haven United States Department of Genetics Yale University New Haven United States Department of Applied Physics Yale University New Haven United States Program in Computing Boise State University Idaho United States Cell Communication Lab Department of Oncology University College London Cancer Institute London United Kingdom Department of Mathematics Yale University New Haven United States Department of Mathematics Boise State University Idaho United States Computational Biology and Bioinformatics Program Yale University New Haven United States Wu-Tsai Institute Yale University New Haven United States Program for Applied Math Yale University New Haven United States

In this paper, we propose HiPoNet, an end-to-end differentiable neural network for regression, classification, and representation learning on high-dimensional point clouds. Single-cell data can have high dimensionality exceeding the capabilities of existing methods for point clouds which are tailored for 3D data. Moreover, modern single-cell and spatial experiments now yield entire cohorts of datasets (i.e. one on every patient), necessitating models that can process large, high-dimensional point clouds at scale. Most current approaches build a single nearest-neighbor graph, discarding important geometric information. In contrast, HiPoNet forms higher-order simplicial complexes through learnable feature reweighting, generating multiple data views that disentangle distinct biological processes. It then employs simplicial wavelet transforms to extract multiscale features—capturing both local and global topology. We empirically show that these components preserve topological information in the learned representations, and that HiPoNet significantly outperforms state-of-the-art point-cloud and graph-based models on single cell. We also show an application of HiPoNet on spatial transcriptomics datasets using spatial co-ordinates as one of the views. Overall, HiPoNet offers a robust and scalable solution for high-dimensional data analysis. © 2025, CC BY.

关键词： Network topology

Qudit Dynamical Decoupling on a Superconducting Quantum Processor

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Tripathi, Vinay Goss, Noah Vezvaee, Arian Nguyen, Long B. Siddiqi, Irfan Lidar, Daniel A. Department of Physics & Astronomy University of Southern California Los AngelesCA90089 United States Center for Quantum Information Science & Technology University of Southern California Los AngelesCA90089 United States Quantum Nanoelectronics Laboratory Department of Physics University of California at Berkeley BerkeleyCA94720 United States Applied Math and Computational Research Division Lawrence Berkeley National Lab BerkeleyCA94720 United States Department of Electrical & Computer Engineering University of Southern California Los AngelesCA90089 United States Department of Chemistry University of Southern California Los AngelesCA90089 United States

Multi-level qudit systems are increasingly being explored as alternatives to traditional qubit systems due to their denser information storage and processing potential. However, qudits are more susceptible to decoherence than qubits due to increased loss channels, noise sensitivity, and crosstalk. To address these challenges, we develop protocols for dynamical decoupling (DD) of qudit systems based on the Heisenberg-Weyl group. We implement and experimentally verify these DD protocols on a superconducting transmon processor that supports qudit operation based on qutrits (d = 3) and ququarts (d = 4). Specifically, we demonstrate single-qudit DD sequences to decouple qutrits and ququarts from system-bath-induced decoherence. We also introduce two-qudit DD sequences designed to suppress the detrimental cross-Kerr couplings between coupled qudits. This allows us to demonstrate a significant improvement in the fidelity of time-evolved qutrit Bell states. Our results highlight the utility of leveraging DD to enable scalable qudit-based quantum computing. © 2024, CC BY.

关键词： Qubits

InGAN: Capturing and Retargeting the “DNA” of a Natural Image

学校读者我要写书评

暂无评论

InGAN: Capturing and Retargeting the “DNA” of a Natural Im...

