检索结果-内蒙古大学图书馆

10th International Conference on Signal Processing and Intelligent Systems, ICSPIS 2024

作者： Azizian, Aref Ahmadian, Rouhollah Ghatee, Mehdi Shamsi, Mostafa Amirkabir University of Technology Dep. of Math. and Computer Science Tehran Iran

ISBN: (纸本)9798331532543

Overfitting and the presence of noisy labels are significant challenges in the training of machine learning models, particularly in complex datasets. This paper introduces a novel checkpointing method designed to mitigate overfitting while maintaining model calibration in the presence of noise. Our approach leverages the entropy of the two largest predicted probabilities from the model's output to determine when to save the model during training. This entropy-based criterion allows for effective checkpointing without necessitating a separate validation set. We evaluate our method across various datasets characterized by different noise levels and model architectures. The results demonstrate that our entropy-based checkpointing offers a more robust solution for managing model performance during training. This work contributes to the ongoing efforts to enhance model reliability in real-world applications where data imperfections are prevalent. © 2024 IEEE.

关键词： Entropy-Based Metric Noisy Labels Overfitting Prevention

来源：评论

学校读者我要写书评

暂无评论

On the classification of weather based on the production of photovoltaic installations 24

On the classification of weather based on the production of ...

引用

24th IEEE International Conference on Data Mining Workshops, ICDMW 2024

作者： Parczyk, Pawel Puslecki, Tobiasz Burduk, Robert Wroclaw University of Science and Technology Dep. of Systems and Computer Networks Wroclaw Poland

ISBN: (纸本)9798331530631

In recent years, there has been an energy transition in which fossil fuel-fired power plants are being replaced by renewable energy sources such as photovoltaics. While the influence of factors such as location and method of installation or the sun's position above the horizon is deterministic, the weather factor is rather random. Weather data can be easily obtained from nearby weather stations, but some weather phenomena are so dynamic that data obtained in this way can become useless. When considering the problem of classifying weather conditions based on production data of photovoltaic installations, we propose two new features extracted from the production data that describe the overall level of sunshine and the variability of sunlight. Three supervised learning methods were used to classify the dataset: CNN, Random Forest classifier and Decision Tree Classifier. Reference values were obtained by generating classification results from the raw output and compared with those generated from the two new features. The experiments showed a statistically significant advantage for the models fitted to the data extended with the new features. © 2024 IEEE.

关键词： Photovoltaic (PV) Renewable Energy Sources (RES) Weather Classification

来源：评论

学校读者我要写书评

暂无评论

When Do Neuromorphic Sensors Outperform cameras? Learning from Dynamic Features 57

When Do Neuromorphic Sensors Outperform cameras? Learning fr...

引用

57th Annual Conference on Information Sciences and Systems, CISS 2023

作者： Deniz, Daniel Ros, Eduardo Fermller, Cornelia Barranco, Francisco University of Granada Dep. of Computer Architecture and Technology Spain University of Maryland Perception and Robotics Group United States

ISBN: (纸本)9781665451819

Visual event sensors only output data when changes in the scene happen at very high frequency. This allows for smartly compressing the scene and thus, enabling real-time operation. Despite these advantages, works in the literature have struggled to show a niche for these event-driven approaches compared to conventional sensors, especially when focusing on accuracy performance. In this work, we show a case that fully exploits event sensor advantages: for manipulation action recognition, learning events achieves superior accuracy and time performance. The recognition of manipulation actions requires extracting and learning features from the hand pose and trajectory and the interaction with the object. As shown in our work, approaches based on event sensors are the best fit for extracting these dynamic features contrarily to conventional approaches based on full frames, which mostly extract spatial features and need to reconstruct the dynamics from sequences of frames. Finally, we show how using a tracker to extract the features to be learned only around the hand, we obtain an approach that is scene- and almost object-agnostic and achieves good time performance with a very limited impact in accuracy. © 2023 IEEE.

关键词： Palmprint recognition

来源：评论

学校读者我要写书评

暂无评论

Real-Time Audio Analysis for Detection of Apnea Intervals in Health-Care System 10

Real-Time Audio Analysis for Detection of Apnea Intervals in...

