检索结果-内蒙古大学图书馆

KDTL: knowledge-distilled transfer learning framework for diagnosing mental disorders using EEG spectrograms

Neural Computing and Applications 2024年第30期36卷 18919-18934页

作者： Singh, Shreyash Jadli, Harshit Padma Priya, R. Surya Prasath, V.B. Department of Software Systems School of Computer Science and Engineering Vellore Institute of Technology Vellore India Division of Biomedical Informatics Cincinnati Children’s Hospital Medical Center CincinnatiOH45229 United States Department of Pediatrics University of Cincinnati College of Medicine CincinnatiOH45257 United States Department of Biomedical Informatics College of Medicine University of Cincinnati CincinnatiOH45267 United States Department of Computer Science University of Cincinnati CincinnatiOH45221 United States

Electroencephalography (EEG) is a well-known modality in neuroscience and is widely used in identifying and classifying neurological disorders. This paper investigates how EEG data can be used along with knowledge distillation-based deep learning models to detect mental disorders like epilepsy and sleep disorders. The EEG signals are converted into time–frequency plots using short-time Fourier transforms. Further, we propose a novel methodology for using knowledge distillation-based transfer learning (KDTL). Knowledge distillation is becoming quite prevalent in the machine learning field and is associated with various applications in our work;we propose its use in the detection of mental disorders from EEG spectrograms. We convert the EEGs using short-term Fourier transform to obtain time–frequency representation and apply teacher-student by first training a large teacher model and use knowledge distillation to train a student model. In our experiments, we found that ConvNext teacher and MobileNet student combination obtained better results. Our proposed KDTL approach is tested on two datasets with multiple cases, namely the Bonn and ISRUC datasets and obtain 98% and 95% accuracies, respectively. Further experimental results show that the overall KDTL methodology can obtain high classification accuracy across both datasets in binary and multiclass classifications and proves to be better than multiple prior works. Further, our KDTL approach provides a way to train lightweight models which have a smaller number of trainable parameters and thereby constitute lower training time overall. Our proposed KDTL-based approach obtained accurate results in diagnosing mental disorders from EEG spectrograms. Compared to other related methods, KDTL outperformed across tasks with obtained good results in both binary and multiclass classifications. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2024.

关键词： Electroencephalography

来源：评论

学校读者我要写书评

暂无评论

The Sophisticated Prognostication of Migraine Aura Using Machine Learning 4

The Sophisticated Prognostication of Migraine Aura Using Mac...

引用

4th International Conference on Emerging Trends in Networks and computer Communications, ETNCC 2024

作者： Samiullah Rehman, Abdul Rehman, Attique Ur Javaid, Sabeen Ali, Tahir Mohammad Mir, Azka Nirsanametla, Yadaiah University of Sialkot Department of Software Engineering Sialkot Pakistan Gulf University for Science and Technology Department of Computer Science Kuwait Lovely Professional University Division of Research and Development Punjab Phagwara144411 India KKR and KSR Institute of Technology and Sciences Andhra Pradesh Guntur India

ISBN: (纸本)9798350353266

Migraine is one of the most disabling diseases in the world and impacting more than one billion individuals. The symptoms such as intensity, Nausea, Vomit, Phonophobia, Photophobia, Visual, Dysphasia, Dysarthria, Vertigo, Sensory, intense to sound, occurs before the migraine. Migraine drains the quality of persons' life. Furthermore, the study of this research was to train machine learning model on migraine aura dataset using different modern approaches that could assist medical patients before occurring migraine and could also give the indications related to migraine. Likewise, by applying random forest algorithm, we got 99.5 percent accuracy using google colab which will be enough to deploy on application for future Project. Moreover, we leveraged traditional framework to complete the research incorporating, Data collection (Online free Kaggle), Preprocessing, Classification, Model training and Comparison of models. As a result, our peak algorithm precision was including, Naive Bayes 94.52%, decision tree 98.3%, K-nearest neighbours 98.6%, random forest 99.5%, Similarly, Rapid miner and Google Colab software are used for the comparison of algorithms and best one is chosen. © 2024 IEEE.

