检索结果-内蒙古大学图书馆

Artificial Intelligence and Smart Systems (ICAIS), International Conference on

作者： Neda Firoz Olga Grigorievna Beresteneva Aksyonov Sergey Vladimirovich Mohammad Sadman Tahsin Faiza Tafannum Institute of Applied Mathematics and Computer Science Tomsk State University Tomsk Russia Department of I.T (Institute of Applied Mathematics and Computer Sciences) Tomsk Polytechnic University Tomsk Russia Department of Computer Sciences & Engineering BRAC University Dhaka Bangladesh

Depression and its symptoms are very common disorders of mental health. They affect the day-to-day activity of the person and degrade the quality of life. The article presents the comparative study of different deep learning models on natural language processing data for detection of depression using textual data. Several studies have been performed for depression detection using artificial intelligence and deep learning state of the art methods. This article investigates the state-of-the-art models and perform hyperparameter tuning for best accuracy results and develop our own hybrid model for detection of depression with improved accuracy scores. The aim of our study is to research and compare the existing findings in deep learning and machine learning models for depression detection and build a precise hybrid model for depression detection with higher accuracy and scores.

关键词： Deep learning Buildings Mental health Predictive models Depression Natural language processing Hybrid power systems

来源：评论

学校读者我要写书评

暂无评论

Upper bounds on minimum size of feedback arc set of directed multigraphs with bounded degree

arXiv

引用

arXiv 2024年

作者： Gutin, Gregory Lei, Hui Yeo, Anders Zhou, Yacong Department of Computer Science Royal Holloway University of London United Kingdom School of Mathematical Sciences and LPMC Nankai University Tianjin300071 China School of Statistics and Data Science LPMC KLMDASR Nankai University Tianjin300071 China Department of Mathematics and Computer Science University of Southern Denmark Denmark Department of Mathematics and Applied Mathematics University of Johannesburg South Africa

An oriented multigraph is a directed multigraph without directed 2-cycles. Let fas(D) denote the minimum size of a feedback arc set in an oriented multigraph D. The degree of a vertex is the sum of its out- and in-degrees. In several papers, upper bounds for fas(D) were obtained for oriented multigraphs D with maximum degree upper-bounded by a constant. Hanauer (2017) conjectured that fas(D) ≤ 2.5n/3 for every oriented multigraph D with n vertices and maximum degree at most 5. We prove a strengthening of the conjecture: fas(D) ≤ m/3 holds for every oriented multigraph D with m arcs and maximum degree at most 5. This bound is tight and improves a bound of Berger and Shor (1990,1997). It would be interesting to determine c such that fas(D) ≤ cn for every oriented multigraph D with n vertices and maximum degree at most 5 such that the bound is tight. We show that 5/7 ≤ c ≤ 24/29 © 2024, CC BY.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Diffusion curvature for estimating local curvature in high dimensional data 22

Diffusion curvature for estimating local curvature in high d...

引用

Proceedings of the 36th International Conference on Neural Information Processing Systems

作者： Dhananjay Bhaskar Kincaid MacDonald Oluwadamilola Fasina Dawson Thomas Bastian Rieck Ian Adelstein Smita Krishnaswamy Department of Genetics Yale University Department of Mathematics Yale University Applied Mathematics Program Yale University Department of Mathematics Department of Physics Yale University Institute of AI for Health Helmholtz Pioneer Campus Helmholtz Munich Department of Computer Science Department of Genetics Applied Mathematics Program Program for Computational Biology and Bioinformatics Yale University

ISBN: (纸本)9781713871088

We introduce a new intrinsic measure of local curvature on point-cloud data called diffusion curvature. Our measure uses the framework of diffusion maps, including the data diffusion operator, to structure point cloud data and define local curvature based on the laziness of a random walk starting at a point or region of the data. We show that this laziness directly relates to volume comparison results from Riemannian geometry. We then extend this scalar curvature notion to an entire quadratic form using neural network estimations based on the diffusion map of point-cloud data. We show applications of both estimations on toy data, single-cell data and on estimating local Hessian matrices of neural network loss landscapes.

关键词：

来源：评论

学校读者我要写书评

暂无评论

An Optimal Framework for SDN Based on Deep Neural Network

引用

computers, Materials & Continua 2022年第10期73卷 1125-1140页

作者： Abdallah Abdallah Mohamad Khairi Ishak Nor Samsiah Sani Imran Khan Fahad RAlbogamy Hirofumi Amano Samih M.Mostafa Department of Industrial Engineering School of Applied Technical Sciences German Jordanian UniversityAmman35247Jordan School of Electrical and Electronic Engineering Universiti Sains MalaysiaNibong Tebal14300Malaysia Center for Artificial Intelligence Technology Faculty of Information Science and TechnologyThe National University of Malaysia(UKM)Bangi43600SelangorMalaysia Department of Electrical Engineering University of Engineering and TechnologyPeshawar814Pakistan Turabah University College Computer Sciences ProgramTaif UniversityTaif21944Saudi Arabia Research Institute for Information Technology Kyushu UniversityFukuoka819-0395Japan Computer Science-Mathematics Department Faculty of ScienceSouth Valley UniversityQena83523Egypt

