International Journal of Computational Science and Engineering (29 papers in press)

Regular Issues

• Self-supervised learning with split batch repetition strategy for long-tail recognition  
  by Zhangze Liao, Liyan Ma, Xiangfeng Luo, Shaorong Xie 
  Abstract: Deep neural networks cannot be well applied to balance testing when the training data present a long tail distribution. Existing works improve the performance of the model in long tail recognition by changing the model training strategy, data expansion, and model structure optimisation. However, they tend to use supervised approaches when training the model representations, which makes the model difficult to learn the features of the tail classes. In this paper, we use self-supervised representation learning (SSRL) to enhance the model's representations and design a three-branch network to merge SSRL with decoupled learning. Each branch adopts different learning goals to enable the model to learn balanced image features in the long-tail data. In addition, we propose a Split Batch Repetition strategy for long-tailed datasets to improve the model. Our experiments on the Imbalance CIFAR-10, Imbalance CIFAR-100, and ImageNet-LT datasets outperform existing similar methods. The ablation experiments prove that our method performs better on more imbalanced datasets. All experiments demonstrate the effectiveness of incorporating the self-supervised representation learning model and split batch repetition strategy.
  Keywords: long-tail recognition; self-supervised learning; decoupled learning; image classification; deep learning; neural network; computer vision;.
  
  
• Alliance: a makespan delay and bill payment cost saving-based resource orchestration framework for the blockchain and hybrid cloud enhanced 6G serverless computing network  
  by Mahfuzul H. Chowdhury 
  Abstract: Serverless computing technology with the function-as-a-service and backend-as-a-service platforms provides on-demand service, high elasticity, automatic scaling, service provider-based server/operating system management, and no idle capacity charges facilities to the users. Owing to limited resources, the traditional research articles on public/private cloud-based serverless computing cannot meet the user's IoT application requirements. The current works are limited only to a single job type and objectives. There is a lack of appropriate resource orchestration schemes for low latency, bill payment, and energy-cost-based serverless computing job execution networks with hybrid cloud and blockchain. To overpower these issues, this paper instigates a low latency, energy-cost, and bill payment-based multiple types of job scheduling, resource orchestration, network, and mathematical model for the blockchain and hybrid cloud-enhanced serverless computing network. The experimental results delineated that up to 70% makespan delay and 30.23% bill payment gain are acquired in the proposed alliance scheme over the baseline scheme.
  Keywords: serverless computing; job scheduling; worker selection; resource orchestration; blockchain; cloud computing; 6G; job execution time; user bill payment.
  DOI: 10.1504/IJCSE.2023.10060388
   
  
• Finite element analysis of the liver subjected to non-invasive indirect mechanical loading  
  by Samar Shaabeth, Amina Kadhem, Hassanain Ali Lafta 
  Abstract: Injuries from non-invasive abdominal blunt trauma represent 75% of no visible bleeding traumas. As a step for additional faster and more reliable diagnostic tool, simulation was performed of karate kick, punch, and pinpoint loading on the front, back, right, and left sides of the abdomen for 31-and 50-years old females and a 45-year-old male 2D segmented computed tomography images. The organs densities and mechanical properties were applied. Mechanical analysis was performed during 0.5 s, 1 s, 1.5 s and 2 s loading time. The results, compatible with previous literature, indicated the affected regions of liver, spleen, pancreas, kidney, and colon by the trauma.
  Keywords: liver; blunt abdominal trauma; BAT; simulation; finite element; indirect loading; non-invasive.
  DOI: 10.1504/IJCSE.2024.10063606
   
  
• Secure forensic image analysis by optimised iterative model with random consensus approaches  
  by S.B. Gurumurthy, Ajit Danti 
  Abstract: Future measurements, software, and scalability testing related to cloud performance are required for forensic image scalability (FIS) optimisations and advancements. An advanced iterative reconstruction model and consensus mechanism must be used to quantitatively evaluate image quality in any blockchain framework since this will have a direct impact on the security and usability of the framework. This work addresses these problems by presenting a fast and efficient forgery detection system based on optimal security, feature extraction, and pre-processing. This will render conventional media security and forensic techniques meaningless. In this work, a random sample consensus (RSC) method is proposed for the analysis of FIS. To ensure that the architecture is as strong and secure as possible, the iterative reconstruction model (IRM) is employed. Initially, one may consider channel processing to be a form of database picture pre-processing. One perspective state that the enhanced chicken swarm optimisation (ECSO) algorithm is used to advance the scaling settings to balance invisibility and power. This RSC’s threshold setting reduces the number of excluded matches as well as the root mean square error (RMSE). Enhancement of scalability as well as picture reconstruction demonstrate the utility of the proposed technology. The simulation findings on multiple retinal image datasets demonstrate that the proposed method further enhances accuracy matching by 10.56% and rate of progress by 30% on average compared with the RSC-IRM strategy.
  Keywords: image reconstruction; scalability; optimisation; image security; forensic image.
  DOI: 10.1504/IJCSE.2024.10064093
   
