检索结果-内蒙古大学图书馆

Robusta: Robust AutoML for Feature Selection via Reinforcement Learning

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Wang, Xiaoyang Li, Bo Zhang, Yibo Kailkhura, Bhavya Nahrstedt, Klara Department of Computer Science University of Illinois at Urbana-Champaign United States Center for Applied Scientific Computing Lawrence Livermore National Laboratory United States

Several AutoML approaches have been proposed to automate the machine learning (ML) process, such as searching for the ML model architectures and hyper-parameters. However, these AutoML pipelines only focus on improving the learning accuracy of benign samples while ignoring the ML model robustness under adversarial attacks. As ML systems are increasingly being used in a variety of mission-critical applications, improving the robustness of ML systems has become of utmost importance. In this paper, we propose the first robust AutoML framework, Robusta–based on reinforcement learning (RL)–to perform feature selection, aiming to select features that lead to both accurate and robust ML systems. We show that a variation of the 0-1 robust loss can be directly optimized via an RL-based combinatorial search in the feature selection scenario. In addition, we employ heuristics to accelerate the search procedure based on feature scoring metrics, which are mutual information scores, tree-based classifiers feature importance scores, F scores, and Integrated Gradient (IG) scores, as well as their combinations. We conduct extensive experiments and show that the proposed framework is able to improve the model robustness by up to 22% while maintaining competitive accuracy on benign samples compared with other feature selection methods. Copyright © 2021, The Authors. All rights reserved.

关键词： Feature extraction

Parallel algorithms for maximizing monotone one-sided σ-smooth functions

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Zhang, Hongxiang Cheng, Yukun Wu, Chenchen Xu, Dachuan Du, Dingzhu Beijing Institute for Scientific and Engineering Computing Beijing University of Technology Beijing100124 China School of Business Suzhou University of Science and Technology Suzhou215009 China College of Science Tianjin University of Technology Tianjin 300384 China Department of Computer Science University of Texa Dallas Dallas75083 United States

In this paper, we study the problem of maximizing a monotone normalized one-sided σ-smooth (OSS for short) function F (x), subject to a convex polytope (no need to downward-closed [1]). A function F (x) is one-sided σ-smooth if (Equation presented), for all x, u ≥ 0, x ≠ 0. This problem was first introduced by Mehrdad et al. [1] to characterize the multilinear extension of some set functions. Different with the serial algorithm with name Jump-Start Continuous Greedy Algorithm by Mehrdad et al. [1], we propose Jump-Start Parallel Greedy (JSPG for short) algorithm, the first parallel algorithm, for this problem. The approximation ratio of JSPG algorithm is proved to be (Equation presented) for any any number α ∈ (0, 1] and ǫ > 0, which improves the approximation ratio of JSCG algorithm in [1]. We also prove that our JSPG algorithm runs in (O(log n/ǫ2)) adaptive rounds and consumes O(n log n/ǫ2) queries, where the number of adaptive rounds and function evaluation queries approximately matches the known results for parallel submodular maximization. In addition, we study the stochastic version of maximizing monotone normalized OSS function, in which the objective function F (x) is defined as F (x) = Ey∼T f(x, y). Here f is a stochastic function with respect to the random variable Y, and y is the realization of Y drawn from a probability distribution T. For this stochastic version, we design Stochastic Parallel-Greedy (SPG) algorithm, which achieves a result of (Equation presented), with the same time complexity of JSPG algorithm. Here (Equation presented) is related to the preset parameters σ, L, D and time t. Copyright © 2022, The Authors. All rights reserved.

