检索结果-内蒙古大学图书馆

arXiv 2021年

作者： Jiang, Haiyan Xiong, Haoyi Wu, Dongrui Liu, Ji Dou, Dejing Big Data Lab Baidu Research School of Artificial Intelligence and Automation Huazhong University of Science and Technology

Principal component analysis (PCA) has been widely used as an effective technique for feature extraction and dimension reduction. In the High Dimension Low Sample Size (HDLSS) setting, one may prefer modified principal components, with penalized loadings, and automated penalty selection by implementing model selection among these different models with varying penalties. The earlier work [1, 2] has proposed penalized PCA, indicating the feasibility of model selection in l2-penalized PCA through the solution path of Ridge regression, however, it is extremely time-consuming because of the intensive calculation of matrix inverse. In this paper, we propose a fast model selection method for penalized PCA, named Approximated Gradient Flow (AgFlow), which lowers the computation complexity through incorporating the implicit regularization effect introduced by (stochastic) gradient flow [3, 4] and obtains the complete solution path of l2-penalized PCA under varying l2-regularization. We perform extensive experiments on real-world datasets. AgFlow outperforms existing methods (Oja [5], Power [6], and Shamir [7] and the vanilla Ridge estimators) in terms of computation costs. Copyright © 2021, The Authors. All rights reserved.

关键词： Principal component analysis

来源：评论

学校读者我要写书评

暂无评论

Federated Learning for Predictive Maintenance and Quality Inspection in Industrial Applications

Federated Learning for Predictive Maintenance and Quality In...

引用

Prognostics and System Health Management Conference (PHM)

作者： Viktorija Pruckovskaja Axel Weissenfeld Clemens Heistracher Anita Graser Julia Kafka Peter Leputsch Daniel Schall Jana Kemnitz Technische Universität Wien Vienna Austria Data Science & Artificial Intelligence AIT Austrian Institute of Technology Vienna Austria Innovation Software Stiwa Automation Attnang-Puchheim Austria Distributed AI Systems Research Siemens Technology Vienna Austria

data-driven machine learning is playing a crucial role in the advancements of Industry 4.0, specifically in enhancing predictive maintenance and quality inspection. Federated learning (FL) enables multiple participants to develop a machine learning model without compromising the privacy and confidentiality of their data. In this paper, we evaluate the performance of different FL aggregation methods and compare them to central and local training approaches. Our study is based on four datasets with varying data distributions. The results indicate that the performance of FL is highly dependent on the data and its distribution among clients. In some scenarios, FL can be an effective alternative to traditional central or local training methods. Additionally, we introduce a new federated learning dataset from a real-world quality inspection setting.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Modeling and Prediction of Ship Trajectory in Inland River Crossing Channels Based on GRU Neural Network

Modeling and Prediction of Ship Trajectory in Inland River C...

引用

2021 International Conference on Security, Pattern Analysis, and Cybernetics, SPAC 2021

作者： Jiang, Junhao Zuo, Yi Li, Tieshan Ma, He Navigation College Dalian Maritime University Dalian China Collaborative Innovation Center of Maritime Big Data and Shipping Artificial General Intelligence Dalian China School of Automation Engineering University of Electronic Science and Technology of China Chengdu China

ISBN: (纸本)9781665443227

With the development of inland river transportation, the navigable conditions of inland waterways are becoming more and more complicated. Therefore, navigation with the aid of navigator alone is not enough. It is also necessary to make reasonable decisions on the navigation of the ship based on the future navigation behavior of the ship. Traditional trajectory prediction methods are difficult to accurately train models with the small amount of Automatic Identification System (AIS) data. In order to solve the problem of the data and the original models, this paper uses the Hermite interpolation method to expand the data. On this basis, we propose a model ship trajectory based on Gated Recurrent Unit (GRU) neural network. In the experiment, we select the AIS data of ships on the Huai'an section of the Beijing-Hangzhou Canal. Finally, Back Propagation Neural Network (BPNN) model and Support Vector Regression (SVR) model are conducted to a comparative experiments. The results show that the GRU neural network trajectory prediction model has high prediction accuracy, feasibility and effectiveness. © 2021 IEEE.

