检索结果-内蒙古大学图书馆

22nd international conference on Software Engineering and Formal Methods, SEFM 2024

作者： von Berg, Benjamin Aichernig, Bernhard K. Rindler, Maximilian Štern, Darko Tappler, Martin Institute of Software Technology Graz University of Technology Graz Austria TU Wien Vienna Austria AVL List GmbH Graz Austria

ISBN: (纸本)9783031773815

Passive automata learning is a method for inferring automaton models from a given collection of observations of system behavior (traces). It has been applied to reactive systems with probabilistic behavior. In particular, IOAlergia is a well known algorithm for inferring models in the form of deterministically labeled Markov decision processes from system traces. The quality of the resulting model depends heavily on the provided data set and suffers if data is scarce. However, in many cases additional knowledge about the system is available. This work aims to incorporate knowledge about system modes into the learning process in order to improve model quality for low-data scenarios. This is done by splitting the traces according to system modes, learning individual models for each mode and combining those models into one model with sub-regions corresponding to individual system modes. In our evaluation on artificial models, our method outperforms the baseline in at least 90 % of cases for all considered metrics. This method was developed to learn generative models of human driving behavior. data from recorded test drives on highways was used to learn a hierarchical stochastic model of typical acceleration behavior of human drivers. In the automotive industry, such models make the simulations of driving emissions more realistic. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Markov processes

来源：评论

学校读者我要写书评

暂无评论

machine learning-Based GPS Multipath Detection Method Using Dual Antennas 13

Machine Learning-Based GPS Multipath Detection Method Using ...

引用

13th Asian Control conference (ASCC)

作者： Kim, Sanghyun Byun, Jungyun Park, Kwansik Yonsei Univ Sch Integrated Technol Incheon South Korea Korea Aerosp Res Inst Daejeon South Korea

ISBN: (纸本)9788993215236

In urban areas, global navigation satellite system (GNSS) signals are often reflected or blocked by buildings, thus resulting in large positioning errors. In this study, we proposed a machine learning approach for global positioning system (GPS) multipath detection that uses dual antennas. A machine learning model that could classify GPS signal reception conditions was trained with several GPS measurements selected as suggested features. We applied five features for machine learning, including a feature obtained from the dual antennas, and evaluated the classification performance of the model, after applying four machine learning algorithms: gradient boosting decision tree (GBDT), random forest, decision tree, and K-nearest neighbor (KNN). It was found that a classification accuracy of 82%-96% was achieved when the test data set was collected at the same locations as those of the training data set. However, when the test data set was collected at locations different from those of the training data, a classification accuracy of 44%-77% was obtained.

关键词： multipath detection global positioning system machine learning dual antennas

来源：评论

学校读者我要写书评

暂无评论

Radiotherapy Treatment Planning: An Integrated Optimization and Reinforcement learning Approach 10th

Radiotherapy Treatment Planning: An Integrated Optimization ...

引用

10th international conference on machine learning, Optimization, and data science, LOD 2024

作者： Anjo, Eva Rocha, Humberto Dias, Joana Faculdade de Ciências e Tecnologia Universidade de Coimbra Coimbra Portugal Faculdade de Economia and CeBER Universidade de Coimbra Coimbra Portugal Faculdade de Economia and INESC Coimbra Universidade de Coimbra Coimbra Portugal

ISBN: (纸本)9783031824807

Radiotherapy treatment planning is one of the most important steps in the radiotherapy treatment workflow. It is responsible for the creation of an individualized treatment for each case, taking into account the characteristics of the volumes to treat and all the surrounding healthy tissues and structures. Typically characterized by a manual, trial-and-error methodology, the selection of irradiation directions and intensities remains a complex endeavor. In this work, a new approach for the automatic generation of radiotherapy treatment plans is described and a proof of concept is considered using prostate cancer cases. In this approach, an unconstrained quadratic optimization problem calculates radiation intensities, with the model parameters being iteratively determined by a fuzzy inference system. To enhance computational efficiency, reinforcement learning, specifically Q-learning, is incorporated. Reinforcement learning enables the fuzzy inference system to dynamically adjust to the evolving requirements of the iterative optimization process. During the learning phase, Q-tables are constructed and subsequently utilized as lookup tables to guide decision-making in each iteration of the optimization approach. In the conducted computational experiments using cross-validation, a noteworthy reduction of up to 66% in the total number of iterations required to achieve a high-quality treatment plan was observed. While acknowledging the preliminary nature of these results, they unequivocally underscore the promising potential of synergizing Reinforcement learning and Optimization methodologies in automating the intricate process of radiotherapy treatment planning. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Leveraging machine learning and Spatial Analysis to Analyze Dynamic Land-Use Changes in Indramayu Regency to Enhance Food Security 11

Leveraging Machine Learning and Spatial Analysis to Analyze ...

