检索结果-内蒙古大学图书馆

Classification of Conversational Sentences Using an Ensemble Pre-Trained Language Model with the Fine-Tuned Parameter

引用

Computers, Materials & Continua 2024年第2期78卷 1669-1686页

作者： R.Sujatha K.Nimala Department of Networking and Communications SRM Institute of Science and TechnologyKattankulathurChengalpattuTamilnadu603203India

Sentence classification is the process of categorizing a sentence based on the context of the *** categorization requires more semantic highlights than other tasks,such as dependence parsing,which requires more syntactic *** existing strategies focus on the general semantics of a conversation without involving the context of the sentence,recognizing the progress and comparing *** ensemble pre-trained language model was taken up here to classify the conversation sentences from the conversation *** conversational sentences are classified into four categories:information,question,directive,and *** classification label sequences are for analyzing the conversation progress and predicting the pecking order of the *** of Bidirectional Encoder for Representation of Transformer(BERT),Robustly Optimized BERT pretraining Approach(RoBERTa),Generative Pre-Trained Transformer(GPT),DistilBERT and Generalized Autoregressive Pretraining for Language Understanding(XLNet)models are trained on conversation corpus with *** tuning approach is carried out for better performance on sentence *** Ensemble of Pre-trained Language Models with a Hyperparameter Tuning(EPLM-HT)system is trained on an annotated conversation *** proposed approach outperformed compared to the base BERT,GPT,DistilBERT and XLNet transformer *** proposed ensemble model with the fine-tuned parameters achieved an F1_score of 0.88.

关键词： Bidirectional encoder for representation of transformer conversation ensemble model fine-tuning generalized autoregressive pretraining for language understanding generative pre-trained transformer hyperparameter tuning natural language processing robustly optimized BERT pretraining approach sentence classification transformer models

来源：评论

学校读者我要写书评

暂无评论

Flameshift Protocol: Revolutionizing Interoperability with Dynamic Asset Recycling for Cross-Chain communications

引用

SN Computer science 2024年第6期5卷 773页

作者： Vijayalakshmi, C. Florence, S. Metilda Department of Networking and Communications SRM Institute of Science and Technology

The burgeoning landscape of blockchain technology has ushered in an era of unprecedented innovation and opportunity. However, one of the most persistent challenges that emerge in this decentralized ecosystem is the lack of efficient and secure interoperability mechanisms across different blockchain networks. The FlameShift Protocol introduces a ground-breaking approach to address this challenge, enabling seamless cross-chain communications through a novel Dynamic Asset Recycling (DAR) mechanism. This paper delves into the intricacies of the FlameShift Protocol, illustrating its capability to not only facilitate secure and efficient cross-chain transactions but also to enhance the liquidity and utility of digital assets across disparate blockchain ecosystems. At the heart of this protocol is the DAR mechanism, which employs a unique auto-burn and reclaim process. This process ensures that when assets are transferred from one chain to another, they are temporarily locked in the source chain, and an equivalent value is minted in the destination chain. This mechanism is designed to maintain a balanced ledger across chains without the need for centralized intermediaries, thereby mitigating the risk of double-spending and other fraudulent activities. Furthermore, the FlameShift Protocol incorporates advanced cryptographic techniques and smart contract functionalities to ensure that the asset transfer process is not only secure but also transparent and verifiable. This paper presents a comprehensive analysis of the protocol’s architecture, security features, and the potential implications for the future of blockchain interoperability. By offering a scalable, secure, and efficient framework for cross-chain communications, the FlameShift Protocol paves the way for a more interconnected and versatile blockchain ecosystem. This study aims to contribute to the ongoing discourse on blockchain interoperability, providing insights and practical solutions for developers, researcher

关键词： Blockchain technology Cross-chain communications Dynamic asset recycling (DAR) FlameShift protocol Interoperability mechanisms

来源：评论

学校读者我要写书评

暂无评论

Hybrid Seagull and Whale Optimization Algorithm-Based Dynamic Clustering Protocol for Improving Network Longevity in Wireless Sensor Networks

引用

China communications 2024年第10期21卷 113-131页

作者： P.Vinoth Kumar K.Venkatesh Department of Networking and Communications School of ComputingCollege of Engineering and TechnologySRM Institute of Science and TechnologyKattankulathur 603203India