关键词： Ships

来源：评论

学校读者我要写书评

暂无评论

Spiking Neural Networks for Temporal Processing: Status Quo and Future Prospects

arXiv

引用

arXiv 2025年

作者： Ma, Chenxiang Chen, Xinyi Li, Yanchen Yang, Qu Wu, Yujie Li, Guoqi Pan, Gang Tang, Huajin Tan, Kay Chen Wu, Jibin Department of Data Science and Artificial Intelligence The Hong Kong Polytechnic University Hong Kong Department of Computing The Hong Kong Polytechnic University Hong Kong Department of Electrical and Computer Engineering National University of Singapore 119077 Singapore Institute of Automation Chinese Academy of Sciences Beijing100045 China State Key Laboratory of Brain-Machine Intelligence College of Computer Science and Technology MOE Frontier Science Center for Brain Science and Brain-Machine Integration Zhejiang University Hangzhou310027 China

Temporal processing is fundamental for both biological and artificial intelligence systems, as it enables the comprehension of dynamic environments and facilitates timely responses. Spiking Neural Networks (SNNs) excel in handling such data with high efficiency, owing to their rich neuronal dynamics and sparse activity patterns. Given the recent surge in the development of SNNs, there is an urgent need for a comprehensive evaluation of their temporal processing capabilities. In this paper, we first conduct an in-depth assessment of commonly used neuromorphic benchmarks, revealing critical limitations in their ability to evaluate the temporal processing capabilities of SNNs. To bridge this gap, we further introduce a benchmark suite consisting of three temporal processing tasks characterized by rich temporal dynamics across multiple timescales. Utilizing this benchmark suite, we perform a thorough evaluation of recently introduced SNN approaches to elucidate the current status of SNNs in temporal processing. Our findings indicate significant advancements in recently developed spiking neuron models and neural architectures regarding their temporal processing capabilities, while also highlighting a performance gap in handling long-range dependencies when compared to state-of-the-art non-spiking models. Finally, we discuss the key challenges and outline potential avenues for future research. Copyright © 2025, The Authors. All rights reserved.

关键词： Spatio-temporal data

来源：评论

学校读者我要写书评

暂无评论

Robust stability analysis for class of Takagi-Sugeno (T-S) fuzzy with stochastic process for sustainable hypersonic vehicles

引用

Information sciences 2023年第1期641卷

作者： Aslam, Muhammad Shamrooz Tiwari, Prayag Pandey, Hari Mohan Band, Shahab S. Artificial Intelligence Research Institute China University of Mining and Technology Xuzhou China School of Automation Guangxi University of Science and Technology Liuzhou545006 China School of Information Technology Halmstad University Sweden Data Science & Artificial Intelligence Department Bournemouth University Fern Barrow Dorset PooleBH12 5BB United Kingdom Future Technology Research Center College of Future National Yunlin University of Science and Technology 123 University Road Section 3 Yunlin Douliou64002 Taiwan

Recently, the rapid development of Unmanned Aerial Vehicles (UAVs) enables ecological conservation, such as low-carbon and "green" transport, which helps environmental sustainability. In order to address control issues in a given region, UAV charging infrastructure is urgently needed. To better achieve this task, an investigation into the T–S fuzzy modeling for Sustainable Hypersonic Vehicles (SHVs) with Markovian jump parameters and H∞ attitude control in three channels was conducted. Initially, the reentry dynamics were transformed into a control–oriented affine nonlinear model. Then, the original T–S local modeling method for SHV was projected by primarily referring to Taylor's expansion and fuzzy linearization methodologies. After the estimation of precision and controller complexity was assumed, the fuzzy model for jump nonlinear systems mainly consisted of two levels: a crisp level and a fuzzy level. The former illustrates the jumps, and the latter a fuzzy level that represents the nonlinearities of the system. Then, a systematic method built in a new coupled Lyapunov function for a stochastic fuzzy controller was used to guarantee the closed–loop system for H∞ gain in the presence of a predefined performance index. Ultimately, numerical simulations were conducted to show how the suggested controller can be successfully applied and functioned in controlling the original attitude dynamics. © 2023 Elsevier Inc.

