检索结果-内蒙古大学图书馆

Methylation Data of Parents in the Prediction of a Preterm Birth: A Machine Learning Approach

SN computer Science 2024年第5期5卷 632页

作者： Jeyananthan, Pratheeba Piyasamara, G.L.D.S. Sachintha, D.C. Department of Computer Engineering Faculty of Engineering University of Jaffna Kilinochchi Sri Lanka

Preterm birth is a serious issue which can affect the whole family, especially the mother both physically and mentally. Further, babies also need to face a lot of short-term and long-term complications, sometimes throughout their life. Annually we have approximately 15 million premature babies worldwide, which is the leading cause of death among children. However, early prediction of a preterm birth can help the clinician to give proper treatments to the mother in order to avoid this complication. Previous biological studies showed that there are epigenetic differences between a preterm baby and a full-term baby, and associations between prenatal risk factors and epigenetic changes. Anyhow it is comparably very hard to get epigenetic data from an infant before labor. Hence, this study analyses the methylation data of father and mother individually in the prediction of the possibility of premature birth using machine learning algorithms such as random forest, support vector machine and K-nearest neighbor. As we have high number of features in this data, mutual information is used to select the relevant features of this study. Different number of features are selected using mutual information and their performances are evaluated using all the three machine learning algorithms to reveal the best number of features. Results indicate that top 15 methylation features taken from the father along with random forest classifier outperform other models with a perfect accuracy (AUC = 1 ± 0.00) in the prediction of premature possibilities. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2024.

关键词： DNA methylation Father-mother–child Feature selection Machine learning models Premature birth

来源：评论

学校读者我要写书评

暂无评论

Artificial Intelligence-Driven Approach for Predicting Maternal Health Risk Factors 9

Artificial Intelligence-Driven Approach for Predicting Mater...

引用

9th South-East Europe Design Automation, computer engineering, computer Networks and Social Media Conference, SEEDA-CECNSM 2024

作者： Hosaain, Mohammad Mobarak Kashem, Mohammod Abdul Nayan, Nasim Mahmud Department of Computer Science and Engineering Gazipur1707 Bangladesh Department of Computer Science and Engineering Dhaka1212 Bangladesh

ISBN: (纸本)9798350342482

The development of machine learning has the potential to significantly improve the identification and treatment of pregnancy-related risks in maternal health. This work uses an extensive dataset to create reliable models that can predict possible health issues with accuracy and reliability. Our models are highly predictive and have transparency and interpretability in their results because of our creative feature engineering, such as featurewiz and chi-square test, and meticulous data processing. The models underwent extensive validation procedures, and in three different categories, each proved better evaluation metrics results alongside accuracy levels above 99% in each class. Ensuring the durability and stability of the models, this remarkable degree of accuracy is further enhanced by ensemble techniques and hyperparameter tuning. Incorporating Explainable AI (XAI) methods, such as LIME and SHAP, has proven crucial in elucidating the models' decision-making process, fostering trust, and enabling their possible use in healthcare environments. The results highlight the potential of ML in predicting maternal health risks and establishing solid foundations for future integration with Medical Cyber-Physical Systems (MCPS), which has the potential to transform the field of predictive analytics in prenatal care completely. © 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

LS-SIFT: Enhancing the Robustness of SIFT During Pose-Invariant Face Recognition by Learning Facial Landmark Specific Mappings

引用

IEEE Access 2024年 12卷 76648-76662页

作者： Lin, Shinfeng D. Linares Otoya, Paulo E. National Dong Hwa University Department of Computer Science and Information Engineering Hualien Taiwan

