检索结果-内蒙古大学图书馆

ADVERSARIAL ATTACKS AND DEFENSES IN FAULT DETECTION AND DIAGNOSIS: A COMPREHENSIVE BENCHMARK ON THE TENNESSEE EASTMAN PROCESS

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Pozdnyakov, Vitaliy Kovalenko, Aleksandr Makarov, Ilya Drobyshevskiy, Mikhail Lukyanov, Kirill AIRI Moscow Russia ISP RAS Research Center for Trusted Artificial Intelligence Moscow Russia Ivannikov Institute for System Programming of the Russian Academy of Sciences Moscow Russia Moscow Russia

Integrating machine learning into Automated Control systems (ACS) enhances decision-making in industrial process management. One of the limitations to the widespread adoption of these technologies in industry is the vulnerability of neural networks to adversarial attacks. This study explores the threats in deploying deep learning models for Fault Detection and Diagnosis (FDD) in ACS using the Tennessee Eastman Process dataset. By evaluating three neural networks with different architectures, we subject them to six types of adversarial attacks and explore five different defense methods. Our results highlight the strong vulnerability of models to adversarial samples and the varying effectiveness of defense strategies. We also propose a new defense strategy based on combining adversarial training and data quantization. This research contributes several insights into securing machine learning within ACS, ensuring robust FDD in industrial processes. © 2024, CC BY-NC-ND.

关键词： Decision making

Convergence analysis of stochastic gradient descent with adaptive preconditioning for non-convex and convex functions

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Pasechnyuk, Dmitrii A. Gasnikov, Alexander Takáč, Martin Mohamed bin Zayed University of Artificial Intelligence United Arab Emirates Moscow Institute of Physics and Technology Russia Kharkevich Institute for Information Transmission Problems RAS Russia Ivannikov Institute for System Programming RAS Russia

Preconditioning is a crucial operation in gradient-based numerical optimisation. It helps decrease the local condition number of a function by appropriately transforming its gradient. For a convex function, where the gradient can be computed exactly, the optimal linear transformation corresponds to the inverse of the Hessian operator, while the optimal convex transformation is the convex conjugate of the function. Different conditions result in variations of these dependencies. Practical algorithms often employ low-rank or stochastic approximations of the inverse Hessian matrix for preconditioning. However, theoretical guarantees for these algorithms typically lack a justification for the defining property of preconditioning. This paper presents a simple theoretical framework that demonstrates, given a smooth function and an available unbiased stochastic approximation of its gradient, that it is possible to refine the dependency of the convergence rate on the Lipschitz constant of the *** Codes 65K10 © 2023, CC BY.

关键词： Stochastic systems

Regularization of linear machine learning problems

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Liu, S. Kabanikhin, S.I. Strijhak, S.V. Novosibirsk State University Novosibirsk Russia Sobolev Institute of Mathematics Siberian Branch of the Russian Academy of Sciences Novosibirsk Russia V.P. Ivannikov Institute of System Programming of the Russian Academy of Sciences Moscow Russia

In this paper, we consider the simplest version of a linear neural network(LNN). Assuming that for training (constructing an optimal weight matrix $Q$)we have a set of training pairs, i.e. we know the input data \\begin{equation}G=\\left\\{g^{\\left(1\ight)},g^{\\left(2\ight)},\\cdots,g^{\\left(K\ight)}\ight\\},\\end{equation} as well as the correct answers to these input data\\begin{equation}H=\\left\\{h^{\\left(1\ight)},h^{\\left(2\ight)},\\cdots,h^{\\left(K\ight)}\ight\\}.\\end{equation} We will study the possibilities of constructing a weight matrix$Q$ of a neural network that will give correct answers to arbitrary input databased on the connection of the specified problem with a system of linearalgebraic equations (SLAE). Consider a class of neural networks in which eachneuron has only one output signal and performs linear operations. We will showhow such LNEs are reduced to SLAEs. Since the questions $G$ and the correctanswers $H$ are known to us, the desired weight matrix $Q$ must satisfy theequations \\begin{equation} Qg^{\\left(k\ight)}=h^{\\left(k\ight)},k=1,2,\\cdots,K. \\end{equation} It is required to restore $Q$. In the generalcase, the matrix $Q$ is rectangular $Q=Q_{MN}=\\left\\{q_{mn}\ight\\}$, $m$ isthe row number, and $g^{\\left(k\ight)}\\in\\mathbb{R}^N$,$h^{\\left(k\ight)}\\in\\mathbb{R}^M$. Let $G_{NK}$ be a matrix composed ofcolumns $g^{\\left(1\ight)}, g^{\\left(2\ight)},\\cdots,g^{\\left(k\ight)}$, and$H_{MK}$ be a matrix composed of columns$h^{\\left(1\ight)},h^{\\left(2\ight)},\\cdots,h^{\\left(k\ight)}$. Then, withrespect to $Q_{MN}$, we obtain a matrix SLAE. \\begin{equation}Q_{MN}G_{NK}=H_{MK}. \\end{equation} This paper will present methods forregularizing the constructed system. © 2024, CC BY.

