检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Gautam, Srishti Boubekki, Ahcene Höhne, Marina M.C. Kampffmeyer, Michael C. Department of Physics and Technology UiT The Arctic University of Norway Norway Machine Learning and Uncertainty Physikalisch-Technische Bundesanstalt Germany Data Science in Bioeconomy Leibniz Institute for Agriculture and Bioeconomy Institute for Computer Science University of Potsdam Germany

Explainable AI (XAI) has unfolded in two distinct research directions with, on the one hand, post-hoc methods that explain the predictions of a pre-trained black-box model and, on the other hand, self-explainable models (SEMs) which are trained directly to provide explanations alongside their predictions. While the latter is preferred in safety-critical scenarios, post-hoc approaches have received the majority of attention until now, owing to their simplicity and ability to explain base models without retraining. Current SEMs, instead, require complex architectures and heavily regularized loss functions, thus necessitating specific and costly training. To address this shortcoming and facilitate wider use of SEMs, we propose a simple yet efficient universal method called KMEx (K-Means Explainer), which can convert any existing pre-trained model into a prototypical SEM. The motivation behind KMEx is to enhance transparency in deep learning-based decision-making via class-prototype-based explanations that are diverse and trustworthy without retraining the base model. We compare models obtained from KMEx to state-of-the-art SEMs using an extensive qualitative evaluation to highlight the strengths and weaknesses of each model, further paving the way toward a more reliable and objective evaluation of SEMs. Copyright © 2023, The Authors. All rights reserved.

关键词： Decision making

来源：评论

学校读者我要写书评

暂无评论

The Implicit Regularization of Dynamical Stability in Stochastic Gradient Descent

arXiv

引用

arXiv 2023年

作者： Wu, Lei Su, Weijie J. School of Mathematical Sciences Peking University Beijing China Center for Machine Learning Research Peking University Beijing China Wharton Statistics and Data Science Department University of Pennsylvania Philadelphia United States

In this paper, we study the implicit regularization of stochastic gradient descent (SGD) through the lens of dynamical stability (Wu et al., 2018). We start by revising existing stability analyses of SGD, showing how the Frobenius norm and trace of Hessian relate to different notions of stability. Notably, if a global minimum is linearly stable for SGD, then the trace of Hessian must be less than or equal to 2/η, where η denotes the learning rate. By contrast, for gradient descent (GD), the stability imposes a similar constraint but only on the largest eigenvalue of Hessian. We then turn to analyze the generalization properties of these stable minima, focusing specifically on two-layer ReLU networks and diagonal linear networks. Notably, we establish the equivalence between these metrics of sharpness and certain parameter norms for the two models, which allows us to show that the stable minima of SGD provably generalize well. By contrast, the stability-induced regularization of GD is provably too weak to ensure satisfactory generalization. This discrepancy provides an explanation of why SGD often generalizes better than GD. Note that the learning rate (LR) plays a pivotal role in the strength of stability-induced regularization. As the LR increases, the regularization effect becomes more pronounced, elucidating why SGD with a larger LR consistently demonstrates superior generalization capabilities. Additionally, numerical experiments are provided to support our theoretical findings. © 2023, CC BY-SA.

关键词： Stability

来源：评论

学校读者我要写书评

暂无评论

A Study on Integrating machine learning Techniques for Waste Management

A Study on Integrating Machine Learning Techniques for Waste...

引用

International Conference on Circuit, Power and Computing Technologies (ICCPCT)

作者： Ruchika Vaidya Rahul Dattangire Divya Biradar Prateek Verma Department of Artificial Intelligence and Machine Learning Faculty of Engineering and Technology Datta Meghe Institute of Higher Education and Research (DU) Wardha Maharashtra India Data Engineering Publicis Sapient Houston Texas USA Department of Computer Science University of Texas Arlignton Texas USA

ISBN: (数字)9798350372816

ISBN: (纸本)9798350372823

The current study focuses on how various machine learning (ML) algorithms can be used in waste management and emphasizes the significance of intelligent technologies in addressing contemporary issues. By leading a broad writing survey, the paper features the meaning of incorporating these advances into waste management practices. The system segment subtleties the information assortment, preprocessing, model determination, and assessment measurements utilized in the review. The accuracy and performance metrics of various classifiers, including K Nearest Neighbors, Linear Support Vector Classifier, Logistic Regression, Decision Tree, Multi-Layer Perceptron, and Random Forest, are demonstrated by the results of experiments with them. Especially noteworthy is the MLP classifier's 91.81% accuracy. The discoveries highlight the viability of ML approaches in ssmart waste management. In addition, a comparison table that summarizes the classifiers' accuracy scores provides insight into their suitability for applications in waste management.

