Computer Sciences & Mathematics Forum

Editorial

Jump to: Other

3 pages, 139 KB

Open AccessEditorial

Preface to the 1st International Conference on Emerging Tech & Innovation (ICETI)

by Dina M. Ibrahim and Jamal Alotaibi

Comput. Sci. Math. Forum 2026, 13(1), 1; https://doi.org/10.3390/cmsf2026013001 - 13 Apr 2026

Viewed by 311

Abstract

n/a Full article

(This article belongs to the Proceedings of The 1st International Conference on Emerging Tech & Innovation (ICETI))

Other

Jump to: Editorial

13 pages, 1961 KB

Open AccessProceeding Paper

Blockchain-Based Secure Data Sharing in Cybersecurity: A Framework for Protecting Sensitive Information

by Raneem Khaled AlFadhel and Mohammad Ali A. Hammoudeh

Comput. Sci. Math. Forum 2026, 13(1), 2; https://doi.org/10.3390/cmsf2026013002 (registering DOI) - 15 Apr 2026

Viewed by 539

Abstract

With the growing volume of sensitive data stored and processed in cloud environments, conventional security models are no longer sufficient to guarantee privacy, integrity, and trust. This paper proposes a blockchain-based framework that integrates Zero-Knowledge Proofs (ZKPs) and homomorphic encryption (HE) to enable [...] Read more.

With the growing volume of sensitive data stored and processed in cloud environments, conventional security models are no longer sufficient to guarantee privacy, integrity, and trust. This paper proposes a blockchain-based framework that integrates Zero-Knowledge Proofs (ZKPs) and homomorphic encryption (HE) to enable secure and privacy-preserving data sharing. ZKPs are employed to verify user access rights without exposing identities or underlying information, while HE allows computations to be performed directly on encrypted data, ensuring confidentiality is preserved throughout the data lifecycle. The proposed framework addresses the limitations of existing approaches that either lack encrypted computation capabilities or expose sensitive data during processing. Formal and informal analyses demonstrate the feasibility of the model in terms of encryption time, ZKP verification latency, and computation overhead. The framework is designed to be applied initially in the healthcare sector and aligns with national digital transformation initiatives such as Saudi Vision 2030. Full article

(This article belongs to the Proceedings of The 1st International Conference on Emerging Tech & Innovation (ICETI))

► Show Figures

Figure 1

16 pages, 454 KB

Open AccessProceeding Paper

Data Encryption Algorithms for Cloud Storage Systems—A Comparative Analysis

by Abdulsalam Ibrahim Almirdasi and Mohamed Tahar Ben Othman

Comput. Sci. Math. Forum 2026, 13(1), 3; https://doi.org/10.3390/cmsf2026013003 - 15 Apr 2026

Viewed by 479

Abstract

Cloud storage systems require strong and efficient encryption methods to ensure data security and reliability. However, selecting the most suitable encryption algorithm remains a challenge due to variations in performance, overhead, and reliability. This study aims to introduce a comparative analysis of five [...] Read more.

Cloud storage systems require strong and efficient encryption methods to ensure data security and reliability. However, selecting the most suitable encryption algorithm remains a challenge due to variations in performance, overhead, and reliability. This study aims to introduce a comparative analysis of five encryption algorithms—Advanced Encryption Standard (AES), Blowfish, Rivest-Shamir-Adleman (RSA), Elliptic Curve Cryptography (ECC), and Advanced Encryption Standard one-time password AES-OTP with RSA hybrid model (AES-OTP with RSA)—to identify the most suitable algorithm to protect sensitive data in cloud storage systems. The evaluation of these algorithms was based on encryption/decryption time, data size overhead, encryption/decryption throughput, performance metrics (accuracy, precision, recall, and F1-score), and error metrics mean square error and mean absolute error (MSE and MAE), using datasets of various sizes. The results indicated that AES provided the fastest encryption and decryption time, minimal overhead, and the highest throughput and accuracy, while Blowfish also performed efficiently but with slightly higher error rates. RSA and ECC, although secure, were slower and demonstrated more overhead. The hybrid AES-OTP with RSA model achieved a good balance between speed and secure key management. This study highlights the trade-offs between speed, security, and storage efficiency, offering guidance in selecting appropriate encryption algorithms for cloud-based data protection. Full article