  
• Unlocking the potential of deepfake generation and detection with a hybrid approach  
  by Shourya Chambial, Tanisha Pandey, Rishabh Budhia, Balakrushna Tripathy, Anurag Tripathy 
  Abstract: With the use of numerous software programs and cutting-edge AI technologies, a lot of phony movies and photos are produced today, although the modification is rarely obvious The videos can be exploited in a variety of unethical ways to frighten, fight, or threaten individuals People should be careful these days to avoid using such techniques to produce phony videos Deep Fake is the name of an AI-based method for creating synthetic versions of human photographs This research paper proposes a hybrid model of Inception ResNetV2 and Xception for the purpose of deepfake detection With the rise of deepfake technology, detecting altered images and videos has become a crucial task for ensuring authenticity and preventing misinformation The hybrid model developed here is using a dataset of fake and real videos and the images and achieved a classification accuracy of over 96 75%.
  Keywords: fake video; convolutional neural network; CNN; RNN.
  DOI: 10.1504/IJCSE.2024.10064094
   
  
• Privacy-preserving SQL queries on cross-organisation databases  
  by Ye Han, Xiaojie Guo, Tong Li, Xiaotao Liu 
  Abstract: In recent years, much industrial interest has been paid to SQL queries on a joint database contributed by several mutually distributed companies or organizations. However, privacy regulations and commercial interest prevent these entities to trivially share their local databases with each other. To enable such SQL queries still, some privacy-preserving technologies should be applied. In this work, we outline a provably secure MPC framework of privacy-preserving SQL queries for industrial applications, such as medical research in hospitals, financial oversight, business cooperation, etc. In particular, this framework is secure against any semi-honest adversary, which is a popular threat model in real-life systems. This framework also models a common efficiency optimization of SQL query plans at the cost of mild leakage.
  Keywords: SQL queries; secure multi-party computation; privacy.
  DOI: 10.1504/IJCSE.2024.10064247
   
  
• A transfer learning approach for adverse drug reactions detection in bio-medical domain based on knowledge graph  
  by Monika Yadav, Prachi Ahlawat, Vijendra Singh 
  Abstract: Among the top causes of mortality, adverse reactions of drugs (ADRs) are dominant. This imposes severe health risks and a significant financial burden on patients. Consequently, timely prediction of possible ADRs of a drug has become an essential concern in the clinical domain. However, it is challenging to recognise the adverse reactions of all drugs using existing ADR data sources. Recently, a semantic-rich knowledge base and machine learning techniques have shown high accuracy in predicting ADRs in advance. This paper introduces a new framework, knowledge graph slot-filling clinical bi-directional encoder representations from transformers (KG-SF Clinical BERT), which takes triples of knowledge graph as text sequences. It applies transformer-based multi-task learning with slot-filling for ADR classification and fine-tuned on bio-medical domain to detect ADRs. The KG-SF clinical BERT brings remarkable performance gain with AUC of 0.88 on drug bank and SIDER datasets and 0.99 AUC on PubMed dataset.
  Keywords: adverse drug reactions; clinical BERT; knowledge graph; KG; transfer learning.
  DOI: 10.1504/IJCSE.2024.10064638
   
