检索结果-内蒙古大学图书馆

2025 International Conference on computer, Electrical and Communication engineering, ICCECE 2025

作者： Gaur, Avantika Chamoli, Snigdh Negi, Priyanka Mishra, Preeti Doon University Department of Computer Science Dehradun India Graphic Era Hill University Department of Computer Science and Engineering Dehradun India

ISBN: (纸本)9798331513894

Artificial intelligence (AI)-driven intrusion detection systems (IDS) are crucial in modern technological environments such as virtual domains for identifying complex attack patterns. Although these machine learning (ML) based security models are highly effective, they remain vulnerable to adversarial attacks that target the ML model and exploit its sensitivity by carrying out small perturbations in the input sample. The proposed framework, RAFM, is designed to evaluate the robustness of deep learning-based and gradient-boosting decision tree-based IDS models. It generates adversarial samples through functionality-preserving byte-level manipulations, such as DOS header modifications and shifting techniques, designed to evade detection while maintaining malware functionality. The framework highlights vulnerabilities in these systems by testing against diverse malware families. It offers valuable insights for developing a more resilient AI-based intrusion detection solution for securing virtual machines (VMs) from adversarial attacks. © 2025 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Multi-source ECG Signal-Based Cardiac-Abnormality Pattern Classification with Fast–Slow Learning DNNs

引用

Journal of The Institution of Engineers (India): Series B 2025年 1-10页

作者： Samant, Sandhya Baloni, Dev Department of Computer Science and Engineering Quantum University Uttarakhand Roorkee247667 India

The classification of cardiac-abnormality patterns with ECG data plays a crucial role in the diagnosis as well as treatment and prognosis of diseases related to the human heart. With the advent of deep learning techniques, particularly convolutional, recurrent, and generative neural networks, there has been a significant advancement in the accuracy and efficiency of cardiac-abnormality pattern classification with electrocardiogram (ECG) data. However, with the availability of multitudes of freely available multi-source ECG data today, more attempts are required to develop new models that can handle and perform well on these datasets simultaneously. In this study, an attempt is made to develop a novel deep learning classification model with multi-source ECG dataset for cardiac abnormality pattern classification. The model uses the power of Transformer networks in their ability to produce low inductive bias towards learning representations and the power of Recurrence networks to memorize a compressed representation of a sequence is that it is beneficial for generalization. The transformers are the fast-stream component due to their sensitivity to sensory input and the RNs are slow-stream component due to their long-term memory sustenance. The multi-source ECG dataset is composed of 4 different and popular 12-lead ECG datasets available publicly for research purposes. The proposed model performed satisfactorily overall on a 27-class classification scenario. © The Institution of Engineers (India) 2025.

关键词： Electrocardiography

来源：评论

学校读者我要写书评

暂无评论

Systematic Selection of a Suitable Data Imputation Technique Based on Data Characteristics

引用

SN computer science 2025年第5期6卷 1-19页

作者： Sebastian, Anu Maria Peter, David Artificial Intelligence Lab Department of Computer Science Cochin University of Science and Technology Kerala Cochin India Department of Computer Science and Engineering Indian Institute of Information Technology Kottayam India

Analyzing incomplete data is one of the prime concerns in data analysis. Discarding the missing records or values might result in inaccurate analysis outcomes or loss of helpful information, especially when the size of the data is small. A preferable alternative is to substitute the missing values using imputation such that the substituted values are very close to the actual missing values and this is a challenging task. In spite of the existence of many imputation algorithms, there is no universal imputation algorithm that can yield the best values for imputing all types of datasets. This is mainly because of the dependence of the imputation algorithm on the inherent properties of the data. These properties include type of data distribution, data size, dimensionality, presence of outliers, data dependency among the attributes, and so on. In the literature, there exists no straightforward method for determining a suitable imputation algorithm based on the data characteristics. The existing practice is to conduct exhaustive experimentation using the available imputation techniques with every dataset and this requires a lot of time and effort. Moreover, the current approaches for checking the suitability of imputations cannot be done when the ground truth data is not available. In this paper, we propose a new method for the systematic selection of a suitable imputation algorithm based on the inherent properties of the dataset which eliminates the need for exhaustive experimentation. Our method determines the imputation technique which consistently gives lower errors while imputing datasets with specific properties. Also, our method is particularly useful when the real-world data do not have the ground truth for missing data to check the imputation performance and suitability. Once the suitability of a DI technique is established based on the data properties, this selection will remain valid for another dataset with similar properties. Thus, our method can save time an

关键词： Data analysis Data distribution Data imputation Data preprocessing Imputation performance Machine learning Missing value handling

来源：评论

学校读者我要写书评

暂无评论

Variable-Length Feedback Codes Over Known and Unknown Channels With Non-Vanishing Error Probabilities

引用

IEEE Transactions on Information Theory 2025年第5期71卷 3271-3286页

作者： Yavas, Recep Can Tan, Vincent Y. F. Descartes Program CNRS@CREATE 138602 Singapore National University of Singapore Department of Computer Science Queenstown 117417 Singapore National University of Singapore Department of Mathematics Department of Electrical and Computer Engineering Queenstown 119077 Singapore