引用

11th international conference on Computer, Control, Informatics and its Applications, IC3INA 2024

作者： Nur, Wawan Hendriawan Al Awaby, Khafidz Asshidqi Ardiyanto, Bayu Saputra, Okta Fajar Pertiwi, Dewi Mustika Haryanti, Dina Firmansyah, Moch Rafli Susilowati, Yuliana Kumoro, Yugo Research Centre for Data and Information Sciences National Research and Innovation Agency Bandung Indonesia Faculty of Agriculture Brawijaya University Malang Indonesia Research Centre for Geoinformatics National Research and Innovation Agency Bandung Indonesia Research Centre for Climate and Atmosphere National Research and Innovation Agency Bandung Indonesia Research Centre for Geological Hazard National Research and Innovation Agency Bandung Indonesia

ISBN: (纸本)9798331542313

Indramayu Regency, a national food barn, has determined sustainable agricultural land to bolster national food security by prohibiting land-use conversion. Studies on land use and land cover changes are essential to manage land cover changes, particularly in sustainable food agricultural areas, and to support food security. This study aims to classifiy land cover in Indramayu Regency using Random Forest and Support Vector machine, and spatial analysis to identify land cover change trends. The study was conducted on Sentinel-2A satellite imagery from 2019 to 2023 using the Google Earth Engine. The results indicate that Random Forest achieved higher accuracy than SVM, with the highest kappa value of 0.9 and overall accuracy (OA) of 93%. Spatial analysis revealed that the most significant land cover change occurred in paddy fields converted to bare land, covering an area of 4301.33 ha, or 860 ha per year. In addition, paddy fields were converted into residential areas, encompassing 1810.35 ha, or 362 ha per year. The use of machine learning algorithms, particularly Random Forest, proved to be effective in analyzing land cover changes in the Indramayu Regency and can contribute to policymaking to enhance food security. © 2024 IEEE.

关键词： Food security Land cover Random Forest SVM

来源：评论

学校读者我要写书评

暂无评论

Semi-Supervised Classifier Ensemble with High-Quality Subspace for High-Dimensional data 13

Semi-Supervised Classifier Ensemble with High-Quality Subspa...

引用

13th international conference on Information science and Technology, ICIST 2023

作者： Li, Guojie Yu, Zhiwen Yang, Kaixiang School of Computer Science & Engineering South China University of Technology Guangzhou China

ISBN: (纸本)9798350313925

Due to the noisy and high-dimensional character of the data combined with a small sample size, semi-supervised classification of high-dimensional data with few labeled samples is a significant problem in machine learning and data mining. To address these challenges, we first propose a novel feature evaluation function that combines the advantages of the Laplacian score and intra-class and inter-class smoothness score to select high-quality features. Subsequently, we combine the selected high-quality features with some randomly selected features to generate a set of high-quality subspaces (HQS). Building upon the concept of HQS, we further propose an ensemble framework (E-HQS) specifically designed for semi-supervised classification tasks on high-dimensional data. We thoroughly validate the effectiveness of our proposed method on 8 real high-dimensional datasets. © 2023 IEEE.

关键词： Feature Selection

来源：评论

学校读者我要写书评

暂无评论

ADFB: Anti-deepfake Framework for Facial Biometric Authentication Systems 6th

ADFB: Anti-deepfake Framework for Facial Biometric Authentic...