  
• Modified glowworm swarm optimisation-based cluster head selection and enhanced energy-efficient clustering protocol for IoT-WSN  
  by T. Kanimozhi, S. Belina V.J. Sara 
  Abstract: The maintenance cost of flat-based wireless sensor network-internet of things (WSN-IoT) is high. Clustering is recommended to reduce message overhead, manage congestion, and simplify topology repairs. A clustering protocol enhances energy efficiency, prolongs network lifespan by grouping nodes into clusters, and reduces transmission distance to the base station (BS). Depending on parameters like quality of service (QoS), energy consumption, and network load, a clustering technique organises nodes into clusters. Each cluster is led by one or more cluster heads (CH) that collect and transmit data to the BS directly or through intermediary nodes. To enhance WSN-based IoT longevity, this study presents an enhanced energy-efficient clustering protocol (EEECP). It establishes an optimal number of clusters, utilises the modified fuzzy C means (MFCM) algorithm to stabilise and reduce sensor node energy consumption, and introduces the modified glowworm swarm optimisation (MGSO) algorithm for CH selection. MGSO incorporates a dynamic threshold for balanced CH longevity within clusters. Performance is evaluated using metrics including first node dies (FND), last node dies (LND), half node dies (HND), weighted first node dies (WFND), energy usage, and network lifetime compared to existing protocols.
  Keywords: modified glowworm swarm optimisation; modified fuzzy C-means algorithm; cluster head; quality of service; enhanced energy-efficient clustering protocol.
  DOI: 10.1504/IJCSE.2024.10064839
   
  
• Optimisation of quantum circuits using cost effective quantum gates  
  by Swathi Mummadi, Bhawana Rudra 
  Abstract: The importance of reversible operations has increased with the emergence of new technologies. Reversible operations are crucial for developing energy-efficient and cost-efficient circuits. The efficiency of a quantum circuit is measured in terms of quantum cost and quantum depth. In this paper, we propose an optimisation algorithm for reversible gates such as the Peres gate, Toffoli gate, and the entanglement purification method. Peres and Toffoli gates play an important role in quantum circuit implementation, and entanglement purification plays a key role in various applications such as quantum teleportation, secure communication, and quantum key distribution. The proposed algorithm optimises the quantum cost and quantum depth to 20% compared with the existing approaches.
  Keywords: reversible computation; quantum computation; quantum csx and sxdg gates; entanglement purification; reversible logic gates.
  DOI: 10.1504/IJCSE.2024.10065245
   
  
• A precise deep learning-based ECG arrhythmia classification scheme using deep bidirectional capsule network classifier  
  by Soumen Ghosh, Satish Chander 
  Abstract: This paper presents a novel approach for accurate ECG signal classification using a deep bidirectional capsule network classifier. ECGs are vital for diagnosing cardiac arrhythmias, and precise classification is crucial for automated heart disease prediction. Initially, ECG signal data is acquired and artifacts are removed using a quantised wavelet threshold method, followed by spectrogram analysis. Features are extracted using the VGG spectral net algorithm, and arrhythmia classification is performed with the deep bidirectional capsule network approach. The study evaluates this technique using the MIT-BIH arrhythmia database, identifying five types of arrhythmias: NOR, RBBB, PVC, LBBB, and APB. The results demonstrate improved accuracy compared to traditional methods, suggesting this deep learning approach could enhance diagnostic capabilities, streamline healthcare workflows, and improve patient outcomes.
  Keywords: electrocardiogram; ECG; deep bidirectional capsule network; DBCN; VGG spectral net; quantised wavelet threshold method; ECG arrhythmias.
  DOI: 10.1504/IJCSE.2024.10065882
   
  
• Deep learning architectures for detection of acute myeloid leukemia  
  by Niteesh K. Ramesh, Vijayakumar Kadappa, Rajeshwari Devi D. Veerabhaskar, Divijendranatha Reddy Sirigiri, Pooja T. Sreedhar 
  Abstract: Acute Myeloid Leukaemia (AML) is a malignancy associated with the rapid development of immature blood cells which fail to perform normal immune functions. AML accounts for almost 80% of all blood-related cancers in adults. Despite the availability of various therapeutic options, the prognosis is poor. Early detection and appropriate diagnosis of AML as well as many other cancers remain unmet clinical necessities that need to be addressed. Artificial intelligence-based approaches find utility across all spectrum of sciences including health care and diagnosis. However, most of the existing work uses traditional machine learning algorithms and smaller datasets, resulting in poor generalization ability of the models which leads to lower detection rates. To address these concerns, we propose two deep learning architectures to detect AML by using publicly available larger image datasets of various blood cells. In contrast to other models, the proposed deep networks employ kernels with diversified shapes for identifying complex patterns. These models require lower parameter count (up to the extent of 97%) and computational requirements, yet they demonstrate improved or competitive performance compared with other deep networks.
  Keywords: artificial intelligence; deep learning; artificial neural networks; acute myeloid leukemia; healthcare.
  DOI: 10.1504/IJCSE.2024.10065883
   