Software-defined networking(SDN)is a new paradigm that promises to change by breaking vertical integration,decoupling network control logic from the underlying routers and switches,promoting(logical)network control centralization,and introducing network ***,the controller is similarly vulnerable to a“single point of failure”,an attacker can execute a distributed denial of service(DDoS)attack that invalidates the controller and compromises the network security in *** address the problem of DDoS traffic detection in SDN,a novel detection approach based on information entropy and deep neural network(DNN)is *** approach contains a DNN-based DDoS traffic detection module and an information-based entropy initial inspection *** initial inspection module detects the suspicious network traffic by computing the information entropy value of the data packet’s source and destination Internet Protocol(IP)addresses,and then identifies it using the DDoS detection module based on *** assaults were found when suspected irregular traffic was *** reveal that the algorithm recognizes DDoS activity at a rate of more than 99%,with a much better accuracy *** false alarm rate(FAR)is much lower than that of the information entropy-based detection ***,the proposed framework can shorten the detection time and improve the resource utilization efficiency.

关键词： Deep neural network computer networks data security optimization

来源：评论

学校读者我要写书评

暂无评论

A Secure Rotation Invariant LBP Feature Computation in Cloud Environment

引用

computers, Materials & Continua 2021年第9期68卷 2979-2993页

作者： Shiqi Wang Mingfang Jiang Jiaohua Qin Hengfu Yang Zhichen Gao College of Computer Science and Information Technology Central South University of Forestry and TechnologyChangsha410004China Department of Information Science and Engineering Hunan First Normal UniversityChangsha410205China Department of Applied Mathematics and Statistics College of Engineering and Applied SciencesStony Brook UniversityNY11794USA

In the era of big data,outsourcing massive data to a remote cloud server is a promising *** storage and computation services can reduce storage costs and computational ***,public cloud storage brings about new privacy and security concerns since the cloud servers can be shared by multiple ***-preserving feature extraction techniques are an effective solution to this *** the Rotation Invariant Local Binary Pattern(RILBP)has been widely used in various image processing fields,we propose a new privacy-preserving outsourcing computation of RILBP over encrypted images in this paper(called PPRILBP).To protect image content,original images are encrypted using block scrambling,pixel circular shift,and pixel diffusion when uploaded to the cloud *** is proved that RILBP features remain unchanged before and after ***,the server can directly extract RILBP features from encrypted *** and experiments confirm that the proposed scheme is secure and effective,and outperforms previous secure LBP feature computing methods.

关键词： Privacy-preserving rotation invariant local binary pattern cloud computing image encryption

来源：评论

学校读者我要写书评

暂无评论

A massively parallel speciation-based differential evolution algorithm applied to the 3D-AB protein structure prediction

A massively parallel speciation-based differential evolution...

引用

作者： Stubs Parpinelli, Rafael Boiani, Mateus Dias, André E. P. Graduate Program in Applied Computing State University of Santa Catarina Santa Catarina Joinville Brazil Rio Grande do Sul Porto Alegre Brazil Computer Science Department State University of Santa Catarina Santa Catarina Joinville Brazil

One of the most challenging problems in Bioinformatics is the finding of a protein conformation and it is known as the Protein Structure Prediction (PSP) problem. The main feature present in the AB off-lattice model is the use of polarity as the main driving force to guide the optimization process. The present work proposes an adaptive evolutionary algorithm based on GPU that is applied to the 3D-AB off-lattice PSP problem. The proposed approach is named cuDSMjDE, and is composed of a Dynamic Speciation-based Mutation Strategies Differential Evolution that employs the jDE mechanism to control F and CR parameters. A crowding strategy is also adopted, increasing competition between individuals and maintaining the diversity of solutions. All routines of the proposed algorithm are developed to run in GPU. During the experiments, eight Protein Data Bank sequences are employed. Results obtained concerning potential energy are competitive when compared with state-of-the-art algorithms. The use of a massively parallel architecture promoted the necessary scalability with speedups up to 708.78 (Figure presented.). © 2021 John Wiley & Sons, Ltd.

关键词： Evolutionary algorithms

来源：评论

学校读者我要写书评

暂无评论

NEW LOGARITHMIC STEP SIZE FOR STOCHASTIC GRADIENT DESCENT

arXiv

引用

arXiv 2024年

作者： Soheil Shamaee, M. Fathi Hafshejani, S. Saeidian, Z. Department of Computer Science Faculty of Mathematical Science University of Kashan Kashan Iran Department of Applied Mathematics Shiraz University of Technology Shiraz Iran Department of Mathematical Sciences University of Kashan Kashan Iran

In this paper, we propose a novel warm restart technique using a new logarithmic step size for the stochastic gradient descent (SGD) approach. For smooth and non-convex functions, we establish an O(1/√T) convergence rate for the SGD. We conduct a comprehensive implementation to demonstrate the efficiency of the newly proposed step size on the FashionMinst, CIFAR10, and CIFAR100 datasets. Moreover, we compare our results with nine other existing approaches and demonstrate that the new logarithmic step size improves test accuracy by 0.9% for the CIFAR100 dataset when we utilize a convolutional neural network (CNN) model. Copyright © 2024, The Authors. All rights reserved.