关键词： Stochastic systems

来源：评论

学校读者我要写书评

暂无评论

The Codecfake dataset and Countermeasures for the Universally Detection of Deepfake Audio

arXiv

引用

arXiv 2024年

作者： Xie, Yuankun Lu, Yi Fu, Ruibo Wen, Zhengqi Wang, Zhiyong Tao, Jianhua Qi, Xin Wang, Xiaopeng Liu, Yukun Cheng, Haonan Ye, Long Sun, Yi State Key Laboratory of Media Convergence and Communication China School of Data Science and Intelligent Media Communication University of China Beijing China Institute of Automation Chinese Academy of Sciences Beijing China School of Artificial Intelligence University of Chinese Academy of Sciences Beijing China Department of Automation Tsinghua University Beijing China Beijing National Research Center for Information Science and Technology Tsinghua University Beijing China School of Cyberspace science and Technology Beijing Institute of Technology Beijing China

With the proliferation of Audio Language Model (ALM) based deepfake audio, there is an urgent need for generalized detection methods. ALM-based deepfake audio currently exhibits widespread, high deception, and type versatility, posing a significant challenge to current audio deepfake detection (ADD) models trained solely on vocoded data. To effectively detect ALM-based deepfake audio, we focus on the mechanism of the ALM-based audio generation method, the conversion from neural codec to waveform. We initially constructed the Codecfake dataset, an open-source, large-scale collection comprising over 1 million audio samples in both English and Chinese, focus on ALM-based audio detection. As countermeasure, to achieve universal detection of deepfake audio and tackle domain ascent bias issue of original sharpness aware minimization (SAM), we propose the CSAM strategy to learn a domain balanced and generalized minima. In our experiments, we first demonstrate that ADD model training with the Codecfake dataset can effectively detects ALM-based audio. Furthermore, our proposed generalization countermeasure yields the lowest average equal error rate (EER) of 0.616% across all test conditions compared to baseline models. The dataset1and associated code2are available online. © 2024, CC BY.

关键词： C (programming language)

来源：评论

学校读者我要写书评

暂无评论

Domain Generalized Solar Cell Defect Segmentation Based on Shape-Aware and Multi-View Meta-Learning

Domain Generalized Solar Cell Defect Segmentation Based on S...

引用

Chinese Control Conference (CCC)

作者： Kun Liu Ying Yang Jingkun Mao Weipeng Liu School of Artificial Intelligence and Data Science Hebei University of Technology Tianjin China Control Engineering Technology Innovation Center of Hebei Province Hebei University of Technology Tianjin P. R. China Department of Automation Tianjin University of Technology Tianjin China

Various defects are inevitably generated in the manufacturing process of solar cells. Deep learning-based methods for defect segmentation under closed situation have achieved remarkable progress. Due to the difference of imaging condition and camera parameter under different production line, there are large differences in brightness distribution of solar cell images. The model trained under closed situation cannot achieve good performance in opened situation such as multi-production lines. In this paper, a new shape-aware and multi-view meta-learning scheme is proposed to improve the model generalization performance for single-domain solar cell defect segmentation. The scheme roots in gradient-based meta-learning. Multi-brightness view information is yielded from the generated augmented images, and explicitly simulating domain shift with virtual meta-train and meta-test during training to mitigate overfitting in single-source domain train and unstable prediction. Importantly, considering tiny and faint solar cell defects are not easily identified under different image brightness conditions, shape edge constraint loss is proposed to encourage shape compactness and shape smoothness of segmentations. Experimental results show that the proposed method has high segmentation accuracy, which outperforms the state-of-the-art methods.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Multi-Scale Attention Semi-Supervised Network Based on Confidence Iteration and Pseudo-Label Optimization for COVID-19 CT Segmentation

Multi-Scale Attention Semi-Supervised Network Based on Confi...

引用

Chinese Control Conference (CCC)

作者： Kun Liu Mei Chen Junfei Xie JingKun Mao WeiPeng Liu School of Artificial Intelligence and Data Science Hebei University of Technology Tianjin China Control Engineering Technology Innovation Center of Hebei Province Hebei University of Technology Tianjin P. R. China Department of Automation Tianjin University of Technology Tianjin China

Computed tomography (CT) image segmentation for COVID-19 lung infection has made great progress. However, accurate segmentation remains a major challenge due to the varying shape and size of COVID-19 infected areas, the more similar tissue of infected lung images to normal lung images, and the additional problem of a limited number of medical images annotated at the pixel level. In this paper, we propose a semi-supervised segmentation network based on confidence iteration and pseudo-label optimization for COVID-19 CT segmentation. In addition, we develop a multi-scale attention module which can achieve attentional feature fusion at different scales. The experiment evaluated on the COVID-19 CT Segmentation dataset, demonstrate our method achieves higher segmentation accuracy than the state-of-art methods.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Optimizing Medication Compliance through Machine Learning and IoT for Personalized Health Management

Optimizing Medication Compliance through Machine Learning an...