关键词： Input output programs

On the Problem of Choosing the Optimal Parameters for the Wind Farm in the Arctic Town of Tiksi 20th

学校读者我要写书评

暂无评论

On the Problem of Choosing the Optimal Parameters for the Wi...

20th International Conference on Mathematical Modeling and Supercomputer Technologies, MMST 2020

作者： Strijhak, Sergei V. Gergel, Victor P. Ivanov, Aleksandr V. Gadal, Sebastien Zh. Ivannikov Institute for System Programming of the RAS Moscow109004 Russia Moscow Aviation Institute Moscow125993 Russia Lobachevsky State University of Nizhny Novgorod Nizhnij Novgorod603950 Russia Aix-Marseille Univ CNRS ESPACE UMR 7300 Univ Nice Sophia Antipolis Avignon Univ Aix-en-Provence13545 France North-Eastern Federal University Yakutsk Sakha Yakutia67000 Russia

ISBN: (纸本)9783030787585

The paper considers the problem of choosing the optimal parameters for the operation of 3 horizontal wind turbines of the wind farm in the town Tiksi, in the Sakha Republic. The open-source WRF-ARW and FLORIS packages are used to calculate the physical parameters in the wind farm. During the calculation, the values of wind velocity, temperature, pressure fields, and the value of the generated power of wind power plants were obtained. The optimization problem is formulated, the objective function is defined, and 2 variable parameters are selected, namely, yaw angle and angle of attack. The calculations were carried out on the computing cluster of ISP RAS. © 2021, Springer Nature Switzerland AG.

关键词： Angle of attack

A Case Study of Valency Patterns for Motion Verbs in Vakh Khanty by Means of Lingvodoc

学校读者我要写书评

暂无评论

A Case Study of Valency Patterns for Motion Verbs in Vakh Kh...

ivannikov ISPRAS Open Conference (ISPRAS)

作者： Irina Novitskaya Victoria Vorobeva Department of Foreign Languages National Research Tomsk State University Tomsk Russian Federation Information Systems Department Ivannikov Institute for System Programming of the RAS TPU Moscow Tomsk Russian Federation

The article presents a study of the syntactic compatibility of verbs of motion in the Vakh dialect of the Khanty language. The study was conducted with the help of the tools of the platform for documentation of the Uralic languages Lingvodoc. The purpose of the article is to describe patterns of syntactic valency that Vakh Khanty verbs of motion have.

关键词： Documentation Syntactics

Deductive verification of unmodified linux kernel library functions

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Efremov, Denis Mandrykin, Mikhail Khoroshilov, Alexey National Research University Higher School of Economics Moscow Russia Ivannikov Institute for System Programming RAS Moscow Russia Moscow Institute of Physics and Technology Moscow Russia Lomonosov Moscow State University Moscow Russia

This paper presents results from the development and evaluation of a deductive verification benchmark consisting of 26 unmodified Linux kernel library functions implementing conventional memory and string operations. The formal contract of the functions was extracted from their source code and was represented in the form of preconditions and postconditions. The correctness of 23 functions was completely proved using AstraVer toolset, although success for 11 functions was achieved using 2 new specification language constructs. Another 2 functions were proved after a minor modification of their source code, while the final one cannot be completely proved using the existing memory model. The benchmark can be used for the testing and evaluation of deductive verification tools and as a starting point for verifying other parts of the Linux kernel. Copyright © 2018, The Authors. All rights reserved.