引用

6th international conference on Recent Innovations in Computing, ICRIC 2023

作者： Salariya, Rizwan Malhotra, Deepti Department of Computer Science and Information Technology Central University of Jammu Jammu and Kashmir Bagla Suchani181143 India

ISBN: (纸本)9789819778614

Facial biometric authentication, which leverages distinctive facial features for identity verification, is a widely used biometric identification method. However, deepfakes synthetic images or videos generated through sophisticated deep learning techniques pose significant challenges to the reliability of these systems by mimicking real facial data. The presence of deepfake-generated data in facial biometric databases threatens their accuracy, potentially leading to incorrect identifications and fraudulent authentications. This research undertakes a comprehensive study and comparative analysis of AI-driven facial biometric techniques published between 2019 and 2024, including an extensive survey and evaluation of existing deepfake facial detection systems. The paper outlines the workflow of a proposed facial biometric authentication system designed to counter deepfake threats, ensuring secure and reliable identity verification. The model achieved an accuracy of 96.59% on a dataset of real and fake images for deepfake detection and an accuracy of 99.56% on a face images dataset for facial biometric authentication. These results demonstrate the effectiveness of the proposed framework in enhancing the security and reliability of facial biometric systems against deepfake threats. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Predicting Crash Injury Severity in Smart Cities: a Novel Computational Approach with Wide and Deep learning Model

引用

international JOURNAL OF INTELLIGENT TRANSPORTATION SYSTEMS RESEARCH 2023年第1期21卷 240-258页

作者： Niyogisubizo, Jovial Liao, Lyuchao Sun, Qi Nziyumva, Eric Wang, Yongqiang Luo, Linsen Lai, Shukun Murwanashyaka, Evariste Fujian Univ Technol Fujian Prov Univ Key Lab Ind Control & Data Anal Fuzhou 350118 Peoples R China Fujian Univ Technol Fujian Prov Univ Engn Res Ctr Intelligent Driving Fuzhou 350118 Peoples R China Tsinghua Univ Sch Vehicle & Mobil Beijing 100084 Peoples R China Changan Univ Sch Elect & Control Engn Xian 710061 Peoples R China Univ Chinese Acad Sci Inst Rock & Soil Mech Wuhan Peoples R China

Smart cities came out as highly knowledgeable bio-networks, offering intelligent services and innovative solutions to urban problems. With rapid development, urbanization, and population pressure, traffic congestion and collisions are increasing substantially in road highway zones. Recently, traffic collisions have become one of the hugest national health problems in many cities of the world. Hence, crash injury severity prediction is vital for informing responsible authorities and the public to find alternative ways of dealing with its adverse effects, accordingly improving traffic safety and reducing traffic congestion. In predicting crash injury severity, researchers have explored and applied several techniques to aid in traffic injury management. However, the performance of many techniques suffers from some inherent limitations, including overgeneralization, lack of interpretability for humans, and low-performance accuracy. To address these issues, this paper proposes a novel computational framework based on improved wide and deep learning methods to predict accurately crash injury severity in the context of smart cities. On the crash dataset of New Zealand cities from 2000 to 2020, the proposed model has demonstrated better performance in comparison with the benchmark algorithms. Moreover, SHAP (SHapley Additive exPlanation) is employed to interpret the results and analyze the importance of each determinant of crash severity. The proposed method can help to prevent traffic crashes in smart cities and take proactive measures before the occurrence, also trauma centers can refer to this information to dispatch proper emergency service equipment swiftly and assist injuries to get direct medical care regardless of the crash location.

关键词： Crash injury severity Deep learning machine learning Traffic accident prediction Smart cities SHAP

来源：评论

学校读者我要写书评

暂无评论

Evidential Federated learning for Skin Lesion Image Classification 27th

Evidential Federated Learning for Skin Lesion Image Classif...

引用

27th international conference on Pattern Recognition, ICPR 2024

作者： Hendrix, Rutger Proietto Salanitri, Federica Spampinato, Concetto Palazzo, Simone Bagci, Ulas Department of Electrical Electronic and Computer Engineering University of Catania Catania Italy Machine and Hybrid Intelligence Lab Department of Radiology Northwestern University Chicago United States

ISBN: (纸本)9783031781094

We introduce FedEvPrompt, a federated learning approach that integrates principles of evidential deep learning, prompt tuning, and knowledge distillation for distributed skin lesion classification. FedEvPrompt leverages two sets of prompts: b-prompts (for low-level basic visual knowledge) and t-prompts (for task-specific knowledge) prepended to frozen pre-trained Vision Transformer (ViT) models trained in an evidential learning framework to maximize class evidences. Crucially, knowledge sharing across federation clients is achieved only through knowledge distillation on attention maps generated by the local ViT models, ensuring enhanced privacy preservation compared to traditional parameter or synthetic image sharing methodologies. FedEvPrompt is optimized within a round-based learning paradigm, where each round involves training local models followed by attention maps sharing with all federation clients. Experimental validation conducted in a real distributed setting, on the ISIC2019 dataset, demonstrates the superior performance of FedEvPrompt against baseline federated learning algorithms and knowledge distillation methods, without sharing model parameters. In conclusion, FedEvPrompt offers a promising approach for federated learning, effectively addressing challenges such as data heterogeneity, imbalance, privacy preservation, and knowledge sharing. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

machine learning Physical Fatigue Estimation Approach Based on IMU and EMG Wearable Sensors 32

Machine Learning Physical Fatigue Estimation Approach Based ...