关键词： Parallel algorithms

MatterChat: A Multi-Modal LLM for Material science

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Tang, Yingheng Xu, Wenbin Cao, Jie Gao, Weilu Farrell, Steven Erichson, Benjamin Mahoney, Michael W. Nonaka, Andy Yao, Zhi Applied Mathematics and Computational Research Division Lawrence Berkeley National Laboratory BerkeleyCA United States National Energy Research Scientific Computing Center Lawrence Berkeley National Laboratory BerkeleyCA United States NSF National AI Institute for Student-AI Teaming University of Colorado at Boulder Boulder United States Department of Electrical and Computer Engineering The University of Utah Salt Lake CityUT United States Scientific Data Division Lawrence Berkeley National Laboratory BerkeleyCA United States International Computer Science Institute BerkeleyCA United States Department of Statistics University of California at Berkeley BerkeleyCA United States

Understanding and predicting the properties of inorganic materials is crucial for accelerating advancements in materials science and driving applications in energy, electronics, and beyond. Integrating material structure data with language-based information through multi-modal large language models (LLMs) offers great potential to support these efforts by enhancing human–AI interaction. However, a key challenge lies in integrating atomic structures at full resolution into LLMs. In this work, we introduce MatterChat, a versatile structure-aware multi-modal LLM that unifies material structural data and textual inputs into a single cohesive model. MatterChat employs a bridging module to effectively align a pretrained universal machine learning interatomic potential with a pretrained LLM, reducing training costs and enhancing flexibility. Our results demonstrate that MatterChat significantly improves performance in material property prediction and human-AI interaction, surpassing general-purpose LLMs such as GPT-4. We also demonstrate its usefulness in applications such as more advanced scientific reasoning and step-by-step material synthesis. © 2025, CC BY.

关键词： Structure (composition)

Group Spam Identification in IOT Home Automation Using Tracking Time Metadata

学校读者我要写书评

暂无评论

Group Spam Identification in IOT Home Automation Using Track...

Communication, computing and Energy Efficient Technologies (I3CEET), International Conference on

作者： Achilov Docent Saurabh Dhanik Pardeep Kumar Jindal Saif O. Husain Diwakar Bhardwaj Varsha Mittal Department of Physics and Chemistry Tashkent Institute of Irrigation and Agricultural Mechanization Engineers National Research University Tashkent Uzbekistan Scientific Researcher Western Caspian University Baku Azerbaijan School of Computing Graphic Era Hill University Bhimtal India Department of Electrical and Communication Engineering Chandigarh Engineering College Chandigarh Group of Colleges Mohali Punjab India Department of Computers Techniques Engineering College of Technical Engineering The Islamic University of Al Diwaniyah Al Diwaniyah Iraq Department of Computer Engineering and Applications GLA University Mathura India Department of Computer Science & Engineering Graphic Era Deemed to be University Dehradun India

ISBN: (数字)9798331541583

ISBN: (纸本)9798331541590

With the stark increase due to smarter homes using online of Things, or IoT devices exchanging info far more frequently than we used to — all wirelessly. IoT devices are affected by several issues like possible cyberattacks, the unpredictability of network connectivity and data leakage since there are many types of IoT devices being used today. OBJECTIVES: To improve the privacy of data provided from web of Things (IoT) smart home devices by using machine training and computational modeling to detect anomalies. At its core the experiment contribution is how reliable Web of Things (WoT), i.e. IoT technologies transferring values from or to home things, reasons are taking different features such as average sq rate error origin and feature value in account of evaluations about internet-connected device efficacy. It is a metric that seems so important when we consider our dependency on every attached Internet of Things (IoT) device, in the network's architecture. The proposed method is applied and validated using a dataset of smart homes along with real temperature information collected for public access. Using Time series evidence collected by iot devices, the suggested method works extremely well as it can detect any anomaly or spam. This technique helps with the general health and safety of your entire structure by creating a solid base for evaluating the acceptability of Sea-of-things (SoT) gadgets in just a smart home network..