International Conference on computer Vision (ICCV)

作者： Assaf Shocher Shai Bagon Phillip Isola Michal Irani Dept. of Computer Science and Applied Math The Weizmann Institute of Science Weizmann Artificial Intelligence Center (WAIC) Computer Science and Artificial Intelligence Lab Massachusetts Institute of Technology

ISBN: (数字)9781728148038

ISBN: (纸本)9781728148045

Generative Adversarial Networks (GANs) typically learn a distribution of images in a large image dataset, and are then able to generate new images from this distribution. However, each natural image has its own internal statistics, captured by its unique distribution of patches. In this paper we propose an "Internal GAN'' (InGAN) - an image-specific GAN - which trains on a single input image and learns its internal distribution of patches. It is then able to synthesize a plethora of new natural images of significantly different sizes, shapes and aspect-ratios - all with the same internal patch-distribution (same "DNA'') as the input image. In particular, despite large changes in global size/shape of the image, all elements inside the image maintain their local size/shape. InGAN is fully unsupervised, requiring no additional data other than the input image itself. Once trained on the input image, it can remap the input to any size or shape in a single feedforward pass, while preserving the same internal patch distribution. InGAN provides a unified framework for a variety of tasks, bridging the gap between textures and natural images.

关键词： Gallium nitride Shape Generators Task analysis DNA Visualization Image reconstruction

InGAN: Capturing and remapping the "DNA" of a natural image

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Shocher, Assaf Bagon, Shai Isola, Phillip Irani, Michal Dept. of Computer Science and Applied Math Weizmann Institute of Science Computer Science and Artificial Intelligence Lab Massachusetts Institute of Technology

Generative Adversarial Networks (GANs) typically learn a distribution of images in a large image dataset, and are then able to generate new images from this distribution. However, each natural image has its own internal statistics, captured by its unique distribution of patches. In this paper we propose an "Internal GAN" (InGAN) – an image-specific GAN – which trains on a single input image and learns its internal distribution of patches. It is then able to synthesize a plethora of new natural images of significantly different sizes, shapes and aspect-ratios all with the same internal patch-distribution (same "DNA") as the input image. In particular, despite large changes in global size/shape of the image, all elements inside the image maintain their local size/shape. InGAN is fully unsupervised, requiring no additional data other than the input image itself. Once trained on the input image, it can remap the input to any size or shape in a single feedforward pass, while preserving the same internal patch distribution. InGAN provides a unified framework for a variety of tasks, bridging the gap between textures and natural images.1 Copyright © 2018, The Authors. All rights reserved.

关键词： Generative adversarial networks

An empirical study on the knowledge management practice in software testing 21

学校读者我要写书评

暂无评论

An empirical study on the knowledge management practice in s...

21st Ibero-American Conference on Software Engineering, CIbSE 2018

作者： Maciel, Claudia P.C. De Souza, Erica Ferreira Vijaykumar, Nandamudi L. De Almeida Falbo, Ricardo Meinerz, Giovani Volnei Felizardo, Katia Romero Department of Computer Federal Technological University of Paraná/UTFPR Cornelio Procópio/PR Brazil Lab. of Comp and Applied Math. National Institute for Space Research INPE Sao José dos Campos/SP Brazil Department of Computer Science Federal University of Espirito Santo UFES Vitória/ES Brazil

ISBN: (纸本)9789584837547

Context: Software testing is a knowledge intensive process, and therefore can benefit from the use of experience gained from past projects. In this context, principles of Knowledge Management (KM) can be applied to promote knowledge capture and sharing as well as the emergence of new knowledge. In spite of this, there are only few studies that present practitioners' perspective about KM initiatives in software testing. Objective: The main goal of this study is to identify the perception of professionals in software testing on the use of KM initiatives in software engineering companies, such as KM adoption, potential benefits or hindrances. Method: A survey was conducted in software development companies. In addition, the survey results were compared with another survey conducted previously in the same context, however, with different objectives and target audience. 39 software companies participated in this research. Results: Testing planning activity and the test case reuse are the aspects that have received more attention by companies. A strategic planning and reuse of existing test cases have a potential to significantly reduce software development costs and time. Conclusion: Applying KM in software companies can bring several benefits in terms of quality of results, reduction costs, time and effort. However, it is still a challenge to be faced and tackled. © 2018 Conferencia Iberoamericana de Software Engineering - CIbSE. All Rights Reserved.

关键词： Software testing