(This article belongs to the Proceedings of The 1st International Conference on Emerging Tech & Innovation (ICETI))

► Show Figures

Figure 1

17 pages, 834 KB

Open AccessProceeding Paper

Deep Learning Approaches for Efficient and Accurate DNA Sequence Alignment Using Large Language Models

by Shefa Alkhowaiter and Mohamed Tahar Ben Othman

Comput. Sci. Math. Forum 2026, 13(1), 4; https://doi.org/10.3390/cmsf2026013004 - 15 Apr 2026

Viewed by 535

Abstract

This study addresses the challenge of DNA sequence similarity analysis by combining deep learning with DNABERT embeddings. Traditional alignment methods based on direct pairwise comparisons often fail to detect deeper biological relationships beyond nucleotide matching. However, DNABERT, a large transformer-based language model, captures [...] Read more.

This study addresses the challenge of DNA sequence similarity analysis by combining deep learning with DNABERT embeddings. Traditional alignment methods based on direct pairwise comparisons often fail to detect deeper biological relationships beyond nucleotide matching. However, DNABERT, a large transformer-based language model, captures contextual and functional patterns within genomic data. We initially used a dataset of 20 human DNA sequences and later expanded it to 70 sequences to enhance statistical reliability. The results showed that DNABERT recovered functional similarities even between sequences with low identity percentages, revealing previously overlooked structural relationships that were hidden by traditional alignments. Quantitative evaluation using precision, recall, and F1 score confirmed the robustness and consistency of the DNABERT-based approach. Overall, this study demonstrates that combining traditional and deep learning-based methods yields a more accurate and interpretable framework for DNA sequence alignment, thereby paving the way for enhanced genomic analysis. Full article

(This article belongs to the Proceedings of The 1st International Conference on Emerging Tech & Innovation (ICETI))

► Show Figures

Figure 1

14 pages, 730 KB

Open AccessProceeding Paper

Lightweight and Transparent Intrusion Detection in the Internet of Medical Things: The Role of Explainable AI

by Rawan Abdulaziz AlRumaih, Tarek Moulahi and Dina M. Ibrahim

Comput. Sci. Math. Forum 2026, 13(1), 5; https://doi.org/10.3390/cmsf2026013005 - 16 Apr 2026

Viewed by 490

Abstract

The rise of the Internet of Medical Things (IoMT) has transformed healthcare through real-time monitoring and improved outcomes but also introduced critical security and privacy challenges. This paper presents a focused survey of Explainable AI (XAI) approaches for intrusion detection in IoMT, emphasizing [...] Read more.

The rise of the Internet of Medical Things (IoMT) has transformed healthcare through real-time monitoring and improved outcomes but also introduced critical security and privacy challenges. This paper presents a focused survey of Explainable AI (XAI) approaches for intrusion detection in IoMT, emphasizing methods that are lightweight, transparent, and deployable under resource constraints. We first clarify XAI terminology and taxonomy (global vs. local scope; ante hoc vs. post hoc; model-agnostic vs. model-specific) and then systematize recent works from the past five years across cybersecurity sub-domains relevant to eHealth. Representative pipelines span classical ML (e.g., LR, RF, SVM, and XGBoost) and deep models (e.g., DNNs and SRU/LSTM), with post hoc explainers, especially SHAP and LIME, dominating practice on benchmark datasets such as CICIDS2017, NSL-KDD, ToN-IoT, WUSTL-EHMS, and CICIoMT2024. Our comparative analysis highlights consistent gains from model ensembling and interpretable feature selection while uncovering key gaps: limited real-world validation, inconsistent explainability metrics, adversarial brittleness, and the computing cost of explanations at the edge. Full article

(This article belongs to the Proceedings of The 1st International Conference on Emerging Tech & Innovation (ICETI))