关键词： Training Data privacy Time series analysis Smart homes Smart meters Telecommunication computing Internet of Things Reliability Object recognition Tuning

Two Methods for Regularized Non-Monotone Submodular Maximization

学校读者我要写书评

暂无评论

SSRN

SSRN 2024年

作者： Sun, Xin Han, Congying Wu, Chenchen Xu, Dachuan Zhou, Yang Du, Ding-Zhu School of Mathematical Sciences University of Chinese Academy of Sciences Beijing100049 China College of Sciecne Tianjin University of Technology Tianjin300072 China Beijing Institute for Scientific and Engineering Computing Beijing University of Technology Beijing100124 China School of Mathematics and Statistics Shandong Normal University Jinan250014 China Department of Computer Science University of Texas at Dallas RichardsonTX75080 United States

In this paper, we investigate regularized non-monotone submodular maximization over a down-monotone family of sets by applying the Lyapunov method and the distorted continuous greedy algorithm. Based on the Lyapunov method, we systematically design parameterized algorithm frameworks in two phases. In the first phase, a theoretical approximation algorithm can be derived, as long as a set of parameters is chosen. Contingent upon the selection of a suitable set of parameters, the approximation guarantee can achieve $(1/e, \frac{\gamma-\gamma/e-1}{\gamma-1})$, where $\gamma\in[0,1)\cup[e,+\infty)$ reflects the relative dominance of the positive and negative parts of the optimal value for the regular term. In the second phase, an implementable algorithm is presented through discretization in $\mathcal{O}(n^3/\epsilon)$ with arbitrarily small loss in guarantee. Additionally, we establish that no polynomial-time algorithm yields a $(0.478,\lambda)$-approximation with $\lambda\in[0,1]$. Furthermore, we demonstrate that the Lyapunov method can also be applied in regularized monotone submodular maximization with the tight $(1-1/e, 1)$-approximation ratio. © 2024, The Authors. All rights reserved.

关键词： Approximation algorithms

High performance hyperspectral image classification using graphics processing units

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Mohammad, Mahmoud Ahmed Hossam Edeen Tolba, Mohammad Fahmy Ebied, Hala Muosher Aziz, Mohammad Hassan Abdel Basic Science Department Faculty of Computer and Information Sciences Ain Shams University Scientific Computing Department Faculty of Computer and Information Sciences Ain Shams University Basic Sciences Department Faculty of Computer and Information Sciences Ain Shams University

Real-time remote sensing applications like search and rescue missions, military target detection, environmental monitoring, hazard prevention and other time-critical applications require onboard real time processing capabilities or autonomous decision making. Some unmanned remote systems like satellites are physically remote from their operators, and all control of the spacecraft and data returned by the spacecraft must be transmitted over a wireless radio link. This link may not be available for extended periods when the satellite is out of line of sight of its ground station. In addition, providing adequate electrical power for these systems is a challenging task because of harsh conditions and high costs of production. Onboard processing addresses these challenges by processing data on-board prior to downlink, instead of storing and forwarding all captured images from onboard sensors to a control station, resulting in the reduction of communication bandwidth and simpler subsequent computations to be performed at ground stations. Therefore, lightweight, small size and low power consumption hardware is essential for onboard real time processing systems. With increasing dimensionality, size and resolution of recent hyperspectral imaging sensors, additional challenges are posed upon remote sensing processing systems and more capable computing architectures are needed. Graphical Processing Units (GPUs) emerged as promising architecture for light weight high performance computing that can address these computational requirements for onboard systems. The goal of this study is to build high performance hyperspectral analysis solutions based on selected high accuracy analysis methods. These solutions are intended to help in the production of complete smart remote sensing systems with low power consumption. We propose accelerated parallel solutions for the well-known recursive hierarchical segmentation (RHSEG) clustering method, using GPUs, hybrid multicore CPU with a GPU

关键词： Graphics processing unit

Learning constitutive models from microstructural simulations via a non-intrusive reduced basis method

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Guo, Theron Rokoš, Ondřej Veroy, Karen Centre for Analysis Scientific Computing and Applications Department of Mathematics and Computer Science Netherlands Mechanics of Materials Department of Mechanical Engineering Eindhoven University of Technology Eindhoven5612 AZ Netherlands