关键词： Random forests

来源：评论

学校读者我要写书评

暂无评论

Optimizing Generative Adversarial Networks Models for Non-Pneumatic Tire Design: A Comparative Analysis and Evaluation 23

Optimizing Generative Adversarial Networks Models for Non-Pn...

引用

2023 Research in Adaptive and Convergent Systems, RACS 2023

作者： Seong, Ju-Yong Ji, Seung-Min Choi, Dong-Hyun Lee, Sungchul Division of Computer Science and Engineering Sunmoon University Chungcheongnam-do Asan-si Korea Republic of Department of Artificial Intelligence and Software Technology Sunmoon University Chungcheongnam-do Asan-si Korea Republic of

ISBN: (纸本)9798400702280

Pneumatic tires1 are used in a wide range of industries. However, they are difficult to design and rely on the knowledge of experienced designers. To aid in the design of pneumatic tires, this paper suggests the use of Generative Adversarial Networks (GAN) models. The 2000 created images for training were used after removing highly sim-ilar images using Mean Squared Error (MSE) and Structural Similarity Index (SSIM). To find the best model for generating patterns of regularly shaped non-pneumatic tires, GAN, Deep Convolu-tional Generative Adversarial Networks (DCGAN), StyleGANv2 ADA and ProjectedGAN were compared and analyzed. In the qualitative evaluation, the GAN, DCGAN, and StyleGAN v2-ADA models showed that the circle shape was distorted and did not maintain a consistent pattern, but ProjectedGAN showed that the circle remained consistent and the pattern was less distorted than the other GAN models. When evaluating quantitative metrics, ProjectedGAN performed best among several techniques that measure the difference be-tween the generated and actual image distributions. © 2023 ACM.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Towards Understanding and Applying Security Assurance Cases for Automotive Systems

arXiv

引用

arXiv 2024年

作者： Mohamad, Mazen Department of Computer Science & Engineering Division of Interaction Design and Software Engineering Chalmers University of Technology Gothenburg University Gothenburg Sweden

Security Assurance Cases (SAC) are structured bodies of arguments and evidence used to reason about security properties of a certain artefact. SAC are gaining focus in the automotive domain as the need for security assurance is growing due to software becoming a main part of vehicles. Market demands for new services and products in the domain require connectivity, and hence, raise security concerns. Regulators and standardisation bodies started recently to require a structured for security assurance of products in the automotive domain, and automotive companies started, hence, to study ways to create and maintain these cases, as well as adopting them in their current way of working. In order to facilitate the adoption of SAC in the automotive domain, we created CASCADE, an approach for creating SAC which have integrated quality assurance and are compliant with the requirements of ISO/SAE-21434, the upcoming cybersecurity standard for automotive systems. CASCADE was created by conducting design science research study in two iterative cycles. The design decisions of CASCADE are based on insights from a qualitative research study which includes a workshop, a survey, and one-to-one interviews, done in collaboration with our industrial partners about the needs and drivers of work in SAC in industry, and a systematic literature review in which we identified gaps between the industrial needs and the state of the art. The evaluation of CASCADE was done with help of security experts from a large automotive OEM. It showed that CASCADE is suitable for integration in industrial product development processes. Additionally, our results show that the elements of CASCADE align well with respect to the way of working at the company, and has the potential to scale to cover the requirements and needs of the company with its large organization and complex products Copyright © 2024, The Authors. All rights reserved.

关键词： Automotive industry

来源：评论

学校读者我要写书评

暂无评论

Classification of Real and Deepfakes Visual Samples with Pre-trained Deep Learning Models 7th

Classification of Real and Deepfakes Visual Samples with Pre...