► Show Figures

Figure 1

14 pages, 2445 KB

Open AccessProceeding Paper

Encephalon_DC: Classification of Brain Diseases Using Deep Learning Techniques

by Leidi M. Saleh Aouto, Lin M. Saleh Aouto, Rawan Khaled Flifel and Dina M. Ibrahim

Comput. Sci. Math. Forum 2026, 13(1), 6; https://doi.org/10.3390/cmsf2026013006 - 16 Apr 2026

Viewed by 286

Abstract

The brain is the most complex organ in the human body, and neurological disorders pose significant diagnostic challenges. This study focuses on three prevalent conditions—Alzheimer’s disease, brain tumors, and Parkinson’s disease—collectively referred to as Encephalon Diseases. We propose a three-level deep learning-based framework, [...] Read more.

The brain is the most complex organ in the human body, and neurological disorders pose significant diagnostic challenges. This study focuses on three prevalent conditions—Alzheimer’s disease, brain tumors, and Parkinson’s disease—collectively referred to as Encephalon Diseases. We propose a three-level deep learning-based framework, termed the Encephalon Diseases Classifier, for automated diagnosis from magnetic resonance imaging (MRI) scans. In Level 1, MRI images are classified as normal or diseased. Level 2 further categorizes diseased cases into one of the three targeted conditions. Level 3 performs stage or subtype classification for Alzheimer’s disease and brain tumors. The framework employs four convolutional neural network (CNN) architectures, namely ResNet152-V2, EfficientNet-B0, DenseNet121, and VGG16, trained on a preprocessed dataset. Experimental results show that ResNet152-V2 achieves the highest accuracy of 100%, while EfficientNet-B0 and DenseNet121 yield comparable performance across all levels. The proposed method demonstrates the potential of multi-level deep learning strategies for precise and scalable Encephalon disease classification. Full article

(This article belongs to the Proceedings of The 1st International Conference on Emerging Tech & Innovation (ICETI))

► Show Figures

Figure 1

8 pages, 594 KB

Open AccessProceeding Paper

Energy-Aware Bid-Based Client Selection for Federated Learning in Resource-Constrained IoT Networks

by Rana Albelaihi

Comput. Sci. Math. Forum 2026, 13(1), 7; https://doi.org/10.3390/cmsf2026013007 - 17 Apr 2026

Viewed by 505

Abstract

Federated learning (FL) enables distributed IoT devices to train machine learning models collaboratively without sharing raw data. However, energy heterogeneity among devices significantly challenges efficient and equitable participation, particularly in resource-constrained networks. This paper introduces BEAF (Bid-based Energy-Aware Federated Learning), a client selection [...] Read more.

Federated learning (FL) enables distributed IoT devices to train machine learning models collaboratively without sharing raw data. However, energy heterogeneity among devices significantly challenges efficient and equitable participation, particularly in resource-constrained networks. This paper introduces BEAF (Bid-based Energy-Aware Federated Learning), a client selection strategy that incorporates the availability of energy and the training utility of the device into a unified selection criterion. Each client independently computes a bid score based on its remaining energy and the relative improvement in local training loss. Clients with the highest utility-per-joule scores are selected to participate in each round. The approach operates without centralized profiling or historical coordination and is compatible with synchronous FL protocols. The evaluation of standard benchmarks shows that BEAF enhances the precision of the global model, reduces total energy consumption, and improves fairness in client participation compared to baseline methods, such as random sampling and selection based on energy thresholds. The method is suitable for deployment in energy-limited environments, including agricultural monitoring and other distributed sensing applications. Full article

(This article belongs to the Proceedings of The 1st International Conference on Emerging Tech & Innovation (ICETI))

► Show Figures

Figure 1

9 pages, 1079 KB

Open AccessProceeding Paper

Spectral Analysis of Neural Network Weight Matrices and the Impact of Weight Conditioning on Optimization Performance

by Abdulnaser Rashid

Comput. Sci. Math. Forum 2026, 13(1), 8; https://doi.org/10.3390/cmsf2026013008 - 16 Apr 2026

Viewed by 371

Abstract

This paper explores the relationship between random matrix theory (RMT) and the use of weight conditioning for training deep neural networks by employing an integrated framework. It has been shown that trained neural networks produce singular value distributions that follow universal distributions prescribed [...] Read more.