  
• Image-text multimodal sentiment analysis method integrating multi-themes and multi-labels  
  by Shunxiang Zhang, LongHui Hu, ShuYu Li, Wenjie Duan, Xiaolong Wang 
  Abstract: The current image-text sentiment analysis models only focus on the content of text and image, and ignore the synergistic effect of themes and labels information on the semantic features of image and text. Therefore, we propose a multi-modal sentiment analysis method integrating multi-themes and multi-labels. Firstly, the global features and local features of the image are obtained by CNN and Faster-RCNN. Bi-LSTM is used to obtain word-level features and sentence-level features of the text, and the Bert is responsible for extracting the theme-label features. Then, the attention network for feature interaction to generate the word-local correlation features, the text's sentence features are combined with the image's global features to generate the joint features of the image-sentence. Finally, these two features are fused with the theme-label features to obtain the results of the sentiment analysis. The experimental results demonstrate that the proposed method can improve the accuracy of image-text sentiment analysis.
  Keywords: multimodal sentiment analysis; multi-theme labels; modal fusion; target detection; attention network.
  DOI: 10.1504/IJCSE.2024.10065952
   
  
• MFA: Web API recommendation based on service multiple feature aggregation  
  by Guobing Zou, Chunhua Zeng, Yue Zhu, Pengtao Li, Song Yang, Shengxiang Hu 
  Abstract: As the number of Web services continues to increase, it has become a challenging problem to provide developers with accurate and efficient Web services that meet Mashup requirements. To solve this problem, various methods have been proposed to recommend APIs to match the needs of mashups, and have achieved great success. However, due to the uneven quality of service descriptions, there are some challenges in feature extraction, utilization of service meta-information, and textual requirements understanding. Therefore, we propose a Web API recommendation method (FMA) based on service multiple feature aggregation. FMA uses the attention mechanism model to mine the semantic features of similar services and enhance the features of Mashup services and Web API services. We conduct extensive experiments on a large-scale real-world dataset called ProgrammableWeb, and the results show that FMA outperforms existing baseline methods on multiple evaluation metrics.
  Keywords: Web API; Mashup service; API recommendation; multiple feature aggregation; attention mechanism.
  DOI: 10.1504/IJCSE.2024.10065996
   
  
• Agreement window algorithm for user controlled and utility supported personal data privacy  
  by Geocey Shejy, Pallavi Chavan 
  Abstract: A good privacy-preserving algorithm must keep the trade-off between privacy preservation and the utility of the data for analysis purposes. As the spectrum of data privacy regulations expands and evolves for fine tuned protection of individual data privacy; both customers and organizations who are the stakeholders of digital services need more supporting research to ensure legal compliance.All existing privacy regulations mention the rights of data users and the responsibilities of data collectors. The work in this article keeps the European Union’s GDPR as a key reference, still, the requirement of the consent of the data owner to release personal data is common in all personal data privacy acts across the globe.The Agreement Window Algorithm (AW Algorithm), helps organizations to do legitimate personal data collection, by ensuring user consent. The agreement window is a conceptual space where the data owner and the service-providing organizations agree to share the data. While sharing
  Keywords: agreement window; differential privacy; set sensitivity; sensitivity factor; personal identifiable information; PII; personal sensitive information; PSI; quasi identifier; QI.
  DOI: 10.1504/IJCSE.2024.10065997
   
  
• Solving overhead transmission line engineering problems with elastic catenary equations  
  by Hervé Ducloux 
  Abstract: Most methods used to solve overhead line sag-tension problems assume that the conductor is inextensible. This assumption leads to the classical catenary curve. In this article, the conductor is assumed to be extensible, which leads to the more accurate elastic catenary curve. First, an original mathematical approach of establishing the equations of this curve is proposed: it uses the slope to connect the abscissa and the ordinate of each point on the curve. Then, it is used to solve sag-tension problems for perfectly linear elastic conductor or composite conductor with nonlinear inelastic behaviour. In each case, all steps of the algorithms proposed are explained. The numerical examples included in this paper show that the number of iterations needed to solve a sag-tension problem is quite small. Comparisons with previous studies are also made to assess the accuracy of the proposed method.
  Keywords: equations of elastic catenary curves; overhead transmission lines; sag-tension problems.
  DOI: 10.1504/IJCSE.2024.10066037
   