引用

10th International Conference on Signal Processing and Intelligent Systems, ICSPIS 2024

作者： Vahdani, Alireza Rahmati, Mahdieh Sharifi, Mehdi Ghatee, Mehdi Amirkabir University of Technology Iran Dep. Of Math. And Computer Science Tehran Iran

ISBN: (纸本)9798331532543

Accurate detection of respiratory pauses is crucial in patients undergoing cataract surgery under sedation, as undetected pauses can lead to serious complications. Conventional monitoring techniques, such as pulse oximetry and capnography, often fail to detect respiratory pauses in a timely manner, limiting their effectiveness in high-stakes clinical settings. To address this gap, we propose an automated method leveraging tracheal sound analysis using a deep learning model based on ResNet18. Our approach incorporates signal processing techniques, noise reduction, Mel-spectrogram extraction, and data augmentation, to optimize the model's performance. The model was evaluated using multiple performance metrics, including AUC of the precision-recall curve, F1 score, etc. It achieved an AUC of 92.5% and an F1 score of 89% for the apnea class, demonstrating the system's potential for precise, real-time respiratory monitoring in clinical environments. © 2024 IEEE.

关键词： Image coding

来源：评论

学校读者我要写书评

暂无评论

ENHANCING E-HEALTH CYBERSECURITY AND RESILIENCE: SHIFTING FROM MONOLITHIC TO MICROSERVICES architecture

UPB Scientific Bulletin, Series C: Electrical Engineering an...

引用

UPB Scientific Bulletin, Series C: Electrical Engineering and computer Science 2025年第1期87卷 21-34页

作者： Contașel, Cristian Rughiniș, Răzvan Trancă, Dumitru Cristian Țurcanu, Dinu Dept. of Computer Science and Engineering National University of Science and Technology POLITEHNICA Bucharest Romania Dep. of Software Engineering and Automatics Technical University of Moldova Chișinău Moldova

This study provides an innovative architectural model for e-Health systems that aims to improve cyber resilience while maintaining high availability under fluctuating traffic loads. We examined typical cybersecurity incidents in the field of e-Health, their correlations with architectural defects, and frequent design patterns in currently operational systems. A testing approach based on our research finds those weaknesses and proposes viable fixes. In this paper, a comprehensive support strategy for transitioning from conventional monolithic architectures to microservices is presented. This change makes use of cloud computing's vertical and horizontal scalability to maximize resource utilization while ensuring system reliability. We also discuss dep.oyment ideas for the new microservices, focusing on operational resilience and cybersecurity in e-Health environments. © 2025, Politechnica University of Bucharest. All rights reserved.

关键词： Electronic health record

来源：评论

学校读者我要写书评

暂无评论

Design and Implementation of Minkowski Feature Selection for Machine Learning Techniques 5

Design and Implementation of Minkowski Feature Selection for...

引用

5th International Conference on Information technology, Applied Mathematics and Statistics, ICITAMS 2023

作者： Hakim, Ekram Al-Shammary, Dhiah Mahdi, Ahmed M. University Of Al-Qadisiyah College Of Computer Science And Information Technology Dep. Computer Science Diwaniyah Iraq

ISBN: (纸本)9798350335118

This paper proposes a design and implementation of a highly efficient feature selection based on Minkowski's mathematical similarity for machine learning. Moreover, the development of the Minkowski feature selection can be general for intelligent datasets and machine learning. Intelligent optimizers and feature selection represent an advanced stage in machine learning and innovative computer applications by reducing the overall number of features to be included in the classification function process. This paper analyses Minkowski's selection of best factors for EEG Bonn University data. Minkowski cancels the random movement and systematically limits the number of features and its point-to-point movement in order to promote exploration and reduce the probability of falling into stagnation. The main measures to analyze this research are Accuracy, precision, Recall, and F1score assessment. Moreover, the accuracy of Minkowski's selection provided better classification results than some of the other approaches (100%) © 2023 IEEE.

关键词： Feature Selection

来源：评论

学校读者我要写书评

暂无评论

Introduction to the HICSS-57 Software technology and Software Development Minitrack

Proceedings of the Annual Hawaii International Conference on...

引用

Proceedings of the Annual Hawaii International Conference on System Sciences 2024年 7697-7699页

作者： Majchrzak, Tim A. Grønli, Tor-Morten Kaindl, Hermann University of Agder Kristiansand Norway Kristiania University College Dep. Technology Oslo Norway TU Wien Inst. of Computer Tech. Vienna Austria

来源：评论

学校读者我要写书评

暂无评论

Gradual Divergence for Seamless Adaptation: A Novel Domain Incremental Learning Method 41

Gradual Divergence for Seamless Adaptation: A Novel Domain I...