The proper functioning of many real-world applications in biometrics and surveillance depends on the robustness of face recognition systems against pose, and illumination variations. In this work, we employ ensemble systems in conjunction with local descriptors to address pose-invariant face recognition (PIFR). Facial landmarks are detected during the first step with a two fold usage. The landmark locations are employed to perform head pose classification (HPC). HPC allows to select only the visible landmarks for further processing. Then, local descriptors are extracted from the selected landmarks within a face image. We are proposing a novel learned descriptor (LS-SIFT) to overcome the robustness limitations of SIFT against large viewpoint variability during face recognition. Second, the extracted descriptors are used to train the base learners comprising an ensemble system for each subject in a face database (one ensemble per subject, one base learner per landmark). A novel GMM-based base learner model, named Mahalanobis Similarity (MS), is introduced in this work. Finally, face recognition is performed based on the ensemble systems' outputs from all the subjects in a face database. During the experimental trials, SIFT and LS-SIFT are employed individually for local feature extraction, whereas GMM and MS are used to build the ensemble systems, in an independent manner, for further comparison. The whole PIFR system is tested on CMU-PIE, Multi-PIE, and FERET databases, outperforming most of the state-of-the-art works regarding images with pose angles in the range of pm 90{o}. © 2013 IEEE.

关键词： Face recognition

来源：评论

学校读者我要写书评

暂无评论

Detecting Braess-like Paradoxes in Cascade Failures via Markov Chains 63

Detecting Braess-like Paradoxes in Cascade Failures via Mark...

引用

63rd IEEE Conference on Decision and Control, CDC 2024

作者： Materassi, Donatello Tomsovic, Kevin University of Minnesota Department of Electrical Engineering and Computer Engineering United States Clemson University Department of Electrical and Computer Engineering United States

ISBN: (纸本)9798350316339

The article wants to draw attention to the potential occurrence of Braess-like phenomena in the context of cascade failures, where certain networked system configurations, which might appear more resilient than others because of the presence of additional redundancies, actually exhibit an unexpected higher degree of fragility. To elucidate and attempt to quantify this counterintuitive phenomenon, the article introduces an approach that leverages basic elements of the theory of Markov chains. Indeed, if it is possible to estimate the probabilities of local failures within the system, a comprehensive Markov Embedding Chain can be defined. This chain not only can capture the interplay of sequences of failure events, but can also provide a means to detect and analyze the latent brittleness inherent in seemingly resilient systems. An intrinsic advantage of this approach lies in its model-agnostic nature, allowing its application across a wide array of infrastructures and distributed systems irrespective of the actual underlying failure model. © 2024 IEEE.

关键词： Markov chains

来源：评论

学校读者我要写书评

暂无评论

An Initial Risk Assessment for Multimodal with LSTM-Based Trust Evaluation Framework for Autonomous Vehicle Security

引用

SN computer Science 2025年第2期6卷 1-15页

作者： Renjith, P.N. Balasubramani, S. Ramesh, K. Patnala, Eswar School of Computer Science and Engineering Vellore Institute of Technology Chennai Campus Chennai India Department of Computer Science and Engineering Koneru Lakshmaiah Education Foundation Andhra Pradesh Vaddeswaram India Department of Computer Science and Engineering Sri Krishna College of Engineering and Technology Coimbatore India

This research aims to develop a new approach to increase the safety and reliability of Autonomous Vehicle (AV) through the proposed risk assessment framework, supported by the trust evaluation approach derived from a machine learning algorithm. This work proposes a comprehensive risk assessment model with real-time sensor data, Vehicle-to-Vehicle (V2V), key communication reliability, environmental cognition, and Improved Long Short-Term Memory (I-LSTM). This framework is multidimensional compared to most conventional models, focusing primarily on assessing individual threat factors. Here, the machine learning portion is designed to evaluate and predict trust by quickly processing vast amounts of information from sensors, signals, and prior operation history. Such an approach to trust assessment enables the system to continuously evaluate the dependability of automated vehicle operation and identify threats to security with improved accuracy. The effectiveness of the performance of this method is substantiated by a number of stand-alone and live scenarios, where the model's performance is higher in perceiving possible risks, avoiding threats, and enhancing safety than the conventional approaches. The findings suggest that the safety regulations of the AVs will be improved by using a multi-dimensional risk assessment model and integrating the I-LSTM trust rating with 95.5% accuracy, 96.7% risk identification, and 180 ms response time. This will cause AV and its surrounding environment to be accepted by the public and used on the roads. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025.