  
• A green pattern-based data encryption solution in the cloud  
  by Farah Abdmeziem, Saida Boukhedouma, Mourad Chabane Oussalah 
  Abstract: In today's business landscape, there's a growing interest in Cloud computing, particularly for extensive data storage. However, a significant security concern arises from entrusting data to cloud providers and relinquishing control to customers. To address this, data encryption on the customer's side is a viable solution, though it can be resource-intensive, especially for large data volumes. This can lead to performance issues and environmental impacts, including increased server carbon footprints. In this work, we propose a solution based on customized encryption/decryption patterns, categorizing data into three sensitivity levels and considering access/update frequency. We also assess environmental implications and cost metrics to highlight the positive impact of our approach compared to other state-of-the-art methods. Experimental results demonstrate that our approach not only facilitates encryption adoption but also effectively balances data confidentiality with practical resource and energy constraints.
  Keywords: data; security; cloud computing; green encryption; encryption-decryption patterns; cost metrics.
  DOI: 10.1504/IJCSE.2024.10066044
   
  
• Design of exercise recommendation model based on clustering collaborative filtering adaptability  
  by Chaoyang Shi, Zhen Zhang 
  Abstract: The study explores the issue of insufficient personalised recommendation ability of exercise systems in online teaching. It combines clustering analysis and collaborative filtering algorithms. K-means clustering is used as the basis for clustering analysis. And the collaborative filtering algorithm is optimised from three aspects: the number of learners working together, the difference in exercise scores, and the difficulty of exercise. A clustering collaborative filtering adaptive exercise recommendation model based on similarity improvement is proposed. The study evaluates the application effectiveness of the model through simulation experiments. The experimental results show that the MAE values formed by the designed algorithm under changes of the nearest neighbours are the lowest among the comparison algorithms, proving its superiority. In the comparison of indicators, the accuracy, recall, and F1 values of the algorithm are all the highest among the comparison algorithms, further verifying its effectiveness. Stability analysis shows that in both sets, the accuracy of the research design model reaches above 0.87, indicating that the model has high stability and accuracy. From this, the model designed in the study has advantages in recommendation effectiveness, which can help students improve learning effectiveness and provides a new approach for learning assistance systems.
  Keywords: clustering; collaborative filtering; online learning; exercise; recommendation; similarity.
  DOI: 10.1504/IJCSE.2024.10066045
   
  
• Scene text detection using robust masks and cascaded classifiers  
  by Houssem Turki, Mohamed Elleuch, Monji Kherallah, Alima Damak 
  Abstract: Detecting text in scenes poses a significant and challenging problem due to the complex character shapes and the diverse degradations present in natural images. It represents the initial and crucial phase that must be successfully completed before the text recognition stage. In this study, we suggest a hybrid approach to tackle this issue, leveraging the maximally stable extremal regions (MSER) algorithm, which gained significant attention in recent research. Despite its popularity, it remains very sensitive to the shape, size, scale and background noise of text characters. To tackle its limits and refine the final result, we focuses on an extended MSER based method. Overall flowchart of the suggested system is divided into three steps: 1) robust masks generation to identify the text candidate regions; 2) feature extraction and selection based on VGG16 deep learning architecture; 3) employing classifiers in a cascaded structure to differentiate between text and non-text areas based on enhanced geometrical pattern characteristics. The effectiveness of the proposed method is demonstrated through an experimental study conducted on various benchmarks, such as ICDAR2013, ICDAR2015, MSRA-TD500, and RRC-MLT.
  Keywords: scene text; text detection; maximally stable extremal regions; MSER; VGG16 deep learning model; random forest; SVM; NCA-based feature selection.
  DOI: 10.1504/IJCSE.2024.10066146
   