引用

Proceedings of the 7th International Conference on Advances in Computing and Data sciences, ICACDS 2023

作者： Nawaz, Marriam Javed, Ali Nazir, Tahira Khan, Muhammad Attique Rajinikanth, Venkatesan Kadry, Seifedine Department of Software Engineering UET Taxila Taxila47050 Pakistan Department of Computing Riphah International University Islamabad Pakistan Department of Computer Science HITEC University Taxila Pakistan Department of Computer Science and Engineering Division of Research and Innovation Saveetha School of Engineering SIMATS Chennai602105 India Department of Applied Data Science Noroff University College Kristiansand4612 Norway Department of Electrical and Computer Engineering Lebanese American University Byblos Lebanon

ISBN: (纸本)9783031379390

Serious security and privacy problems have arisen as a result of significant advancements in the creation of deepfakes. Attackers can easily replace a person’s face with the target person’s face in an image using sophisticated Deep learning (DL) algorithms to spoof their identity. Deepfakes detection algorithms have been proposed in response to the growing concerns about the potential harm caused by deepfakes. However, a reliable deepfakes detector that can keep up with contemporary deepfakes creation techniques is required. In this work, we have proposed an end-to-end methodology for detecting manipulated visual content. We used multiple CNN models i.e., ResNet18, ResNet50, DenseNet65, DenseNet77, and DenseNet100 along with the SVM classifier to compute effective cues from the input facial faces to distinguish between actual and altered content. We have also applied the concept of transfer learning to solve the issue of model over-fitting and improve generalizability against different manipulation algorithms. A comparison study is carried out to evaluate the performance of several feature extractors. Through thorough experiments performed using the Deepfakes Detection Challenge dataset, our results demonstrated that DenseNet100 surpasses the other CNN models by better recognizing deepfakes. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Learning algorithms

来源：评论

学校读者我要写书评

暂无评论

The Future of Differentiated Thyroid Cancer Recurrence Prediction Using a Machine Learning Framework Advancements, Challenges, and Prospects 4

The Future of Differentiated Thyroid Cancer Recurrence Predi...

引用

4th International Conference on Emerging Trends in Networks and computer Communications, ETNCC 2024

作者： Imtiaz, Irsa Rehman, Attique Ur Javaid, Sabeen Ali, Tahir Mohammad Mir, Azka Masud, Mehedi Nirsanametla, Yadaiah University of Sialkot Department of Software Engineering Sialkot Pakistan GULF University for Sciences and Technology Department of Software Engineering Mubarak Al-Abdullah Kuwait Taif University Department of Computer Science Taif Saudi Arabia Lovely Professional University Division of Research and Development Phagwara144411 India KKR and KSR institute of technology and sciences Andhra Pradesh Guntur India

ISBN: (纸本)9798350353266

Differentiated thyroid cancer originates in the thyroid gland, which is positioned in the front of the neck. The thyroid gland generates hormones that control metabolism, heart rate, and other bodily functions. Differentiated thyroid carcinoma (DTC) recurrence is a major challenge in clinical therapy. Early detection and treatment play an important role in reducing the impact of thyroid cancer recurrence. The development of precise prediction algorithms is demanding. The prognosis, diagnosis, and treatment of differentiated thyroid carcinoma have been the subject of extensive investigation. Scholars have investigated diverse methodologies to forecast the likelihood of getting thyroid cancer, refine early identification techniques, and augment therapeutic results. Machine learning (ML) frameworks have emerged as useful tools in this setting, with the potential to improve prediction accuracy and patient outcomes. This paper provides a detailed evaluation of the current state and future directions of DTC recurrence prediction using machine learning. We examine current advances in machine learning techniques, data sources, and feature selection approaches used in DTC recurrence prediction models. Several machine-learning algorithms have been applied. We have suggested the model containing the classifier with the highest accuracy after comparing the accuracy percentage of the various classifiers that were produced. Our suggested model has a 98.17% accuracy rate with Bagging. Finally, we propose solutions to these difficulties and emphasize ML's potential to revolutionize the landscape of DTC recurrence prediction. We hope that this analysis will provide insights into the developing role of machine learning in DTC management and motivate more research in this crucial area. © 2024 IEEE.