关键词： Autonomous vehicles LSTM model Machine learning Risk assessment Trust evaluation

来源：评论

学校读者我要写书评

暂无评论

Designing Secure IoT Systems: A Survey of Existing Approaches and Future Directions 2

Designing Secure IoT Systems: A Survey of Existing Approache...

引用

2nd International Conference on Emerging Trends in Information Technology and engineering, ic-ETITE 2024

作者： Praveen, T.N.V.S. Odugu, Rama Devi Lingamgunta, Sumalatha Jntuk Department of Computer Science &engineering Kakinada India Lakireddy Balireddy College of Engineering Department of Computer Science &engineering Mylavaram India

ISBN: (纸本)9798350328202

The Internet of Things (IoT) has emerged as a transformative technology, connecting a wide array of devices and enabling seamless communication and data exchange. However, the rapid proliferation of IoT devices has brought forth significant security challenges, making them susceptible to various cyber-attacks. This research paper explores the design of security schemes for the detection and prevention of attacks on IoT systems. It examines the major threats faced by IoT deployments and proposes comprehensive security measures to safeguard these interconnected devices and networks. The proposed schemes encompass a multi-layered approach, including intrusion detection systems, secure communication protocols, and encryption techniques, to mitigate the risks and enhance the overall security posture of IoT ecosystems. © 2024 IEEE.

关键词： Data visualization

来源：评论

学校读者我要写书评

暂无评论

Document Summarizer: A Machine Learning approach to PDF summarization 4

Document Summarizer: A Machine Learning approach to PDF summ...

引用

4th International Conference on Evolutionary Computing and Mobile Sustainable Networks, ICECMSN 2024

作者： Peddarapu, Rama Krishna Anogna, Sudagoni Harshini, M.N.S. Sri Sireesha, J. Rathna, K. Guna Department of Computer Science and Engineering Faculty of Computer Sciences and Engineering VNR Vignana Jyothi Institute of Engineering and Technology Telangana Hyderabad India Department of Computer Science and Engineering VNR Vignana Jyothi Institute of Engineering and Technology Telangana Hyderabad India

To address the need for summarizing and extracting information efficiently, this paper highlights the growing challenge posed by the increasing number of PDF files. Reading lengthy documents is a tedious and time-consuming task. To save time and quickly comprehend the key points of a PDF, a PDF summarizer tool has been developed to tackle these issues. In today's professional environments, gathering and managing data from documents is critical. This article introduces an innovative solution called DocSum, which automates the process of summarizing extracted data. Powered by *** Core, the DocSum system allows users to upload PDF documents for processing and receive concise summaries in return. The system features a user-friendly interface that encourages interaction and engagement, utilizing AI and machine learning techniques to streamline document handling. Users can request specific summaries, enabling efficient document management workflows. By empowering users to seamlessly interact with vast amounts of information, DocSum enhances productivity and explores new ways to optimize document management. This solution is ideal for professionals seeking to stay informed and manage data effectively, while also keeping track of advancements in document handling technologies across multiple industries. It is not only helpful for professionals but also helpful for students who need quick revision or examinations, teachers to gain brief knowledge on the concepts. © 2025 The Authors. Published by Elsevier B.V.

关键词： Data handling

来源：评论

学校读者我要写书评

暂无评论

CTRL: Clustering Training Losses for Label Error Detection

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2024年第8期5卷 4121-4135页

作者： Yue, Chang Jha, Niraj K. Princeton University Department of Electrical and Computer Engineering PrincetonNJ08540 United States

In supervised machine learning, use of correct labels is extremely important to ensure high accuracy. Unfortunately, most datasets contain corrupted labels. Machine learning models trained on such datasets do not generalize well. Thus, detecting the label errors can significantly increase their efficacy. We propose a novel framework, called CTRL11CTRL is open-source: https://***/chang-yue/ctrl. (Clustering TRaining Losses for label error detection), to detect label errors in multiclass datasets. It detects label errors in two steps based on the observation that models learn clean and noisy labels in different ways. First, we train a neural network (NN) using the noisy training dataset and obtain the loss curve for each sample. Then, we apply clustering algorithms to the training losses to group samples into two categories: cleanly labeled and noisily labeled. After label error detection, we remove samples with noisy labels and retrain the model. Our experimental results demonstrate state-of-the-art error detection accuracy on both image and tabular datasets under labeling noise. We also use a theoretical analysis to provide insights into why CTRL performs so well. © 2020 IEEE.