This paper explores the relationship between random matrix theory (RMT) and the use of weight conditioning for training deep neural networks by employing an integrated framework. It has been shown that trained neural networks produce singular value distributions that follow universal distributions prescribed by RMT; however, the presence of non-universal outliers in the distribution can contain significant information particular to the task being performed. In addition, this research investigates how the application of diagonal row equilibration as a form of conditioning affects spectral behavior and optimization stability within deep neural networks. The results show that through conditioning, the random bulk of the singular value decomposition (SVD) spectrum is effectively compressed into a narrow band about the value 1, significantly reducing the Marchenko–Pastur bounds. The results also support the claim that weight conditioning retains the informative nature of the spectral outliers. The experimental results show that weight condition numbers (κ(W)) decreased from extremely ill-conditioned regimes of approximately 10³ to 10⁴ to almost 1.0, producing smoother training landscapes, a quicker convergence rate, and an improved ability for gradients to propagate. These results suggest that conditioning weights can be thought of as an implicit spectral regularize linking RMT evidence and concepts to the practical optimization of deep learning methods. Full article

(This article belongs to the Proceedings of The 1st International Conference on Emerging Tech & Innovation (ICETI))

► Show Figures

Figure 1

12 pages, 619 KB

Open AccessProceeding Paper

DSGCNN-DA: A Deep Stacked Graph Convolutional Neural Network with Dynamic Aggregation for Malware Behavioral Learning

by Ghida Almusned, Lama Almutairi, Emna Benmohamed and Rana Albelaihi

Comput. Sci. Math. Forum 2026, 13(1), 9; https://doi.org/10.3390/cmsf2026013009 - 15 Apr 2026

Viewed by 184

Abstract

Malware remains a major threat to computer systems, posing serious risks to security and privacy by stealing sensitive data, disrupting services, and compromising system integrity. Traditional detection methods are often ineffective against rapidly evolving malware. In response, data-driven deep learning has emerged as [...] Read more.

Malware remains a major threat to computer systems, posing serious risks to security and privacy by stealing sensitive data, disrupting services, and compromising system integrity. Traditional detection methods are often ineffective against rapidly evolving malware. In response, data-driven deep learning has emerged as a powerful alternative. Recent models have demonstrated promising performance in detecting malicious behavior by learning from these behavioral traces. Behavior-based detection represents a significant advancement in the fight against malware. This paper introduces a deep stacked Graph Convolutional Network (GCN) for effective malware behavioral analysis. The aggregation of multiple GCN layers and blocks results in dynamically performed Jumping Knowledge (JK) method, especially Long Short-Term Memory (LSTM). LSTM-based JK dynamically selects and weights the most informative GCN layers for each node to improve the model’s ability. Experimental results demonstrate the superior performance of our deep stacked Graph Convolutional Network with Dynamic Aggregation (DSGCN-DA) model, achieving an accuracy of 98.93% on the API-Call-Sequences dataset, outperforming the state-of-the-art approaches. Full article

(This article belongs to the Proceedings of The 1st International Conference on Emerging Tech & Innovation (ICETI))

► Show Figures

Figure 1

18 pages, 1062 KB

Open AccessProceeding Paper

Heterogeneous Federated Learning Model for Recognizing Human Activity

by Nwadher S. Alblihed and Dina M. Ibrahim

Comput. Sci. Math. Forum 2026, 13(1), 10; https://doi.org/10.3390/cmsf2026013010 - 17 Apr 2026

Viewed by 154

Abstract

A range of sensors are used by human activity recognition (HAR) to identify the activities that people complete each day. The recognition of human activities has benefited greatly from machine learning (ML), as it has made many human activities more easily recorded. Unfortunately, [...] Read more.

