检索结果-内蒙古大学图书馆

Small sample learning based on probability-informed neural networks for SAR image segmentation

Neural Computing and Applications 2025年第14期37卷 8285-8308页

作者： Dostovalova, Anastasia Gorshenin, Andrey Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences Vavilova Str 44/2 Moscow119333 Russia

The paper introduces a probability-informed methodology for the segmentation of synthetic aperture radar (SAR) images in the case of small sample learning. It assumes that the amount of training data is limited to several hundred or thousand elements, which prevents the effective training of state-of-the-art neural network (NN) models. This is a typical problem for real SAR images whose characteristics depend significantly on the sensors used to produce them and cannot always be repeated within open available datasets. To solve this problem, we propose NN models called Probability-Informed Neural Networks (PrINNs). As part of our approach, we introduce the use of probability models as a source of additional features for data. Specifically, the training dataset is enriched by modeling the pixel brightness using a finite normal mixture. We prove that such an extension can reduce errors in the learning process theoretically. The resulting enriched dataset is segmented using attention-based convolutional NNs or visual transformers. Then, post-processing is implemented based on another probability model—quadtree, which is a special case of random Markov fields. As we have theoretically demonstrated, this part of PrINNs is analogous to the graph-convolutional NNs with fixed weights. Using open SAR images obtained by different radars (namely, Sentinel-1, Capella, ESAR and HRSID) with various types of underlying surfaces, the possibility of improving segmentation quality based on PrINNs is demonstrated. We tested various combinations of methods from the PrINNs architecture, and in all cases, the PrINN approach we proposed was superior to any other combination of these methods. From the point of view of the achieved accuracy metrics, the mean F1 score increased up to 19.24%, and the median F1 score was improved up to 9.57%. Some further architectural improvements to PrINNs are also discussed in the paper. © The Author(s), under exclusive licence to Springer-Verlag London Ltd

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

Identification of an Endogenous Production Function for the U.S. Economy 15th

Identification of an Endogenous Production Function for t...

引用

15th International Conference on Optimization and Applications, OPTIMA 2024

作者： Olenev, Nicholas Federal Research Center "Computer Science and Control" of the Russian Academy of Science Moscow Russia

ISBN: (纸本)9783031791185

The paper identifies an endogenous production function for the U.S. economy, represented by the distribution of production capacities across technologies with finite lifetimes. Technology characteristics (initial labor and capital intensity) are determined during the creation of the production unit. The microeconomic description of the model is based on the hypothesis that production capacity decreases with age, while the number of jobs remains the same until the production unit is liquidated when its lifetime limit is exceeded. Initial labor and capital intensity may decrease under the influence of scientific and technological progress. The parameters of the production function built on the basis of microeconomic description are determined by parallel calculations indirectly by comparing the model-calculated and statistical time series of macro-indicators of the U.S. economy in 1970–2022. A preliminary economic interpretation of the obtained results is presented. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Endogenous production function Parameter identification Production capacity U.S. economy Vintage capacity model

来源：评论

学校读者我要写书评

暂无评论

Modeling the Vibrations of an Annular Membrane with a Nonsteady General Boundary Condition

引用

Russian Engineering research 2025年第1期45卷 92-96页

作者： Kostikov, Yu. A. Romanenkov, A.M. Moscow Aviation Institute Moscow Russia Federal Research Center "Computer Science and Control " Russian Academy of Sciences Moscow Russia

Abstract: The initial boundary value problem regarding vibrations of an annular membrane is considered. Nonsteady boundary conditions are specified, and there is no distributed load. This is a nonclassical formulation, since rigid attachment conditions at the edges are usually specified. By appropriate replacement of the function to be determined, the initial problem may be reduced to an inhomogeneous problem with trivial boundary conditions of general form. Application of the standard Fourier method then allows us to express the solution as a functional series based on linear combinations of Bessel functions of the first and second kinds. © Allerton Press, Inc. 2025.

关键词： Boundary value problems

来源：评论

学校读者我要写书评

暂无评论

Solution of a One-Dimensional Parabolic Problem with a Nonlocal Boundary Condition

引用

Russian Engineering research 2025年第1期45卷 101-105页

作者： Danilina, I.A. Kostikov, Yu. A. Romanenkov, A.M. Moscow Aviation Institute Moscow Russia Federal Research Center "Computer Science and Control " Russian Academy of Sciences Moscow Russia

Abstract: For a one-dimensional boundary problem associated with a linear parabolic equation in the presence of a nonlocal spatial condition, necessary and sufficient conditions for the existence of a solution are established. Structural proof of the existence of a solution is presented. By means of Laguerre polynomials, a solution is explicitly constructed in the form of a functional series. © Allerton Press, Inc. 2025.

关键词： Boundary conditions

来源：评论

学校读者我要写书评

暂无评论

Forced Resonant Damping Vibrations of a Viscoelastic Beam

引用

Russian Engineering research 2025年第1期45卷 88-91页

作者： Kostikov, Yu. A. Romanenkov, A.M. Skachkov, A.S. Moscow Aviation Institute Moscow Russia Federal Research Center "Computer Science and Control " Russian Academy of Sciences Moscow Russia

Abstract: The initial model boundary value problem regarding vibrations of a viscoelastic beam with damping of Voigt type is considered. A classical formulation of a mixed problem for a fifth-order linear partial differential equation is adopted. The equation includes a term with a mixed derivative: first-order with respect to the time and fourth-order with respect to the spatial variable. On the basis of integral assessment, the energy identity is established, and the presence of damping vibrations is qualitatively demonstrated. By Galerkin projection, an exact solution of this problem is obtained. An algorithm is determined for calculation of the exact solution as a functional series. An explicit resonant term is obtained in general form. © Allerton Press, Inc. 2025.