关键词： Diseases

来源：评论

学校读者我要写书评

暂无评论

Incremental Instance Segmentation for Cluttered Baggage Threat Detection

Incremental Instance Segmentation for Cluttered Baggage Thre...

引用

2023 IEEE International Conference on Computational Intelligence and Virtual Environments for Measurement Systems and Applications, CIVEMSA 2023

作者： Nasim, Ammara Velayudhan, Divya Ahmed, Abdelfatah Hassan Hassan, Taimur Akcay, Samet Akram, Muhammad Usman Werghi, Naoufel National University of Sciences and Technology Department of Computer and Software Engineering Pakistan Khalifa University KUCARS and C2PS Department of Electrical Engineering and Computer Science United Arab Emirates Abu Dhabi University Department of Electrical and Computer Engineering Abu Dhabi United Arab Emirates Intel Corporation RandD Division United Kingdom

ISBN: (纸本)9781665434454

Identification of contraband items from highly oc-cluded baggage of air travelers is a challenging task even for human experts with very high experience. Many researchers have been working rigorously to develop computer vision-based techniques for baggage screening through X-ray images. Nu-merous machine learning and deep learning-based frameworks have been proposed by researchers in the last two decades. However, all of these techniques face limitations in segmenting prohibited items from highly occluded and cluttered baggage. In this paper, we propose a novel framework based on semantic segmentation to automatically detect concealed prohibited items from X-ray baggage scans. Furthermore, to detect different overlapping instances of the same contraband item, we propose an instance-Aware segmentation model that enables the semantic segmentation model to identify multiple instances of the same threat category through incremental learning without requiring additional overhead. The proposed framework is computationally lighter compared to other similar approaches as it requires min-imal training examples and leverages previous knowledge. The proposed model has outperformed state-of-The-Art instance seg-mentation techniques when tested on publicly available GDXray and SIXray datasets, giving mean average precision scores of 0.50 and 0.47, respectively. In addition, the proposed framework leads other instance segmentation baseline models in terms of mean inference time. © 2023 IEEE.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Optimizing Heart Failure Predictive Accuracy: An Effective Approach Using SMOTE Techniques 4

Optimizing Heart Failure Predictive Accuracy: An Effective A...

引用

4th International Conference on Emerging Trends in Networks and computer Communications, ETNCC 2024

作者： Baber, Ahmed Ahmed, Faizan Rehman, Attique Ur Javaid, Sabeen Alshammeri, Menwa Mir, Azka Kumar, Deepak University of Sialkot Department of Software Engineering Sialkot Pakistan College of Computer and Information Sciences Al Jouf University Department of Computer Science Saudi Arabia Lovely Professional University Division of Research and Development Phagwara144411 India KKR and KSR institute of technology and sciences Andhra Pradesh Guntur India

ISBN: (纸本)9798350353266

Heart failure occurs when the heart's lower cham-bers (ventricles) weaken and can't pump blood effectively. The heart contains 2 ventricles that are left and right. When left ventricular failure occurs, it causes shortness of breath and fatigue, and when right ventricular failure occurs, it causes peripheral and abdominal fluid accumulation. Heart failure is a disorder in breathing and oxygen supply to the whole body. As we know, Heart is responsible for circulating blood all over the body parts. When it stops pumping, blood can no longer circulate effectively causing fluid to build up in the lungs causing difficulty in breathing and in the legs, causing swelling. The fluid buildup can cause shortness of breath and swelling of the legs and feet. As a result, fluctuation in breathing takes place and, in the end, the patient dies. This study investigates the use of machine learning tools, like k-Nearest Neighbors, Decision Trees, Naïve Bayes, and Random Forests, to identify heart failure. We are proposing a model for Heart Failure Prediction using a machine learning algorithm and a dataset to determine the performance in terms of accuracy. As a result, the Random Forest classifier along with the SMOTE operator provides the highest accuracy of 91.12%, precision of 87.88%, recall of 95.39%, and F1 measure of 91.48% among the models considered. This research is a very important contribution to healthcare for effective predictions of this disease, as the number of patients is increasing day by day. The results hold significance for future modifications too in medical diagnostics and helping material for more accurate detection methodologies. © 2024 IEEE.