引用

International Conference on Communication and Electronics Systems (ICCES)

作者： P. Balaji I.K. Mohitha K. Prashanthini K. Sivaprakash P. Rithick S. Kasthuri Department of Electronics and Instrumentation Engineering Sri Ramakrishna Engineering College Coimbatore India Department of Electronics and Communication Engineering KGiSL Institute of Technology Coimbatore India Department of Robotics and Automation Engineering Sri Ramakrishna Engineering College Coimbatore India Department of Artificial Intelligence and Data Science SNS College of Engineering Coimbatore India

ISBN: (数字)9798350377972

ISBN: (纸本)9798350377989

Machine learning (ML) and Internet of Things (IoT) technologies have revolutionized the healthcare industry, particularly in the field of medication administration. Inadequate adherence to medication is a serious worldwide issue that can result in poor health and higher medical expenses. Through continuous monitoring and analysis of patient-specific parameters, such as vital signs, medication history, and lifestyle factors, the system dynamically adjusts medication schedules and dosages to optimize therapeutic efficacy while minimizing adverse effects. The system tracks medicine usage, vital signs, and ambient variables using data gathered from Internet of Things (IoT) devices like wearable devices and smart pill bottles. An ML model, trained to anticipate potential adherence problems, is then provided with this preprocessed data. Personalized interventions, like medicine reminders, educational content, and progress reports, are delivered by the Intelligent Medication Compliance System (IMAC) based on these predictions. The integration of IoT-enabled smart sensors and connectivity allows for smooth communication between patients, healthcare professionals, and medicine dispensers, permitting prompt interventions, prescription reminders, and remote monitoring. Additionally, the system makes it easier for patients and medical professionals to communicate, allowing for prompt modifications to treatment programs. Serving as an anticipatory system, the IMAC provides personalized interventions. Predicting possible adherence problems, such as missing doses or inadvertent drug interactions, is done by the model through data analysis. The proposed IMAS seeks to empower patients, personalized medication management, and eventually improve medication adherence by fusing real-time IoT data with ML-driven insights.

关键词： Industries Schedules Accuracy Medical services Machine learning Predictive models data models Internet of Things Wearable devices Smart phones

来源：评论

学校读者我要写书评

暂无评论

Prediction of ship fuel consumption based on Elastic network regression model

Prediction of ship fuel consumption based on Elastic network...

引用

2021 International Conference on Information, Cybernetics, and Computational Social Systems, ICCSS 2021

作者： Li, Shanshan Li, Xinyu Zuo, Yi Li, Tieshan Dalian Maritime University Navig. Coll. Collab. Innov. Center of Maritime Big Data and Marine Artificial General Intelligence Dalian China University of Electronic Science and Technology of China College of Automation Engineering Chengdu China

ISBN: (纸本)9781665402453

Predicting the fuel consumption of ships sailing under different navigation conditions and improving the operation efficiency of shipping industry has become an important topic. There are many characteristic variables affecting ship fuel consumption during navigation, such as trim, draft, wind speed, wind direction and so on. And some variables are highly correlated, which is easy to produce multicollinearity problems. It makes the fuel consumption prediction complex. The study established an Elastic network regression model by combining the least absolute contraction and selection operator (LASSO) and Ridge regression algorithm. The model reduces the complexity and improves the interpretability and accuracy by selecting the characteristic variables affecting ship fuel consumption. The study is verified by the navigation data of a ferry within two months. The results show that compared with long short term memory (LSTM) and back-propagation neural network (BPNN), the Elastic network regression model can not only explain the relationship between fuel consumption and variables, but also predict fuel consumption more accurately and effectively. © 2021 IEEE.

关键词： Regression analysis

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：