关键词： Waste management Surveys Productivity Measurement Accuracy Reviews Vectors

来源：评论

学校读者我要写书评

暂无评论

Advanced Process Control in Manufacturing Using IoT Devices and Artificial Neural Networks 4

Advanced Process Control in Manufacturing Using IoT Devices ...

引用

4th International Conference on Sustainable Expert Systems, ICSES 2024

作者： Sumathi, V. Ramesh, S. Basavaraddi, Chethan Chandra S. Visumathi, J. Ishwarya, M.V. Srinivasan, S. Sri Sairam Engineering College Department of Mathematics Tamil Nadu Chennai India School of Computing College of Engineering and Technology Srm Institute of Science and Technology Department of Computing Technologies Kattankulathur Tamil Nadu Chennai India R&d Don Bosco Institute of Technology Vtu Department of Artificial Intelligence and Machine Learning Karnataka Belagavi India Veltech Rangarajan Dr. Sagunthala R&d Institute of Science and Technology Department of Information Technology Tamil Nadu Chennai India Agni College of Technology Artificial intelligence and Data science Department Tamil Nadu Chennai India Saveetha School of Engineering Saveetha Institute of Medical and Technical Sciences Saveetha University Department of Biomedical Engineering Tamil Nadu Chennai India

ISBN: (纸本)9798331540364

Optimization of industrial activities is significantly helped by Advanced Process Control (APC), which increases efficiency, decreases costs, and improves product quality. Artificial Neural Networks (ANNs) and the Internet of Things (IoT) have greatly improved APC in manufacturing. It examines how APC systems may include ANNs and IoT devices in industrial processes. By connecting to a wide range of sensors placed strategically throughout a production facility, IoT devices may gather data in real time, allowing for managing and monitoring crucial parameters. To APC systems, ANNs are invaluable for predictive modeling and control due to their adaptability to changing process conditions and capacity to represent complicated nonlinear interactions. Manufacturers may gain better control over their operations, improving productivity and product quality, by using data from the IoT to train and update ANN models. Recent advances in APC using IoT devices and ANNs are reviewed in this study. Their applications in various industrial industries are discussed, and case examples are shown to demonstrate their usefulness. Further improvements in APC for manufacturing using IoT devices and ANNs may be guided by highlighting the field's problems and future research goals. Challenges with data security, integration challenges, and reliability impact the accuracy of predictive modeling, scalability, and performance of IoT systems. © 2024 IEEE.

关键词： Adaptive control systems

来源：评论

学校读者我要写书评

暂无评论

Towards Enhancing Deep Active learning with Weak Supervision and Constrained Clustering

Towards Enhancing Deep Active Learning with Weak Supervision...

引用

2023 Workshop on Interactive Adaptive learning, IAL@ECML-PKDD 2023

作者： Aßenmacher, Matthias Rauch, Lukas Goschenhofer, Jann Stephan, Andreas Bischl, Bernd Roth, Benjamin Sick, Bernhard Department of Statistics LMU Munich Ludwigstr. 33 MunichD-80539 Germany LMU Munich Germany University of Kassel Wilhelmshöher Allee 73 KasselD-34121 Germany Fraunhofer IIS Erlangen Germany Research Group Data Mining and Machine Learning University of Vienna Kolingasse 14-16 ViennaA-1090 Austria UniVie Doctoral School Computer Science Vienna Austria

Three fields revolving around the question of how to cope with limited amounts of labeled data are Deep Active learning (DAL), deep Constrained Clustering (CC), and Weakly Supervised learning (WSL). DAL tackles the problem by adaptively posing the question of which data samples to annotate next in order to achieve the best incremental learning improvement, although it suffers from several limitations that hinder its deployment in practical settings. We point out how CC algorithms and WSL could be employed to overcome these limitations and increase the practical applicability of DAL research. Specifically, we discuss the opportunities to use the class discovery capabilities of CC and the possibility of further reducing human annotation efforts by utilizing WSL. We argue that the practical applicability of DAL algorithms will benefit from employing CC and WSL methods for the learning and labeling process. We inspect the overlaps between the three research areas and identify relevant and exciting research questions at the intersection of these areas. © 2023 Copyright for this paper by its authors.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Automatic Human Life Rescue Analysis Model with Incorporation of Deep learning Algorithm

Automatic Human Life Rescue Analysis Model with Incorporatio...