Energy efficiency is the prime concern in Wireless Sensor Networks(WSNs) as maximized energy consumption without essentially limits the energy stability and network lifetime. Clustering is the significant approach essential for minimizing unnecessary transmission energy consumption with sustained network lifetime. This clustering process is identified as the Non-deterministic Polynomial(NP)-hard optimization problems which has the maximized probability of being solved through metaheuristic *** adoption of hybrid metaheuristic algorithm concentrates on the identification of the optimal or nearoptimal solutions which aids in better energy stability during Cluster Head(CH) selection. In this paper,Hybrid Seagull and Whale Optimization Algorithmbased Dynamic Clustering Protocol(HSWOA-DCP)is proposed with the exploitation benefits of WOA and exploration merits of SEOA to optimal CH selection for maintaining energy stability with prolonged network lifetime. This HSWOA-DCP adopted the modified version of SEagull Optimization Algorithm(SEOA) to handle the problem of premature convergence and computational accuracy which is maximally possible during CH selection. The inclusion of SEOA into WOA improved the global searching capability during the selection of CH and prevents worst fitness nodes from being selected as CH, since the spiral attacking behavior of SEOA is similar to the bubble-net characteristics of WOA. This CH selection integrates the spiral attacking principles of SEOA and contraction surrounding mechanism of WOA for improving computation accuracy to prevent frequent election process. It also included the strategy of levy flight strategy into SEOA for potentially avoiding premature convergence to attain better trade-off between the rate of exploration and exploitation in a more effective manner. The simulation results of the proposed HSWOADCP confirmed better network survivability rate, network residual energy and network overall throughput on par wi

关键词： clustering energy stability network lifetime seagull optimization algorithm(SEOA) whale optimization algorithm(WOA) wireless sensor networks(WSNs)

来源：评论

学校读者我要写书评

暂无评论

A new perspective exploration of machine learning algorithms for defending Side-Channel attacks

引用

Multimedia Tools and Applications 2025年第16期84卷 17021-17040页

作者： Ramakrishnan, Vedhavathy Thoguluva Seethapathy, Murugaanandam Sundar, Ramesh Dhandapani, Saveetha Munusamy, Sundarrajan Balasubrahmaniam, Lakshmi Dhevi Department of Networking and Communications School of Computing Faculty of Engineering and Technology SRM Institute of Science and Technology Kattankulathur India

Machine learning algorithms are used in various real-time applications, where security is one of the major problems. Security is applied in various aspects of the application in cloud computing. One of the security issues is a side-channel attack, which is not a common problem that can be solved by a point solution. The emerging computing industries need a better solution for side-channel attack detection since the amount of data used in communication and sharing on the Internet is enormous. Several old research works have concluded that machine learning algorithms are highly suitable for data analytics. Hence, this paper has aimed to propose a better machine-learning model for side-channel attack detection. This paper explains some machine learning algorithms from a technical point of view. The explanation identifies that the Convolution Neural Network algorithm is selected as a better algorithm for side-channel attack detection. The proposed CNN model has experimented with Python software to verify the results. Since CNN is more efficient in data analytics, the analysis result produces more accurate predictions than earlier algorithms. The performance of the proposed CNN is compared with the earlier algorithms for evaluating it. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Data Analytics

来源：评论

学校读者我要写书评

暂无评论

Medical image security by crypto watermarking using enhanced chaos and fruit fly optimization algorithm with SWT and SVD

引用

Multimedia Tools and Applications 2024年第27期83卷 70451-70476页

作者： Abirami, R. Malathy, C. Department of Networking and Communications School of Computing SRM Institute of Science and Technology Kattankulathur603203 India

Telemedicine is the field that uses medical images for diagnosing various diseases. The transmitting and storing of medical images via the cloud-based network must meet several stringent criteria, including confidentiality, validity, and security, and are viewed as very sensitive, irrespective of the image processing context. Medical image copyright protection has been essential since little alterations can even put the lives of patients in danger. Hence numerous significant watermarking methods need to be developed. Watermarking conceals sensitive data by embedding it in a harmless medium, such as a cover. The challenging issue of quick and extremely safe image encryption can be solved more effectively with chaos-based cryptography. Thus the image is encrypted by using enhanced chaos with the Fruit Fly Optimization Algorithm (FFOA). The work proposes an efficient medical image watermarking approach using the Two-level Stationary Wavelet Transform (SWT) and Singular Value Decomposition (SVD) technique on chaotic encrypted medical images. Finally, the proposed method is compared with existing methods to demonstrate higher performance. To check robustness and imperceptibility results, the work is examined under various attacks and produces good results in Peak Signal-to-Noise Ratio (PSNR) and Normalized Cross Correlation (NCC) measures. For different kinds of medical images, the PSNR of the suggested methodology is greater than 40 dB, and NCC values are close to 1 illustrating the technique's superior efficacy. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Singular value decomposition