关键词： Neural network models

来源：评论

学校读者我要写书评

暂无评论

Extending adversarial attacks to produce adversarial class probability distributions

The Journal of Machine Learning Research

引用

The Journal of Machine Learning Research 2023年第1期24卷 491-532页

作者： Jon Vadillo Roberto Santana Jose A. Lozano Department of Computer Science and Artificial Intelligence University of the Basque Country UPV/EHU Donostia Spain Department of Computer Science and Artificial Intelligence University of the Basque Country UPV/EHU Donostia Spain and Basque Center for Applied Mathematics Bilbao Spain

Despite the remarkable performance and generalization levels of deep learning models in a wide range of artificial intelligence tasks, it has been demonstrated that these models can be easily fooled by the addition of imperceptible yet malicious perturbations to natural inputs. These altered inputs are known in the literature as adversarial examples. In this paper, we propose a novel probabilistic framework to generalize and extend adversarial attacks in order to produce a desired probability distribution for the classes when we apply the attack method to a large number of inputs. This novel attack paradigm provides the adversary with greater control over the target model, thereby exposing, in a wide range of scenarios, threats against deep learning models that cannot be conducted by the conventional paradigms. We introduce four different strategies to efficiently generate such attacks, and illustrate our approach by extending multiple adversarial attack algorithms. We also experimentally validate our approach for the spoken command classification task and the Tweet emotion classification task, two exemplary machine learning problems in the audio and text domain, respectively. Our results demonstrate that we can closely approximate any probability distribution for the classes while maintaining a high fooling rate and even prevent the attacks from being detected by label-shift detection methods.

关键词： adversarial examples deep neural networks robust classification class probability distributions linear programming

来源：评论

学校读者我要写书评

暂无评论

SPLINE BASED METHODS FOR FUNCTIONAL DATA ON MULTIVARIATE DOMAINS

arXiv

引用

arXiv 2023年

作者： Basna, Rani Nassar, Hiba Podgórski, Krzysztof Department of Statistics Lund University Sweden Cognitive Systems Department of Applied Mathematics and Computer Science Technical University of Denmark Denmark

Functional data analysis is typically performed in two steps: first, functionally representing discrete observations, and then applying functional methods to the so-represented data. The initial choice of a functional representation may have a significant impact on the second phase of the analysis, as shown in recent research, where data-driven spline bases outperformed the predefined rigid choice of functional representation. The method chooses an initial functional basis by an efficient placement of the knots using a simple machine-learning algorithm. The approach does not apply directly when the data are defined on domains of a higher dimension than one such as, for example, images. The reason is that in higher dimensions the convenient and numerically efficient spline bases are obtained as tensor bases from 1D spline bases that require knots that are located on a lattice. This does not allow for a flexible knot placement that was fundamental for the 1D approach. The goal of this research is to propose two modified approaches that circumvent the problem by coding the irregular knot selection into their densities and utilizing these densities through the topology of the spaces of splines. This allows for regular grids for the knots and thus facilitates using the spline tensor bases. It is tested on 1D data showing that its performance is comparable to or better than the previous methods. © 2023, CC BY.

关键词： Tensors

来源：评论

学校读者我要写书评

暂无评论

Globally injective ReLU networks

The Journal of Machine Learning Research

引用

The Journal of Machine Learning Research 2022年第1期23卷 4544-4598页

作者： Michael Puthawala Konik Kothari Matti Lassas Ivan Dokmanić Maarten De Hoop Department of Computational and Applied Mathematics Rice University Houston TX Department of Computer Science University of Illinois at Urbana-Champaign Champaign IL Department of Mathematics and Statistics University of Helsinki Finland Department of Mathematics and Computer Science University of Basel Basel Switzerland

Injectivity plays an important role in generative models where it enables inference; in inverse problems and compressed sensing with generative priors it is a precursor to well posedness. We establish sharp characterizations of injectivity of fully-connected and convolutional ReLU layers and networks. First, through a layerwise analysis, we show that an expansivity factor of two is necessary and sufficient for injectivity by constructing appropriate weight matrices. We show that global injectivity with iid Gaussian matrices, a commonly used tractable model, requires larger expansivity between 3.4 and 10.5. We also characterize the stability of inverting an injective network via worst-case Lipschitz constants of the inverse. We then use arguments from differential topology to study injectivity of deep networks and prove that any Lipschitz map can be approximated by an injective ReLU network. Finally, using an argument based on random projections, we show that an end-to-end--rather than layerwise--doubling of the dimension suffices for injectivity. Our results establish a theoretical basis for the study of nonlinear inverse and inference problems using neural networks.

关键词： bilipschitz networks injective generative models random neural networks deep learning universal approximation

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：