  
• Constructing stock portfolio with transformer  
  by Li Jinyuan, Linkai Luo 
  Abstract: Machine learning methods have been applied to quantitative investing, yet the application of Transformer models remains limited. Stock prices are influenced by both long-term and short-term features. Existing methods usually treat the influencing factors as a whole and do not distinguish them. In this paper, we introduce a Transformer encoder-decoder architecture tailored for the capture of long-term and short-term features. By partitioning historical data into long-term and short-term parts, the encoder module concentrates on extracting long-term features, while the decoder concentrates on short-term features and the integration of long and short-term features. Portfolios are then constructed from the top N predicted stocks. Experimental results show that the proposed Transformer model outperforms the existing state-of-the-art methods, LSTM, RNN, and GRU models, with improvements of 26%, 19%, and 14% in annualised returns for long-short portfolio combinations, respectively. It indicates the benefits of extracting long-term and short-term features separately.
  Keywords: stock portfolio; transformer; factor model.
  DOI: 10.1504/IJCSE.2024.10066373
   
  
• Multilingual language classification model for offensive comments categorisation in social media using HAMMC tree search with enhanced optimisation technique  
  by B. Aarthi, Balika J. Chelliah 
  Abstract: The exponential rise of social media platforms has led to a surge in offensive content, highlighting the necessity for effectively detecting and managing such comments. This necessitates precise and advanced online social networks (OSN) categorisation and optimisation methods. This study introduces and assesses a novel technique for automatically categorising texts, supporting over 60 languages, without relying on a pre-annotated data set. The technique employs multilingual methods based on the randomised explicit semantic analysis (ESA) strategy. To combat the inherently multilingual nature of social media content, the paper introduces an innovative classification and optimisation strategy named hybrid adaptive Markov chain Monte Carlo tree search (HAMCMTS) with enhanced eagle Aquila optimiser (EEAO). The study uses three publicly available datasets to identify negative or offensive comments in various languages, offering a comprehensive analysis in this field. The proposed approach holds potential for diverse applications, particularly in multilingual categorisation tasks such as monitoring disaster-related communications on social media to improve visibility and trust. Moreover, it incorporates a sophisticated mechanism to bolster the dependability of its recommendations.
  Keywords: negative or offensive comments; multilingual languages; explicit semantic analysis; ESA; enhanced eagle Aquila optimiser; EEAO.
  DOI: 10.1504/IJCSE.2024.10066586
   
  
• A verifiable and secure DNN classification model over encrypted data  
  by Weixun Li, Guanghui Sun, Yajun Wang, Long Yuan, Minghui Gao, Yan Dong, Chen Wang 
  Abstract: Outsourcing deep neural networks (DNNs) offers relief for client overhead but sparks concerns over sensitive data privacy. Current methods, like homomorphic encryption and federated learning, aim to safeguard privacy but often falter in preserving it and verifying gradient integrity. In response, we introduce DNNCM-ED, a novel verifiable and secure DNN classification model operating on encrypted data. Our approach establishes a jointly trained DNN model between client and server, with the client encrypting local gradients via double masking before aggregation at the server. We also devise secure communication protocols for fundamental operations crucial in constructing the DNN classification model. Additionally, we craft verifiable algorithms tailored to DNNCM-ED, ensuring privacy of local gradients and integrity verification. Extensive property and performance analyses underscore DNNCM-ED's superiority in accuracy, time efficiency, and communication overhead. Through these advancements, DNNCM-ED addresses critical shortcomings in existing privacy-preserving outsourcing methods while providing robust confidentiality and integrity verification.
  Keywords: deep neural networks; encrypted data; homomorphic encryption.
  DOI: 10.1504/IJCSE.2024.10066642
   
  
• Enhancing multi-view ensemble learning with zig-zag pattern-based feature set partitioning  
  by Aditya Kumar, Jainath Yadav 
  Abstract: This study suggests a novel approach called Zig-Zag Pattern-Based Feature Set Partitioning. The method involves two steps: first, calculating feature correlations using Pearson's coefficient, and second, ranking features based on mean correlation and arranging them in a zig-zag pattern. The zig-zag pattern ensures diverse and balanced feature subsets, improving model generalization and reducing over- fitting. Experimental results on ten high-dimensional datasets show the practical significance of the suggested strategy, which show that it outperforms previous strategies in accuracy and generalization. This approach advances multi-view ensemble learning, offering a practical solution for improving ensemble model performance in complex data analysis tasks.
  Keywords: feature set partitioning; views construction; ensemble learning; zig-zag partitioning; classification; multi-view ensemble learning.
  DOI: 10.1504/IJCSE.2024.10066742
   