关键词： Nearest neighbor search

来源：评论

学校读者我要写书评

暂无评论

EPM: Evolutionary Perception Method for Anomaly Detection in Noisy Dynamic Graphs

引用

IEEE Transactions on Knowledge and Data engineering 2025年

作者： Wang, Huan Chen, Junyang Wu, Yirui Leung, Victor C. M. Wang, Di King Abdullah University of Science and Technology Division of Computer Electrical and Mathematical Sciences and Engineering Saudi Arabia Huazhong Agricultural University College of Informatics China Shenzhen University College of Computer Science and Software Engineering China Hohai University College of Computer and Information China

With the rapid expansion of interactions across various domains such as knowledge graphs and social networks, anomaly detection in dynamic graphs has become increasingly critical for mitigating potential risks. However, existing anomaly detection methods often assume noise-free dynamic graphs, overlooking the prevalence of noisy dynamic graphs in real-world applications. Specifically, noisy dynamic graphs affected by structural noises-such as spurious and missing nodes and edges-struggle to consistently provide reliable structural evidence for anomaly detection. To tackle this challenge, we propose an Evolutionary Perception Method (EPM) for identifying anomalous nodes in noisy dynamic graphs by resisting the interference of structural noises. EPM primarily consists of two components: a dynamic fitter and a filtering reviser. The dynamic fitter characterizes the interaction dynamics of nodes that removes and generates links at each period as a multiple superposition state, utilizing various link prediction algorithms to fit evolutionary mechanisms. Additionally, the filtering reviser designs evolutional entropies to quantify the evolutional uncertainty in multiple superposition states, further designing the Kalman filter to optimize these entropies. Extensive experiments show that the proposed EPM method surpasses state-of-the-art approaches in detecting anomalous nodes in noisy dynamic graphs. © 1989-2012 IEEE.

关键词： Kalman filters

来源：评论

学校读者我要写书评

暂无评论

Rethinking Cognition: Morphological Info-Computation and the Embodied Paradigm in Life and Artificial Intelligence

arXiv

引用

arXiv 2024年

作者： Dodig-Crnkovic, Gordana Department of Computer Science and Engineering Chalmers University of Technology Gothenburg Sweden Division of Computer Science and Software Engineering School of Innovation Design and Engineering Mälardalen University Västerås Sweden

This study aims to place Lorenzo Magnani’s Eco-Cognitive Computationalism within the broader context of current work on information, computation, and cognition. Traditionally, cognition was believed to be exclusive to humans and a result of brain activity. However, recent studies reveal it as a fundamental characteristic of all life forms, ranging from single cells to complex multicellular organisms and their networks. Yet, the literature and general understanding of cognition still largely remain human-brain-focused, leading to conceptual gaps and incoherency. This paper presents a variety of computational (information processing) approaches, including an info-computational approach to cognition, where natural structures represent information and dynamical processes on natural structures are regarded as computation, relative to an observing cognizing agent. We model cognition as a web of concurrent morphological computations, driven by processes of self-assembly, self-organization, and autopoiesis across physical, chemical, and biological domains. We examine recent findings linking morphological computation, morphogenesis, agency, basal cognition, extended evolutionary synthesis, and active inference. We establish a connection to Magnani’s Eco-Cognitive Computationalism and the idea of computational domestication of ignorant entities. Novel theoretical and applied insights question the boundaries of conventional computational models of cognition. The traditional models prioritize symbolic processing and often neglect the inherent constraints and potentialities in the physical embodiment of agents on different levels of organization. Gaining a better info-computational grasp of cognitive embodiment is crucial for the advancement of fields such as biology, evolutionary studies, artificial intelligence, robotics, medicine, and more. © 2024, CC BY-NC-ND.

关键词：

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：