引用

Sustainable Communication Networks and Application (ICSCNA), International Conference on

作者： V.K.G. Kalaiselvi Prithikhaa S Divya Darshini J S Bharath Reddy Anandkumar B Shanmugasundaram Hariharan Department of Information Technology Sri Sairam Engineering College Tamilnadu India Department of CSE KG Reddy College of Engineering and Technology Telangana India Department of Artificial Intelligence and Machine Learning Vardhaman College of Engineering India Department of Artificial Intelligence and Data Science Vardhaman College of Engineering Hyderabad India

ISBN: (数字)9798331530013

ISBN: (纸本)9798331530020

In today's fast-paced world, every passing moment can mean the difference between life and death. To address this critical issue, we have developed an innovative application designed to provide rapid emergency medical response. Ambulance services play a pivotal role in emergency healthcare by offering vital transportation and care. However, limited access to ambulance contact information often hampers patients' ability to secure timely assistance during emergencies. With the widespread adoption of smartphones and mobile applications, the proposed solution integrates React Native with a Convolutional Neural Network (CNN) to create an advanced mobile app for law enforcement, capable of real-time car accident detection. Utilizing datasets from Kaggle, images are resized to optimize CNN training. Through transfer learning, the CNN architecture effectively extracts spatial features, enabling accurate accident prediction, validated by robust evaluation metrics such as precision and recall. The React Native app facilitates swift notifications to law enforcement, utilizing cutting-edge deep learning technology to enhance road safety and improve emergency response efficiency. This project aims to streamline emergency services, minimizing response and transport times, thereby saving lives and fostering safer communities.

关键词： Deep learning Training Law enforcement Medical services Emergency services Road safety Mobile applications Convolutional neural networks Time factors Accidents

来源：评论

学校读者我要写书评

暂无评论

MAKING THERMODYNAMIC MODELS OF MIXTURES PREDICTIVE BY machine learning: MATRIX COMPLETION OF PAIR INTERACTIONS

arXiv

引用

arXiv 2022年

作者： Jirasek, Fabian Bamler, Robert Fellenz, Sophie Bortz, Michael Kloft, Marius Mandt, Stephan Hasse, Hans TU Kaiserslautern Kaiserslautern Germany Data Science and Machine Learning University of Tübingen Tübingen Germany Machine Learning Group TU Kaiserslautern Kaiserslautern Germany Kaiserslautern Germany Department of Computer Science University of California Irvine Irvine United States

Predictive models of thermodynamic properties of mixtures are paramount in chemical engineering and chemistry. Classical thermodynamic models are successful in generalizing over (continuous) conditions like temperature and concentration. On the other hand, matrix completion methods (MCMs) from machine learning successfully generalize over (discrete) binary systems;these MCMs can make predictions without any data for a given binary system by implicitly learning commonalities across systems. In the present work, we combine the strengths from both worlds in a hybrid approach. The underlying idea is to predict the pair-interaction energies, as they are used in basically all physical models of liquid mixtures, by an MCM. As an example, we embed an MCM into UNIQUAC, a widely-used physical model for the Gibbs excess energy. We train the resulting hybrid model in a Bayesian machine-learning framework on experimental data for activity coefficients in binary systems of 1,146 components from the Dortmund data Bank. We thereby obtain, for the first time, a complete set of UNIQUAC parameters for all binary systems of these components, which allows us to predict, in principle, activity coefficients at arbitrary temperature and composition for any combination of these components, not only for binary but also for multi-component systems. The hybrid model even outperforms the best available physical model for predicting activity coefficients, the modified UNIFAC (Dortmund) model. Copyright © 2022, The Authors. All rights reserved.