  
• EIUAPA: an efficient and imperceptible universal adversarial attack on audio classification models  
  by Huifeng Li, Pengzhou Jia, Weixun Li, Bin Ma, Bo Li, Dexin Wu, Haoran Li 
  Abstract: The domain of Audio Classification Models is emerging as a significant paradigm, albeit susceptible to universal adversarial attacks. These attacks involve the insertion of a single optimal perturbation into all audio samples, leading to incorrect predictions. Nonetheless, existing attack methodologies are hindered by inefficiencies and imperceptibility challenges. In order to streamline the attack process effectively, we propose a two-step strategy EIUAPA that offers an optimal initiation point for the perturbation optimization process, resulting in a notable decrease in generation time. To maintain imperceptibility, we present a range of metrics focusing on perturbation concealment, serving as benchmarks for optimization. These metrics ensure that perturbations are concealed not only in the frequency and time domains but also remain statistically indistinguishable. Experimental results demonstrate that our method generates UAPs 87.5% and 86.8% faster than baseline methods, with improved Signalto-Noise Ratio (SNR) and Attack Success Rate (ASR) scores.
  Keywords: adversarial attack; artificial intelligence; security and privacy; audio classification; deep learning.
  DOI: 10.1504/IJCSE.2024.10066803
   
  
• Using generative adversarial network for music transformation  
  by Cheng-Han Wu, Yu-Cheng Lin, Pimpa Cheewaprakobkit, Wan-Chin Ting, Timothy K. Shih 
  Abstract: In this study, we propose a generative adversarial network (GAN) framework for music style transfer. Initially, a dataset of traditional Jiangnan songs is pre-processed into two categories: complete compositions and corresponding musical phrases (starting and ending notes), which are then converted into piano-roll images. The CycleGAN model is then used to train these images until the model converges to establish a music style transfer model. The goal is to allow users to input only the starting and ending notes of each measure as a musical phrase, and the system will convert this phrase into complete musical compositions in the Jiangnan style. Then we use a deep learning framework and music expertise for data processing, enhancing the quality and utility of our conversions. At the same time, we have established music style assessment metrics based on the statistical data of the dataset, providing an effective method for evaluating music styles.
  Keywords: music transformation; generative adversarial network; GAN; automatic music generation; music style transfer.
  DOI: 10.1504/IJCSE.2024.10066865
   
  
• Scalable malicious URL detection technique for smishing attacks  
  by Razvan Stoleriu, Catalin Negru, Bogdan-Costel Mocanu, Emil-Andrei Constantinescu, Alexandra-Elena Mocanu, Florin Pop 
  Abstract: Nowadays, smartphones are used daily and use sensitive information making people more vulnerable to cyber-security attacks. The easiest way for attackers to access a smartphone is through SMS phishing (smishing) using URL shortening services. In this paper, we propose a scalable technique to detect malicious URLs in smishing attacks based on a Cloud-Edge architecture, using threat intelligence platforms (e.g., VirusTotal, PhishTank), and Machine Learning algorithms that classify the URLs based on their features. We used a public dataset for training and proposed new features to improve it. We evaluated our proposed ML model against JRip, PART, J48, and Random Forest algorithms. Our model has improved performance compared to similar solutions, obtaining an accuracy of approximately 97%. To showcase the effectiveness of our solution, we implement an Android application that detects malicious short URLs in SMS messages and notifies the user concerning their legitimacy.
  Keywords: smishing attacks; malicious URLs; edge-cloud computing; threat intelligence; machine learning.
  DOI: 10.1504/IJCSE.2024.10067011
   
  
• Performance analysis and comparison of jellyfish optimisation-based maximum power point tracking controller for partial shading condition  
  by Dilip Yadav, Nidhi Singh 
  Abstract: This paper addresses the critical challenge of partial shading conditions (PSC) in photovoltaic systems, which significantly affect the efficiency of PV panels. Conventional methods often fail to optimise output under partial shading condition, prompting the need for innovative approaches. The study proposes the jelly-fish optimisation algorithm for maximum power point tracking, comparing its effectiveness with various existing MPPT controllers including incremental conductance, modified incremental conductance, perturbance and observation, particle swarm optimisation, cuckoo search algorithm, grey wolf optimisation, and whale search optimisation techniques. The study reveals the limitations of conventional techniques in optimising power output under PSC. The findings highlight the superiority of the jellyfish-based MPPT, achieving an impressive efficiency of 99.89% with a minimal tracking time of 0.14 seconds, surpassing other MPPT controllers. This work advances the field by highlighting the jellyfish algorithms effectiveness and guiding future research toward more efficient MPPT methods.
  Keywords: cuckoo search; jellyfish optimisation algorithm; maximum power point tracking; partial shading condition; particle swarm optimisation.
  DOI: 10.1504/IJCSE.2024.10067209
   