We study variable-length feedback (VLF) codes with noiseless feedback for discrete memoryless channels. We present a novel non-asymptotic bound, which analyzes the average error probability and average decoding time of our modified Yamamoto-Itoh scheme. We then optimize the parameters of our code in the asymptotic regime where the average error probability ∈ remains a constant as the average decoding time N approaches infinity. Our second-order achievability bound is an improvement of Polyanskiy et al.’s (2011) achievability bound. We also develop a universal VLF code that does not rely on the knowledge of the underlying channel parameters. Our universal VLF code employs the empirical mutual information as its decoding metric and universalizes the code by Polyanskiy et al. (2011). We derive a second-order achievability bound for universal VLF codes. Our results for both VLF and universal VLF codes are extended to the additive white Gaussian noise channel with an average power constraint. The former yields an improvement over Truong and Tan’s (2017) achievability bound. The proof of our results for universal VLF codes uses a refined version of the method of types and an asymptotic expansion from the nonlinear renewal theory literature. © 1963-2012 IEEE.

关键词： White noise

来源：评论

学校读者我要写书评

暂无评论

A novel hybrid loss-based Encoder–Decoder model for accurate Pulmonary Embolism segmentation

引用

International Journal of Information Technology (Singapore) 2025年第3期17卷 1663-1677页

作者： Vadhera, Renu Sharma, Meghna Department of Computer Science and Engineering The NorthCap University Haryana Gurugram 122017 India

Pulmonary embolism (PE) is diagnosed early and accurately to ensure minimal danger at an advanced stage. This approach extends the advanced techniques for preprocessing, including normalization, slice filtering and resizing. It combines an architecture with skip connections and upsampling toward capturing that extensive detailed contextual information. The loss function used in the model is a combination of SSIM and Dice loss to balance consistency with regard to structural detail and optimization of pixel overlap. It is estimated on a PE challenge dataset (CT scans), where the mean Dice coefficient reached 0.9407, Jaccard similarity 0.9286, sensitivity 0.9324. This methodology outperforms the state of art models. All this shows that the model has a good potential for being applied in clinical practice to automate PE detection. © Bharati Vidyapeeth's Institute of computer Applications and Management 2025.

关键词： CNN based model Deep learning Encoder decoder network Hybrid loss Pulmonary embolism Segmentation UNET

来源：评论

学校读者我要写书评

暂无评论

A Novel Deep Reinforcement Learning Approach for Stability-Based Parallel Disassembly Sequence Planning Problem

引用

SN computer science 2025年第3期6卷 1-19页

作者： Kumar, M. R. Mahesh Ravi, Chandrasekar Department of Computer Science and Engineering National Institute of Technology Puducherry Karaikal India

Parallel Disassembly Sequence Planning (DSP) deals with obtaining the order of disassembling the product with multiple parts disassembling simultaneously. Existing studies derive the product’s disassembly order based on the requirement for liaison data. However, the absence of stability data does not ensure that all the product parts are disassembled without damage. Since the order of retrieval of the product can also be obtained from stability data, this reduces the requirement of an initial precedence graph based on liaison data. The complexity of the problem increases if we adopt different representations for precedence and stability conditions of the product for constraint checking throughout the disassembly process. This paper constructs a stability-based precedence graph based on only the product’s stability data. The developed graph represents the precedence of the stable parts for disassembly, and the same is utilized to formulate the problem as a model-free deep Reinforcement Learning (RL) problem. A novel Exploration–Exploitation (EE) strategy-based Deep Q-Learning (DQL) algorithm is proposed to parallel DSP minimizing direction changes. Our effective reward mechanism accelerates the learning process for all the conditions and objectives of the problem considered. Based on past performance, the proposed algorithm dynamically decides the exploration and exploitation strategy to avoid local optima. We applied the proposed algorithm on five different-scale products and compared the results with uniform-sampling-based model-free deep RL algorithms. It showed significant improvements, particularly in cumulative rewards and disassembling all the parts for large-scale products with early convergence. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025.

关键词： Deep Q-learning Deep Q-Networks Disassembly sequence planning Parallel disassembly Precedence graph Stability graph

来源：评论

学校读者我要写书评

暂无评论

Security to text (S2T): multi-layered based security approaches for secret text content

引用

Multimedia Tools and Applications 2025年第15期84卷 15151-15171页

作者： Kashid, Shamal Awasthi, Lalit K. Berwal, Krishan Department of Computer Science and Engineering National Institute of Technology Uttarakhand IN Srinagar India Faculty of Communication Engineering Military College of Telecommunication Engineering IN Mhow India