引用

32nd Irish conference on Artificial Intelligence and Cognitive science, AICS 2024

作者： Nair, Suraj P. Sica, Marco Tedesco, Salvatore Visentin, Andrea School of Computer Science & IT University College Cork Ireland Tyndall National Institute University College Cork Ireland SFI Insight Centre for Data Analytics University College Cork Ireland

Physical fatigue refers to a state of exhaustion or reduced capacity for physical performance due to prolonged exertion, repetitive movements, or lack of rest. It is a multifaceted condition that can severely impact performance, especially in activities requiring sustained effort, precision, or concentration. In physical tasks, fatigue manifests as a decrease in muscle strength, coordination, and endurance, leading to diminished performance and an increased risk of injury. Detecting physical fatigue is crucial in a variety of domains: professional sports, collaborative robotics, construction, and more. This research introduces a novel framework for predicting fatigue during shoulder movements using data collected from wearable inertial measurement units and electromyography sensors. By integrating the Borg Scale, a subjective measure of perceived exertion, our approach uniquely combines objective sensor data with user-reported fatigue levels, creating a more holistic fatigue assessment model. The primary aim of this study is to develop a predictive model capable of accurately estimating fatigue, as measured by the Borg Scale. An investigation of the best machine learning algorithm for this task ensures that the chosen method provides the most reliable predictions. Furthermore, by systematically reducing the number of sensors and analyzing the impact on model performance, it is possible to find a minimal sensor configuration that maintains the model’s predictive power while reducing complexity and cost. The Ridge Regression model, after hyperparameter tuning, outperformed other models, achieving a mean absolute error of 2.417 in predicting fatigue. This preliminary study shows the potential of integrating data from different inertial and electromyography sensors for fatigue prediction in shoulder movements, with potential applications in occupational safety. © 2024 Copyright for this paper by its authors.

关键词： Regression analysis

来源：评论

学校读者我要写书评

暂无评论

data-DRIVEN AERODYNAMIC OPTIMIZATION OF A 15-STAGE COMPRESSOR WITH TANDEM OUTLET GUIDE VANES 69

DATA-DRIVEN AERODYNAMIC OPTIMIZATION OF A 15-STAGE COMPRESSO...

引用

69th ASME Turbomachinery Technical conference and Exposition (ASME Turbo Expo) (GT)

作者： Liu, Yijie Chen, Jiang Liu, Yi Xiang, Hang Beihang Univ Sch Power & Energy Engn Beijing 100091 Peoples R China

ISBN: (纸本)9780791888087

The aerodynamic design and performance improvement of compressors in gas turbine development, especially in multi-stage axial flow compressors with numerous design parameters, pose challenges relying on manual design experience. To address computational inefficiency in aerodynamic optimization of multi-stage turbomachinery, an aerodynamic optimization platform is developed based on the Direct Free-Form Deformation (DFFD) method. This method, offering full-geometry three-dimensional parameterization with arbitrary degrees of freedom, conducts direct manipulation of control vertex displacement, generating smooth shapes, reducing control variables, and achieving dimensionality reduction. The optimization framework combines the DFFD method with a Pre-screen data-driven single-objective optimization algorithm using differential evolution and a hybrid surrogate model. Introducing machine learning techniques significantly reduces Computational Fluid Dynamics (CFD) simulations, enhancing iteration efficiency. applied to a 15-stage gas turbine compressor, the platform resulted in a 0.37 percentage points efficiency improvement, reducing parameterized variables to 32. Utilizing isentropic efficiency as the pre-screening criterion with a population size of 200 and 560 true evaluations, the entire optimization process took 79.2 hours. This data-driven aerodynamic optimization platform demonstrates efficiency and stability in three-dimensional rapid parameterized optimization for highly loaded multi-stage compressors, presenting a novel approach for such optimization challenges.

关键词： Gas Turbine Aerodynamic Optimization DFFD Parameterization Pre-SADE

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：