引用

41st International Conference on Machine Learning, ICML 2024

作者： Jeeveswaran, Kishaan Arani, Elahe Zonooz, Bahram Dep. of Mathematics and Computer Science Eindhoven University of Technology Netherlands Wayve Technologies Ltd London United Kingdom

Domain incremental learning (DIL) poses a significant challenge in real-world scenarios, as models need to be sequentially trained on diverse domains over time, all the while avoiding catastrophic forgetting. Mitigating representation drift, which refers to the phenomenon of learned representations undergoing changes as the model adapts to new tasks, can help alleviate catastrophic forgetting. In this study, we propose a novel DIL method named DARE, featuring a three-stage training process: Divergence, Adaptation, and REfinement. This process gradually adapts the representations associated with new tasks into the feature space spanned by samples from previous tasks, simultaneously integrating task-specific decision boundaries. Additionally, we introduce a novel strategy for buffer sampling and demonstrate the effectiveness of our proposed method, combined with this sampling strategy, in reducing representation drift within the feature encoder. This contribution effectively alleviates catastrophic forgetting across multiple DIL benchmarks. Furthermore, our approach prevents sudden representation drift at task boundaries, resulting in a well-calibrated DIL model that maintains the performance on previous tasks. Copyright 2024 by the author(s)

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

BEYOND UNIMODAL LEARNING: THE IMPORTANCE OF INTEGRATING MULTIPLE MODALITIES FOR LIFELONG LEARNING 3

BEYOND UNIMODAL LEARNING: THE IMPORTANCE OF INTEGRATING MULT...

引用

3rd Conference on Lifelong Learning Agents, CoLLAs 2024

作者： Sarfraz, Fahad Zonooz, Bahram Arani, Elahe Dep. of Mathematics and Computer Science Eindhoven University of Technology Netherlands Wayve Technologies Ltd London United Kingdom

While humans excel at continual learning (CL), deep neural networks (DNNs) exhibit catastrophic forgetting. A salient feature of the brain that allows effective CL is that it utilizes multiple modalities for learning and inference, which is underexplored in DNNs. Therefore, we study the role and interactions of multiple modalities in mitigating forgetting and introduce a benchmark for multimodal continual learning. Our findings demonstrate that leveraging multiple views and complementary information from multiple modalities enables the model to learn more accurate and robust representations. This makes the model less vulnerable to modality-specific regularities and considerably mitigates forgetting. Furthermore, we observe that individual modalities exhibit varying degrees of robustness to distribution shift. Finally, we propose a method for integrating and aligning the information from different modalities by utilizing the relational structural similarities between the data points in each modality. Our method sets a strong baseline that enables both single- and multimodal inference. Our study provides a promising case for further exploring the role of multiple modalities in enabling CL and provides a standard benchmark for future research.1 © 2024 CoLLAs. All Rights Reserved.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Study on the Robotic Arm Trajectory Compensation System with Deviation Sensing 3

Study on the Robotic Arm Trajectory Compensation System with...

引用

3rd International Conference on Mechatronics and Automation technology, ICMAT 2024

作者： Liu, Chao-Chung Liu, Jun-Chi Liu, Chao-Shu Dep. of Computer Science Zhaoqing University Zhaoqing526061 China Dep. of Industrial Design National United University Miaoli360301 Taiwan Dep. of Mechanical Engineering National Kaohsiung University of Science and Technology Kaohsiung80778 Taiwan

ISBN: (纸本)9781643685649

Robotic arms play an important role in the automation industry. Robotic arms plan processing paths based on the shape of the workpiece. However, the processing path often deviates due to inaccurate placement of the workpiece. To solve the issue of deviations, this paper develops a trajectory compensation system for a robotic arm with offset perception. This system uses the concept of workpiece boundary projection and designs an offset calculation algorithm by comparing the dimensions of the workpiece drawings, correcting errors caused by missing corners of the workpiece. In terms of operation, the system uses photoelectric sensors to quickly scan the workpiece boundary and immediately calculate the displacement and rotational offset of the workpiece. Subsequently, when deviations occur due to different placement positions of the workpiece, the robotic arm's motion path and action must be corrected and compensated to maintain precision. Tests on selected workpieces show that the corrected rotational angle error is less than 0.4 degrees, and the translational error is less than 0.7 millimeters. Since these correction error values are lower than the 1.0 mm error threshold for the robotic arm's trajectory tracking control. The offset-sensing robotic arm trajectory compensation system developed in this study can satisfy the requirements of smart automation production lines and arrive at the manufacturer's application standards. © 2024 The Authors.

关键词： Robotic arms

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：