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

An Improved Finite-time Analysis of Temporal Difference learning with Deep Neural Networks

arXiv

引用

arXiv 2024年

作者： Ke, Zhifa Wen, Zaiwen Zhang, Junyu Center for Data Science Peking University China Beijing International Center for Mathematical Research Center for Machine Learning Research Changsha Institute for Computing and Digital Economy Beijing China Department of Industrial Systems Engineering and Management National University of Singapore Singapore

Temporal difference (TD) learning algorithms with neural network function parameterization have well-established empirical success in many practical large-scale reinforcement learning tasks. However, theoretical understanding of these algorithms remains challenging due to the nonlinearity of the action-value approximation. In this paper, we develop an improved non-asymptotic analysis of the neural TD method with a general L-layer neural network. New proof techniques are developed and an improved new Õ(ϵ−1) sample complexity is derived. To our best knowledge, this is the first finite-time analysis of neural TD that achieves an Õ(ϵ−1) complexity under the Markovian sampling, as opposed to the best known Õ(ϵ−2) complexity in the existing literature. Copyright © 2024, The Authors. All rights reserved.

关键词： Asymptotic analysis

来源：评论

学校读者我要写书评

暂无评论

An improved finite-time analysis of temporal difference learning with deep neural networks 24

An improved finite-time analysis of temporal difference lear...

引用

Proceedings of the 41st International Conference on machine learning

作者： Zhifa Ke Zaiwen Wen Junyu Zhang Center for Data Science Peking University China Beijing International Center for Mathematical Research Center for Machine Learning Research and Changsha Institute for Computing and Digital Economy Beijing China Department of Industrial Systems Engineering and Management National University of Singapore Singapore

Temporal difference (TD) learning algorithms with neural network function parameterization have well-established empirical success in many practical large-scale reinforcement learning tasks. However, theoretical understanding of these algorithms remains challenging due to the nonlinearity of the action-value approximation. In this paper, we develop an improved nonasymptotic analysis of the neural TD method with a general L-layer neural network. New proof techniques are developed and an improved new Õ(ε-1) sample complexity is derived. To our best knowledge, this is the first finite-time analysis of neural TD that achieves an Õ(ε-1) complexity under the Markovian sampling, as opposed to the best known Õ(ε-2) complexity in the existing literature.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Rainfall Prediction using Novel Approach for Enhancing Crop Yield 2

Rainfall Prediction using Novel Approach for Enhancing Crop ...

引用

2nd International Conference on Advanced Computing and Communication Technologies, ICACCTech 2024

作者： Seelam, Visweswararao Vikruthi, Sriharsha Bhagavatham, Naresh Kumar Nagendrudu, P.V.V.S.D. Srinivasarao, Goda Chakravarthi M, Ayyappa PACE Institute of Technology and Sciences Dept of CSE India B V Raju Institute of Technology Dept of CSE Telangana Narsapur Medak502313 India Vignana bharathi Institute of Technology Dept of CSE Hyderabad India Aditya University Department of Artificial Intelligence and Machine Learning Surampalem India KL University Vaddeswaram Dept of CSE Andhra Pradesh Guntur India KKR & KSR Institute of Technology & Sciences Dept of CSE-Data Science Andhra Pradesh Guntur India

ISBN: (纸本)9798331519056

In India, agriculture is one of the fastest-growing fields to improve the country's economy. Based on several factors like climatic conditions, predicting rainfall in these stages is one of the most significant tasks. In recent years, unexpected and unpredictable weather conditions have been observed in some places. This study mainly focused on developing the Hybrid algorithm (HA) to predict rainfall in the early stages and find accurate weather conditions. The proposed approach also called as Novel Approach that improves the algorithm's strength by adopting the advanced preprocessing and feature extraction technique. The training model random forest (RF) also used to find the accurate patterns from the given dataset samples. The preprocessing technique focused on cleaning and transforming the raw data to remove the noise from the input dataset, improving the data quality. The proposed approach combines Autoencoders and Linear Regression (LR) to accurately predict rainfall. Extensive experiments are carried out to evaluate the performance of the hybrid algorithm using historical meteorological data from multiple regions with varying climatic conditions. The results show that the hybrid algorithm outperforms traditional methods and individual machine-learning models in terms of accuracy, robustness, and generalization. © 2024 IEEE.

关键词： Adversarial machine 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：