A range of sensors are used by human activity recognition (HAR) to identify the activities that people complete each day. The recognition of human activities has benefited greatly from machine learning (ML), as it has made many human activities more easily recorded. Unfortunately, a centralized approach is used in many HAR applications, which might compromise user privacy. One must use deep learning (DL) using different algorithms and models to analyze the data generated from ML. Another kind of ML is distributed ML, called federated learning (FL), which tries to distribute ML models across edge devices. Thus, this study presents an FL model to support HAR by building a generic model and using user-based training data without data sharing. Through developing heterogeneous local models, each client takes the most suitable DL model to the client. This study uses three different DL models to develop the local model: Convolutional Neural Network (CNN), Residual Network (ResNet), and Long Short-term Memory (LSTM). Moreover, different numbers of clients are experimented with: two, five, and ten clients. The UniMiB SHAR dataset is used to apply the experiments. As a result, using five clients with three mixed DL models gives the highest Accuracy of 90.8%. Full article

(This article belongs to the Proceedings of The 1st International Conference on Emerging Tech & Innovation (ICETI))

► Show Figures

Figure 1

8 pages, 361 KB

Open AccessProceeding Paper

A Computational Model for Animal Language Processing: Translating Canine and Feline Behavior into Human-Readable Communication

by Deepa Sonal, Md Alimul Haque, Sultan Ahmad, Sultan Alqahtani and A. E. M. Eljialy

Comput. Sci. Math. Forum 2026, 13(1), 11; https://doi.org/10.3390/cmsf2026013011 - 17 Apr 2026

Viewed by 363

Abstract

Humans have always been curious about what animals are trying to communicate, especially our closest companions—dogs and cats. While we often rely on instinct and observation to understand their needs and feelings, this method can be inaccurate or limited. This research introduces a [...] Read more.

Humans have always been curious about what animals are trying to communicate, especially our closest companions—dogs and cats. While we often rely on instinct and observation to understand their needs and feelings, this method can be inaccurate or limited. This research introduces a new computational model designed to translate the behaviors of dogs and cats into simple, human-readable messages. By combining data from their body language, sounds, facial expressions, and movements, the model uses advanced machine learning and deep learning techniques to identify what the animal might be feeling or trying to express. We collect and analyze real-world behavioral data from pets, then train the system to interpret signals like barking, meowing, tail movements, or posture changes. The final output could be a sentence or voice alert that helps pet owners understand things like “I’m hungry,” “I’m scared,” or “I want to play.” This approach not only improves how we care for pets but also enhances emotional connection and communication between humans and animals. It opens new doors for technology in pet care, training, and veterinary support. Full article

(This article belongs to the Proceedings of The 1st International Conference on Emerging Tech & Innovation (ICETI))

► Show Figures

Figure 1

14 pages, 1608 KB

Open AccessProceeding Paper

Explainable Intrusion Detection System Using Prototypical Network and Recursive Feature Elimination

by Wessam F. Abouzaid, Ebrahim A. Ramadan and Nermeen G. Rezk

Comput. Sci. Math. Forum 2026, 13(1), 12; https://doi.org/10.3390/cmsf2026013012 - 22 Apr 2026

Viewed by 217

Abstract

This study explores the use of traditional machine learning and deep learning algorithms to develop efficient Intrusion Detection Systems (IDSs). It evaluates data using the NSL-KDD dataset, which contains both normal and attack traffic. The research compares the performance of various classifiers, including [...] Read more.

This study explores the use of traditional machine learning and deep learning algorithms to develop efficient Intrusion Detection Systems (IDSs). It evaluates data using the NSL-KDD dataset, which contains both normal and attack traffic. The research compares the performance of various classifiers, including Random Forest, Extreme Gradient Boosting, LightGBM, and Prototypical Networks. Recursive Feature Elimination is used for feature selection to enhance decision-making and model performance. The models are assessed using multiple metrics, such as accuracy, precision, recall, F-score, ROC curves, and confusion matrices. In addition, Explainable AI techniques like SHAP and LIME are employed to interpret predictions, making the IDS more transparent and reliable. Results indicate that few-shot learning models, particularly Prototypical Networks, combined with Recursive Feature Elimination techniques, outperform traditional models, achieving up to 98% accuracy. This approach enhances IDS applications in IoT by enabling more accurate threat detection, improving decision-making, and identifying key intrusion parameters. Full article