In order to optimally design materials, it is crucial to understand the structure-property relations in the material by analyzing the effect of microstructure parameters on the macroscopic properties. In computational homogenization, the microstructure is thus explicitly modeled inside the macrostructure, leading to a coupled two-scale formulation. Unfortunately, the high computational costs of such multiscale simulations often render the solution of design, optimization, or inverse problems infeasible. To address this issue, we propose in this work a non-intrusive reduced basis method to construct inexpensive surrogates for parametrized microscale problems;the method is specifically well-suited for multiscale simulations since the coupled simulation is decoupled into two independent problems: (1) solving the microscopic problem for different (loading or material) parameters and learning a surrogate model from the data;and (2) solving the macroscopic problem with the learned material model. The proposed method has three key features. First, the microscopic stress field can be fully recovered, which is useful for instance for revealing local stress concentrations inside the microstructure. Second, the method is able to accurately predict the stress field for a wide range of material parameters;furthermore, the derivatives of the effective stress with respect to the material parameters are available and can be readily utilized in solving optimization problems. Finally, it is more data efficient, i.e. requiring less training data, as compared to directly performing a regression on the effective stress. To construct the surrogate model, first, a proper orthogonal decomposition is performed on precomputed microscopic stress field snapshots to find a reduced basis for the stress. Second, a regression is employed to infer the coefficients of the reduced basis approximation for any arbitrary parameter value, thus enabling a rapid online evaluation of the microscopic stress.

关键词： Microstructure

Scalar field comparison with topological descriptors: Properties and applications for scientific visualization

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Yan, Lin Masood, Talha Bin Sridharamurthy, Raghavendra Rasheed, Farhan Natarajan, Vijay Hotz, Ingrid Wang, Bei Scientific Computing and Imaging Institute University of Utah United States Linköping University Norrköping Sweden Department of Computer Science and Automation Indian Institute of Science Bangalore India

In topological data analysis and visualization, topological descriptors such as persistence diagrams, merge trees, contour trees, Reeb graphs, and Morse-Smale complexes play an essential role in capturing the shape of scalar field data. We present a state-of-the-art report on scalar field comparison using topological descriptors. We provide a taxonomy of existing approaches based on visualization tasks associated with three categories of data: single fields, time-varying fields, and ensembles. These tasks include symmetry detection, periodicity detection, key event/feature detection, feature tracking, clustering, and structure statistics. Our main contributions include the formulation of a set of desirable mathematical and computational properties of comparative measures, and the classification of visualization tasks and applications that are enabled by these measures. Copyright © 2021, The Authors. All rights reserved.

关键词： Visualization

End-to-end performance modeling of distributed GPU applications 34

学校读者我要写书评

暂无评论

End-to-end performance modeling of distributed GPU applicati...

34th ACM International Conference on Supercomputing, ICS 2020

作者： Choi, Jaemin Richards, David F. Kale, Laxmikant V. Bhatele, Abhinav Department of Computer Science University of Illinois at Urbana-Champaign UrbanaIL United States Center for Applied Scientific Computing Lawrence Livermore National Laboratory LivermoreCA United States

ISBN: (纸本)9781450379830

With the growing number of GPU-based supercomputing platforms and GPU-enabled applications, the ability to accurately model the performance of such applications is becoming increasingly important. Most current performance models for GPU-enabled applications are limited to single node performance. In this work, we propose a methodology for end-to-end performance modeling of distributed GPU applications. Our work strives to create performance models that are both accurate and easily applicable to any distributed GPU application. We combine trace-driven simulation of MPI communication using the TraceR-CODES framework with a profiling-based roofline model for GPU kernels. We make substantial modifications to these models to capture the complex effects of both on-node and off-node networks in today's multi-GPU supercomputers. We validate our model against empirical data from GPU platforms and also vary tunable parameters of our model to observe how they might affect application performance. © 2020 ACM.

关键词： Graphics processing unit