来源：评论

学校读者我要写书评

暂无评论

Multimodal deep learning for chronic kidney disease prediction: leveraging feature selection algorithms and ensemble models

引用

International Journal of Computers and Applications 2023年第10期45卷 647-659页

作者： Subashini, N.J. Venkatesh, K. Department of Networking and Communications SRM Institute of Science and Technology Chennai India

This research presents an advanced approach to enhance disease diagnosis using imbalanced medical datasets. Feature selection techniques, LASSO and Relief, are applied to identify relevant features from the UCI dataset and missing values are handled appropriately. To address the class imbalance, SMOTEENN is used, creating a new combined dataset with selected features. Three deep learning models, FNNs, LSTMs, and GBMs, are employed and trained on the combined dataset, achieving remarkable accuracy (1.0). Evaluating the models on LASSO and Relief datasets independently, FNN/MLP obtains perfect accuracy, GBM performs well (0.9888 on LASSO and 1.0 on Relief), and LSTM shows good results (0.9663 on LASSO and 1.0 on Relief). This study demonstrates the effectiveness of combining LASSO and Relief for feature selection and highlights the impact of SMOTEENN on model performance. The achieved accuracy with all models on the combined dataset showcases deep learning's potential for accurate disease diagnosis even with imbalanced data, offering promising insights for robust medical diagnosis systems. © 2023 Informa UK Limited, trading as Taylor & Francis Group.

关键词： Feature Selection

来源：评论

学校读者我要写书评

暂无评论

Heterogeneous Convolutional Neural Networks for Emotion Recognition Combined with Multimodal Factorised Bilinear Pooling and Mobile Application Recommendation

International Journal of Interactive Mobile Technologies

引用

International Journal of Interactive Mobile Technologies 2023年第16期17卷 129-142页

作者： Saisanthiya, D. Supraja, P. Department of Networking and Communications SRM University Tamil Nadu Kattankulathur India Department of Networking and Communications in SRM Institute of Science and Technology Kattankulathur India

The field of emotion recognition has garnered considerable interest due to its diverse applications in mental health, personalised advertising and enhancing user experiences. This research paper introduces a unique and innovative method for emotion recognition by integrating heterogeneous convolutional neural networks (CNNs) with multimodal factorised bilinear pooling. Furthermore, the paper also incorporates the integration of mobile application recommendations as part of the overall approach. The proposed method leverages the power of CNNs to extract high-level features from different modalities, including facial expressions, speech signals and physiological signals. By using heterogeneous CNNs, each modality is processed independently to capture modality-specific emotional cues effectively. To fuse the extracted features, multimodal factorised bilinear pooling is employed, which captures the complex interactions between different modalities while reducing the computational complexity. This pooling technique efficiently combines the modality-specific features, resulting in a compact and discriminative representation of the emotional state. In addition to emotion recognition, this paper also introduces the integration of mobile app recommendations. By leveraging the recognised emotion, the system recommends relevant mobile applications that are tailored to the user’s emotional state. This integration enhances user experience and facilitates emotion regulation through the utilisation of appropriate mobile apps. Experimental evaluations are conducted on benchmark emotion recognition datasets, including the DEAP and MAHNOB_HCI datasets. The findings of the study highlight the effectiveness of the proposed methodology in terms of accuracy and robustness, surpassing existing approaches in the field. Additionally, the integration of the mobile app recommendation system showcases encouraging outcomes by offering personalised recommendations tailored to the user’s emotiona

关键词： Recommender systems

来源：评论

学校读者我要写书评

暂无评论

Informal settlement mapping from very high-resolution satellite data using a hybrid deep learning framework

引用

Neural Computing and Applications 2024年第4期37卷 2877-2889页

作者： Prabhu, Ravi Department of Networking and Communications School of Computing SRM Institute of Science and Technology Tamilnadu Kattankulathur Chennai India