关键词： Formal verification

Control of Overfitting with Physics

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Kozyrev, Sergei V. Lopatin, Ilya A. Pechen, Alexander N. Steklov Mathematical Institute of Russian Academy of Sciences Gubkina St. 8 Moscow119991 Russia Ivannikov Institute for System Programming The Russian Academy of Sciences Alexandra Solzhenitsyna str. 25 Moscow109004 Russia

While there are many works on the applications of machine learning, not so many of them are trying to understand the theoretical justifications to explain their efficiency. In this work, overfitting control (or generalization property) in machine learning is explained using analogies from physics and biology. For stochastic gradient Langevin dynamics, we show that the Eyring formula of kinetic theory allows to control overfitting in the algorithmic stability approach—when wide minima of the risk function with low free energy correspond to low overfitting. For the generative adversarial network (GAN) model, we establish an analogy between GAN and the predator–prey model in biology. An application of this analogy allows us to explain the selection of wide likelihood maxima and overfitting reduction for GANs. Copyright © 2024, The Authors. All rights reserved.

关键词： Generative adversarial networks

Real-time analytics, hybrid transactional/analytical processing, in-memory data management, and non-volatile memory

学校读者我要写书评

暂无评论

Real-time analytics, hybrid transactional/analytical process...

ivannikov ISPRAS Open Conference (ISPRAS)

作者： Sergey D. Kuznetsov Pavel E. Velikhov Qiang Fu Ivannikov Institute for System Programming of the RAS Lomonosov Moscow State University Moscow Institute of Physics and Technology (State University) National Research University Higher School of Economics Moscow Russia Huawei Russian Research Institute Moscow Russia

ISBN: (数字)9781665412919

ISBN: (纸本)9781665448215

These days, real-time analytics is one of the most often used notions in the world of databases. Broadly, this term means very fast analytics over very fresh data. Usually the term comes together with other popular terms, hybrid transactional/analytical processing (HTAP) and in-memory data processing. The reason is that the simplest way to provide fresh operational data for analysis is to combine in one system both transactional and analytical processing. The most effective way to provide fast transactional and analytical processing is to store an entire database in memory. So on the one hand, these three terms are related but on the other hand, each of them has its own right to life. In this paper, we provide an overview of several in-memory data management systems that are not HTAP systems. Some of them are purely transactional, some are purely analytical, and some support real-time analytics. Then we overview nine in-memory HTAP DBMSs, some of which don't support real-time analytics. Existing real-time in-memory HTAP DBMSs have very diverse and interesting architectures although they use a number of common approaches: multiversion concurrency control, multicore parallelization, advanced query optimization, just in time compilation, etc. Additionally, we are interested whether these systems use non-volatile memory, and, if yes, in what manner. We conclude that an emergence of new generation of NVM will greatly stimulate its use in in-memory HTAP systems.

关键词： Nonvolatile memory Multicore processing Query processing Distributed databases Transforms Real-time systems Concurrency control

Stochastic Gradient Methods with Preconditioned Updates

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Sadiev, Abdurakhmon Beznosikov, Aleksandr Almansoori, Abdulla Jasem Kamzolov, Dmitry Tappenden, Rachael Takáč, Martin Ivannikov Institute for System Programming Moscow Russia Moscow Institute of Physics and Technology Moscow Russia Mohamed bin Zayed University of Artificial Intelligence Abu Dhabi United Arab Emirates University of Canterbury Christchurch New Zealand

This work considers the non-convex finite sum minimization problem. There are several algorithms for such problems, but existing methods often work poorly when the problem is badly scaled and/or ill-conditioned, and a primary goal of this work is to introduce methods that alleviate this issue. Thus, here we include a preconditioner based on Hutchinson’s approach to approximating the diagonal of the Hessian, and couple it with several gradient-based methods to give new ‘scaled’ algorithms: Scaled SARAH and Scaled L-SVRG. Theoretical complexity guarantees under smoothness assumptions are presented. We prove linear convergence when both smoothness and the PL condition are assumed. Our adaptively scaled methods use approximate partial second-order curvature information and, therefore, can better mitigate the impact of badly scaled problems. This improved practical performance is demonstrated in the numerical experiments also presented in this work. Copyright © 2022, The Authors. All rights reserved.

关键词： Gradient methods