关键词： Maximum principle

来源：评论

学校读者我要写书评

暂无评论

Modeling the Growth of Household Savings with a Nonlocal Boundary Condition

引用

Russian Engineering research 2025年第1期45卷 97-100页

作者： Kostikov, Yu. A. Romanenkov, A.M. Parenkina, V.I. Moscow Aviation Institute Moscow Russia Federal Research Center "Computer Science and Control " Russian Academy of Sciences Moscow Russia

Abstract: A stochastic model of the growth of household (family) savings is considered. Conditions are identified such that there exists a solution to the mixed parabolic problem for the distribution density of households in terms of savings. Using Laguerre polynomials, a solution is constructed explicitly in the form of a functional series for an initial boundary value problem with a nonlocal boundary condition. An economic interpretation of this nonlocal condition is outlined. © Allerton Press, Inc. 2025.

关键词： Boundary value problems

来源：评论

学校读者我要写书评

暂无评论

Motion of an Electrically Conducting Viscous Liquid at a Porous Wall with Liquid Injection

引用

Russian Engineering research 2025年第1期45卷 106-109页

作者： Vysk, N.D. Gurchenkov, A.A. Romanenkov, A.M. Moscow Aviation Institute Moscow Russia Federal Research Center "Computer Science and Control " Russian Academy of Sciences Moscow Russia

The motion of viscous electrically conducting incompressible liquid is investigated, in circumstances where the liquid rotates initially in a solid mass at constant speed together with a porous boundary wall (plate) until the instantaneous onset of longitudinal vibrations in the wall. The wall is at an arbitrary angle with respect to the axis of rotation. An unsteady flux is induced by the following factors: the longitudinal vibrations;injection (removal) of the liquid perpendicular to the surface of the porous wall;and a constant magnetic field perpendicular to the wall, which is also switched on instantaneously. Exact equations are derived for the velocity field and pressure field of the liquid. The magnetic field induced in the flux of electrically conducting liquid is determined. A hydrodynamic paradox is established on the basis of the results. © Allerton Press, Inc. 2025.

关键词： Magnetic fields

来源：评论

学校读者我要写书评

暂无评论

A Model of Investment Policy of Firms 15th

A Model of Investment Policy of Firms

引用

15th International Conference on Optimization and Applications, OPTIMA 2024

作者： Olenev, Nicholas Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences Moscow Russia

ISBN: (纸本)9783031791185

The microeconomic description of the investment policy of a firm in a market-type economic system is presented. Firms differ from each other by the moment of creation and are limited liability companies. At the moment of creation, the firm’s production capacity and its technology are set. Investments are made at the expense of borrowed funds provided by the banking system. An optimal control problem of economic activity of the firm is set and solved. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Banking System Investment Policy Optimal control Production capacity

来源：评论

学校读者我要写书评

暂无评论

Improving Feasibility of Optimal control via Obtaining High-Precision Model 4th

Improving Feasibility of Optimal Control via Obtaining Hig...

引用

4th International Conference on Numerical Computations: Theory and Algorithms, NUMTA 2023

作者： Shmalko, Elizaveta Prokopiev, Igor Diveev, Askhat Yamshanov, Konstantin Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences Moscow Russia

ISBN: (纸本)9783031812408

In the absence of any observation system or low veracity of the data, it is possible to provide control over a limited time interval basing on a high-precision control object model used. The paper proposes to use a multilayer artificial neural network (ANN) to obtain such a model. In this case, instead of the ordinary differential equation (ODE) system in the Cauchy form, we get a mixed ODE-ANN mathematical model, the parameters of which are tuned to the dynamics of a particular control object. The paper describes the process of identification of the control object model including obtaining a sufficient volume of the training sample. General data properties are formulated to obtain a good model from the point of view of use in control problems. The problem of optimal control for an object described with ANN-based model is formulated and numerical approach for its solution is proposed. An example of solving the optimal control problem for a mobile robot based on the identified neural network model is given. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Multilayer neural networks

来源：评论

学校读者我要写书评

暂无评论

Toward Supervised Deep Gaussian Mixture Models⋆ 27th

Toward Supervised Deep Gaussian Mixture Models⋆

引用

27th International Conference on Distributed computer and Communication Networks, DCCN 2024

作者： Gorshenin, Andrey Federal Research Center "Computer Science and Control" of the Russian Academy of Sciences Moscow Russia

ISBN: (纸本)9783031808524

The paper presents for the first time a methodology for solving supervised learning problems, such as classification and regression, based on deep Gaussian mixture models (DGMMs). We use a self-supervised approach to construct a classifier as well as a semi-supervised one for a regressor. More than 20 public UCI datasets with various parameters were used for testing. It has been demonstrated that the greatest increase in classification accuracy of 37.69% is achieved by using the ensemble of DGMM and extreme gradient boosting (XGBoost). The accuracy of this method exceeds that of the combination of GMM and SVM by 14.51%. The DGMM regression (DGMMR) analogue of the Gaussian mixture model regression (GMMR) is introduced as a semi-supervised learning algorithm. On the test data, the best results were shown by the ensemble of DGMMR and XGBoost regression. The accuracy of this method exceeded the combination with support vector machines regression (SVR), as well as variants of GMMR with SVR and linear regression with SVR by 3.58%, 11.63% and 32.78%, respectively. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Self-supervised learning

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：