(This article belongs to the Proceedings of The 1st International Conference on Emerging Tech & Innovation (ICETI))

► Show Figures

Figure 1

13 pages, 1654 KB

Open AccessProceeding Paper

Multifractal Analysis in Healthcare: A Review of Techniques, Applications, and Future Perspectives

by Ahlem Aziz and Necmi Serkan Tezel

Comput. Sci. Math. Forum 2026, 13(1), 13; https://doi.org/10.3390/cmsf2026013013 (registering DOI) - 22 Apr 2026

Viewed by 286

Abstract

Complex biological and medical systems often exhibit irregular and self-similar structures that can be effectively analyzed using fractal and multifractal frameworks. This study aims to provide a comprehensive overview of multifractal analysis as a mathematical tool for characterizing complex biomedical patterns and improving [...] Read more.

Complex biological and medical systems often exhibit irregular and self-similar structures that can be effectively analyzed using fractal and multifractal frameworks. This study aims to provide a comprehensive overview of multifractal analysis as a mathematical tool for characterizing complex biomedical patterns and improving disease diagnosis. The methods discussed include the Wavelet Transform Modulus Maxima (WTMM) and box-counting techniques, which quantify local scaling behaviors and heterogeneity within medical images. A review of recent studies demonstrates that multifractal parameters have successfully differentiated between normal and pathological tissues in diseases such as cancer, cardiac disorders, and Alzheimer’s disease. This paper also examines the integration of artificial intelligence, particularly machine learning algorithms, with multifractal features to enhance diagnostic accuracy and automate image interpretation. The results indicate that this hybrid approach improves the reliability and sensitivity of early disease detection. In conclusion, multifractal analysis, when systematically applied and combined with AI, offers a promising complementary framework for advancing precision medicine and supporting clinical decision-making. Full article

(This article belongs to the Proceedings of The 1st International Conference on Emerging Tech & Innovation (ICETI))

► Show Figures

Figure 1

6 pages, 176 KB

Open AccessProceeding Paper

Can You Trust Your Copilot? A Privacy Scorecard for AI Coding Assistants

by Amir Al-Maamari

Comput. Sci. Math. Forum 2026, 13(1), 14; https://doi.org/10.3390/cmsf2026013014 - 25 May 2026

Viewed by 151

Abstract

The rapid integration of AI-powered coding assistants into developer workflows has raised significant privacy and trust concerns. As developers entrust proprietary code to services like OpenAI’s GPT, Google’s Gemini, and GitHub Copilot, the unclear data handling practices of these tools create security and [...] Read more.

The rapid integration of AI-powered coding assistants into developer workflows has raised significant privacy and trust concerns. As developers entrust proprietary code to services like OpenAI’s GPT, Google’s Gemini, and GitHub Copilot, the unclear data handling practices of these tools create security and compliance risks. This paper addresses this challenge by introducing and applying a novel, expert-validated privacy scorecard. The methodology involves a detailed analysis of four document types—from legal policies to external audits—to score five leading assistants against 14 weighted criteria. A legal expert and a data protection officer refined these criteria and their weighting. The results reveal a distinct hierarchy of privacy protections, with a 20-point gap between the highest- and lowest-ranked tools. The analysis uncovers common industry weaknesses, including the pervasive use of opt-out consent for model training and a near-universal failure to filter secrets from user prompts proactively. The resulting scorecard provides actionable guidance for developers and organizations, enabling evidence-based tool selection. This work establishes a new benchmark for transparency and advocates for a shift towards more user-centric privacy standards in the AI industry. Full article

(This article belongs to the Proceedings of The 1st International Conference on Emerging Tech & Innovation (ICETI))

► Show Figures

Figure 1

Journal Menu

Journal Browser

Comput. Sci. Math. Forum, 2026, ICETI 2025

The 1st International Conference on Emerging Tech & Innovation (ICETI)

Editorial

Other

Further Information

Guidelines

MDPI Initiatives

Follow MDPI