关键词： Neural networks

来源：评论

学校读者我要写书评

暂无评论

Image description using tags latent concepts in convolutional neural networks

引用

Multimedia Tools and Applications 2024年 1-24页

作者： Barati, Alireza Farsi, Hassan Mohamadzadeh, Sajad Department of Electrical and Computer Engineering University of Birjand Shahid Avini Highway Birjand Iran

Automatic image annotation systems are employed to describe the visual content of images via tag assignment. Most of these image description systems use deep convolutional neural networks as feature extractors or multi-label classification systems. However, the majority of these systems do not optimally employ latent variables and image concepts within the side information. The idea of this study is to add a latent variable to the CNN-based annotation problem. A Bayesian modeling framework is used for this purpose. High-Level Semantic Concepts of tags are used as an instance of a latent variable to implement the proposed model. The Bayesian modeling result interpretation leads to localizing the entire image annotation process and breaks it down into smaller problems. Thus, a CNN-based sub-network is trained independently for each sub-problem and converted to an expert in annotation for a particular high-level semantic concept. To estimate tags in the test phase, the tags obtained from all sub-networks are combined in weighted form. The weight of each sub-net is proportional to the relation of its high-level semantic concept with the test image content. These weights are obtained from a separate general CNN as a classifier of concept. At the end, a tag refinement method is used to remove noisy tags. The obtained results (F1 values are 40.3, 35.2, and 38.1 for Corel5k, ESPGame, and IAPRTC12 datasets, respectively) confirm the effectiveness of the proposed method and can be comparable with the state-of-the-art annotation systems. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2024.

关键词： Image annotation

来源：评论

学校读者我要写书评

暂无评论

A Dual-Stage Semi-Supervised Pre-Training Approach for Medical Image Segmentation

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2024年第2期5卷 556-565页

作者： Aralikatti, Rajath C Pawan, S.J. Rajan, Jeny National Institute of Technology Karnataka Department of Computer Science and Engineering Surathkal575025 India

Deep neural networks have played a vital role in developing automated methods for addressing medical image segmentation. However, their reliance on labeled data impedes the practicability. Semi-Supervised learning is gaining attention for its intrinsic ability to extract valuable information from labeled and unlabeled data with improved performance. Recently, consistency regularization methods have gained interest due to their efficient learning procedures. They are, however, confined to data or network-level perturbations, negating the benefit of having both forms in a single framework. In light of this, we ask an intriguing but unexplored question: Can we have both network-level and data-level perturbation in the semi-supervised framework? To this end, we present a holistic approach that integrates data-level perturbation in the model pre-training stage, followed by implicit network-level perturbation in the fine-tuning stage. Furthermore, we incorporate networks with manifold learning paradigms throughout the training to facilitate the formation of robust data representations by ensuring local and global semantic affinities adhering to the theory of consensus. Notably, this may be the first attempt in the semi-supervised medical image segmentation archetype to use data and network-level perturbation with a model pre-training strategy. We extensively validated the efficacy of the proposed framework on three benchmark datasets, namely the Automated Cardiac Diagnosis Challenge, ISIC-2018, and Left Atrial Segmentation Challenge datasets, subjected to severely low-sampled labeled data. Notably, in ACDC (4%), ISIC-2018 (5%), and LA (6%) labeled cases, the proposed method outperforms the second-best method by 2.95%, 1.31%, and 0.71% in the Dice Similarity Metric. © 2023 IEEE.

关键词： Feature extraction

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：