Informal settlements are becoming increasingly common in the global south, and their locations are often inaccurately represented in government statistics and maps. In remote sensing (RS), classifying informal settlements from very high-resolution satellite images presents a significant challenge due to their unique spectral signatures. This article proposes a novel hybrid framework, multi-shape-multisize-morphological profiles-dual-channel convolutional neural network (MSh-MSi-MP-DC-CNN) architecture, which fully exploits the feature extraction capabilities of both morphological profiles (MP) and convolutional neural networks (CNN) for detecting informal settlements from RS images. Three different images were used to evaluate the proposed MSh-MSi-MP-DC-CNN architecture. One was taken by the WorldView-2 dataset (1.82 m resolution) of Tiruppur in India, and the other two were taken by the WorldView-3 dataset (0.31 m resolution) of Kibera in Kenya and Medellin in Colombia. By extracting both spatial and structural features, the proposed CNN architecture proves robust and scalable for RS images of complex urban slum environments. It also generates significantly higher classification accuracy with less computing time compared to other state-of-the-art CNN methods considered in this work. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2024.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Neuro-facial Fusion for Emotion AI: Improved Federated Learning GAN for Collaborative Multimodal Emotion Recognition

引用

IEIE Transactions on Smart Processing and Computing 2024年第1期13卷 61-68页

作者： Saisanthiya, D. Supraja, P. Department of Networking and Communications Faculty of Engineering and Technology SRM Institute of Science and Technology Tamil Nadu Kattankulathur603203 India

In the context of artificial intelligence technology, an emotion recognition (ER) has numerous roles in human lives. On the other hand, the emotion recognition techniques most currently used perform poorly in recognizing emotions, which limits their wide spread use in practical applications. A Collaborative Multimodal Emotion Recognition through Improved Federated Learning Generative Adversarial Network (MER-IFLGAN) for facial expressions and electro encephalogram (EEG) signals was proposed to reduce this issue. Multi-resolution binarized image feature extraction (MBIFE) was initially used for facial expression feature extraction. The EEG features were extracted using the Dwarf Mongoose Optimization (DMO) algorithm. Finally, IFLGAN completes the Emotion recognition task. The proposed technique was simulated in MATLAB. The proposed technique achieved 25.45% and 19.71% higher accuracy and a 32.01% and 39.11% shorter average processing time compared to the existing models, like EEG based Cross-subject and Cross-modal Model (CSCM) for Multimodal Emotion Recognition (MERCSCM) and Long-Short Term Memory Model (LSTM) for EEG Emotion Recognition (MERLSTM), respectively. The experimental results of the proposed model shows that complementing EEG signals with the features of facial expression could identify four types of emotions: happy, sad, fear, and neutral. Further more, the IFLGAN classifier can enhance the capacity of multimodal emotion recognition. © 2024 institute of Electronics Engineers of Korea. All rights reserved.

关键词： Emotion Recognition

来源：评论

学校读者我要写书评

暂无评论

Feature Selection with Deep Belief Network for Epileptic Seizure Detection on EEG Signals

引用

Computers, Materials & Continua 2023年第5期75卷 4101-4118页

作者： Srikanth Cherukuvada R.Kayalvizhi Department of Networking and Communications School of ComputingSRM Institute of Science and TechnologyKattankulathurChennai603203India

The term Epilepsy refers to a most commonly occurring brain disorder after a *** identification of incoming seizures significantly impacts the lives of people with *** detection of epileptic seizures(ES)has dramatically improved the life quality of the *** Electroencephalogram(EEG)related seizure detection mechanisms encountered several difficulties in *** EEGs are the non-stationary signal,and seizure patternswould changewith patients and recording ***,EEG data were disposed to wide noise varieties that adversely moved the recognition accuracy of *** intelligence(AI)methods in the domain of ES analysis use traditional deep learning(DL),and machine learning(ML)*** article introduces an Oppositional Aquila Optimizer-based Feature Selection with Deep Belief Network for Epileptic Seizure Detection(OAOFS-DBNECD)technique using EEG *** primary aim of the presented OAOFS-DBNECD system is to categorize and classify the presence of *** suggested OAOFS-DBNECD technique transforms the EEG signals *** format at the initial ***,the OAOFS technique selects an optimal subset of features using the preprocessed *** seizure classification,the presented OAOFS-DBNECD technique applies Artificial Ecosystem Optimizer(AEO)with a deep belief network(DBN)*** extensive range of simulations was performed on the benchmark dataset to ensure the enhanced performance of the presented OAOFS-DBNECD *** comparison study shows the significant outcomes of the OAOFS-DBNECD approach over other *** addition,the result of the suggested approach has been evaluated using the CHB-MIT database,and the findings demonstrate accuracy of 97.81%.These findings confirmed the best seizure categorization accuracy on the EEG data considered.

关键词： Seizure detection EEG signals machine learning deep learning feature selection

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：