In the digital world, text data is produced in an unstructured manner across various communication channels. Extracting valuable information from such data with security is crucial and requires the development of techniques in text mining, information retrieval, and natural language processing (NLP). To solve this issue, we introduce two novel approaches: keyword extraction (KE) and a multi-layered secret sharing scheme (MLSS) to provide security to extracted keywords rather than overall text documents. The KE approach encompasses a sequence of text pre-processing procedures, including tokenization, stopword removal, stemming, and bag of words representation, followed by indexing. This methodology aims to efficiently extract keywords from text datasets. For this research work, we have proposed three datasets, including text messages, whatsapp messages, and electronic mail. MLSS enhances the security of extracted textual data by leveraging text pre-processing steps. This scheme ensures better confidentiality and the non-revealment of sensitive information. Additionally, we evaluate our KE model on our dataset as well as on standard datasets. Experimental results demonstrate the effectiveness of our proposed security to text (S2T) model, which outperform existing state-of-the-art approaches. The model obtains a 100% correlation between the reconstructed text and the original text. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Extraction

来源：评论

学校读者我要写书评

暂无评论

Meta-learning Assisted Graph Representation Learning for Downstream Learning Tasks 5th

Meta-learning Assisted Graph Representation Learning for Do...

引用

5th International Conference on Computing and Network Communications, CoCoNet 2023

作者： Bhat, Nischal R. Kulkarni, Prerana Sanjay Menon, Ananya Meda, Smruthika S. Das, Bhaskarjyoti Department of Computer Science and Engineering PES University Karnataka Bengaluru India Department of Computer Science and Engineering in AI & ML PES University Karnataka Bengaluru India

ISBN: (纸本)9789819747108

The Graph Transformer Network (GTN) is a state-of-the-art solution for graph representation learning, which utilizes the transformer architecture to capture attention and long-range dependencies in a non-euclidean dataset such as a graph. However, even with GTN in a downstream learning task, there are usually a few data points that are harder to classify than others and are at risk of being wrongly classified. Metric-based meta-learning, which encompasses many meta-learning algorithms, proves promising in such scenarios. Through this work, we propose a novel 2-stage pipeline combining a Graph Transformer Network with metric-based meta-learning methodologies to make downstream classification tasks easier. The work described in this article investigates and evaluates this pipeline in two specific learning setups using two Metric Learning frameworks, i.e., a Siamese Network for binary classification of graphs and a Prototypical Network for multi-class classification of nodes, both appended to a representation learning block of Graph Transformer Network. The experiments described in this article conclusively prove the efficacy of such a pipeline both in full-shot and few-shot learning scenarios. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Proposal and Implementation of Neural Network-Based Approach for Secure Speaker Recognition Using Spectrograms

Proposal and Implementation of Neural Network-Based Approach...

引用

2025 International Conference on Pervasive Computational Technologies, ICPCT 2025

作者： Bindlish, Arindam Gujral, Vaibhav Madaan, Rosy Kumar, Praveen Kunskapsskolan Gurgaon India Mriirs Department of Computer Science and Engineering Faridabad India Astana It University Department of Computer Engineering Astana Kazakhstan

ISBN: (纸本)9798331508685

In this era of Artificial Intelligence, research on Speaker Recognition systems has taken the forefront. In this paper, an efficient deep learning pipeline has been proposed. A total of 7500 audio samples from Kaggle's speaker recognition dataset have been used to generate Spectrograms, Mel Spectrograms, and Mel-Frequency Cepstrum Coefficients (MFCC), which in turn have been used to train a ResNet 18 Deep Neural Network from scratch. Using different regularization techniques such as dropout, early stopping, and weight decay, high accuracies were obtained. The results generated from the ResNet by using Mel spectrograms, classic spectrograms, and MFCCs were compared alongside each other to check which one among them gave the best testing accuracy. The proposed approach is designed and implemented with additional feature enhanced by providing a secure feature selection system from audio signals. To implement this, the proposed system is equipped with AES security algorithm on splitting of data set into ratio of 60:30:10. The Classic Spectrograms predicted the speakers with 92.96% accuracy. Mel Spectrograms were comparatively more efficient with 93.75% testing accuracy. The results obtained from the model trained with the MFCC features were the best of all, at 95.2%. Further, the usage of the Advanced Encryption Standard (AES) algorithm has also been proposed for a secure speaker recognition task. © 2025 IEEE.

关键词： Spectrographs

来源：评论

学校读者我要写书评

暂无评论

Evaluation of Hybrid Encryption Method to Secure Healthcare Data 3rd

Evaluation of Hybrid Encryption Method to Secure Healthcare ...

引用

3rd International Conference on Machine Vision and Augmented Intelligence, MAI 2023

作者： Das, Sangjukta Singh, Maheshwari Prasad Namasudra, Suyel Department of Computer Science and Engineering National Institute of Technology Patna Bihar India Department of Computer Science and Engineering National Institute of Technology Agartala Agartala India

ISBN: (纸本)9789819743582

Technological advancements have brought a new era of growth for the healthcare industry. Nowadays, the security of healthcare data and the preservation of user privacy inside smart healthcare systems are being severely compromised by cyber attackers. This work introduces an encryption method that makes use of a hybrid encryption approach by combining both asymmetric and symmetric encryption methods to improve healthcare data security. This scheme solves the key escrow problem and its effectiveness is analyzed through performance evaluations. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Electronic health record

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：