  
• Hybrid predictive modelling for insurance premium retention: integrating statistical and AI techniques  
  by Ahmed A. Khalil, Zaiming Liu 
  Abstract: This research highlights the critical role of forecasting in the insurance industry and emphasises the premium retention ratio (PRR) as a key internal performance indicator for evaluating insurance company operations. Traditional time series models like ARIMA and Exponential Smoothing face limitations in capturing complex data patterns. To address this, the study proposes a hybrid predictive model that combines statistical time series models (ARIMA, EXP) with advanced AI techniques (ANN, SVR) to enhance PRR prediction accuracy in Egypts Fire, Marine, and Aviation insurance sectors. Using 80% of data for training (19892015) and 20% for testing (20162021), the study demonstrates that hybrid models, particularly ARIMA-ANN and EXP-ANN, outperform conventional models. The findings suggest that incorporating ANN into these models significantly improves prediction accuracy. This research offers a novel approach to forecasting in the Egyptian insurance market and provides publicly accessible datasets for further comparative studies across different countries.
  Keywords: artificial neural network; ANN; artificial intelligence; autoregressive integrated moving average; ARIMA; exponential smooth; Egyptian insurance market; statistical time series; support vector machine; SVM; insurance.
  DOI: 10.1504/IJCSE.2024.10067258
   
  
• GraphBiGRU model for anti-money laundering based on preference-based reinforcement learning via the label filtering loop mechanism  
  by Meng Li, Xinqiao Su, Lu Jia, Rongbo You 
  Abstract: Anti-Money Laundering (AML) in Bitcoin transactions remains challenging since Bitcoin data has a complex graph structure and sequential nature, with many unknown labels and an imbalanced distribution of licit and illicit transactions. To address these challenging issues, we propose a novel reinforcement learning-based GraphBiGRU model via the label filtering loop mechanism to detect illicit transactions in the Bitcoin blockchain. Specifically, we first constructed the GraphBiGRU network to learn the graph structure and temporal information of Bitcoin data. Then, we introduced the label filtering loop mechanism, which encouraged the GraphBiGRU to select reliable pseudo-labeled samples that reduced data noise interference. In addition, we investigated a preference-based reinforcement learning strategy that enabled the GraphBiGRU to better identify illicit transactions, thereby improving performance on imbalanced datasets. Finally, we conducted experiments on the Elliptic dataset, demonstrating that our method achieved state-of-the-art performance, especially with a limited labelled dataset.
  Keywords: anti-money laundering; illicit transactions; GraphBiGRU; label filtering loop mechanism; pseudo-labeled samples; preference-based reinforcement learning; elliptic dataset.
  DOI: 10.1504/IJCSE.2024.10068029
   
  
• Efficient traffic management in the internet of vehicles through enhanced routing and deep learning  
  by Arundhati Sahoo, Asis Kumar Tripathy 
  Abstract: In the Internet of Vehicles (IoV), vehicles are treated as sophisticated smart devices with robust communication systems. IoV uses cellular technology and internet access for vehicle-to-vehicle communication. However, traditional routing algorithms struggle with rapid vehicle movements and varying road conditions, leading to instability and inefficiency, especially in congested traffic. This study proposes a unique approach called the Improved Greedy-Bi directional Long Short-Term Memory (I-GBiLSTM) predictor, which integrates an Improved Greedy Perimeter Stateless Routing Algorithm to enhance link stability within 5G networks by incorporating real-time data on vehicle movements and road conditions and traffic patterns. Additionally, a BiLSTM neural network has been enhanced by incorporating a 1-dimensional Convolutional Autoencoder (1D-CNNAE) and a Temporal Transformer Encoder (TTE) for monitoring and predicting traffic data, enabling unique feature extraction. Experimental results demonstrate that I-GBiLSTM is superior to the other existing protocols, achieving a 99% delivery ratio, 100 routing overhead, 180 ms end-to-end delay, and 98.2% prediction accuracy.
  Keywords: traffic management; internet of vehicles; IoV; routing; deep learning; network traffic prediction.
  DOI: 10.1504/IJCSE.2024.10068349