2026. 1^st Issue

Volume XVIII, Number 1

Table of contents 

Full issue  

 

PAPERS FROM OPEN CALL

Rajakani M., Beaulah Jeyavathana R., and Kavitha R. J.
Relay Pursuit-Vathana: A Novel Optimization Approach for Feature Selection in Software Defect Prediction 
Software defect prediction plays a crucial role in ensuring the quality and reliability of software systems. Feature selection, the process of identifying the most relevant features from a large set of potential features which is essential for building effective defect prediction models. In this paper, we propose a novel feature selection model based on the RelayPursuit- Vathana (RP-Vathana) optimization algorithm, inspired by relay races and pursuit dynamics in biological systems. The proposed model aims to identify an optimal subset of features for software defect prediction, maximizing the predictive performance of the resulting classification model. The RP-Vathana algorithm was integrated with a Naïve Bayes classifier and benchmarked on three datasets (PC5, JM1, KC2) to validate its effectiveness in feature selection for defect prediction. The results show that RP-Vathana significantly outperforms existing wrapper-based methods, obtaining mean accuracies of 94.28%, 93.69%, and 96.35% on PC5, JM1, and KC2, respectively, compared to the 83−90% range of rival techniques. While the parameter-free design improves usability, the algorithm's performance on highly noisy or very small datasets warrants future investigation into hybrid extensions for enhanced robustness.

DOI: 10.36244/ICJ.2026.1.1
Download 

 

Oussama Lagnfdi, Marouane Myyara, and Anouar Darif
A New Deep Learning-Based Approach for IoT Task Offloading in Multi-access Edge Computing 
The exponential growth of Internet of Things (IoT) devices and the growing demand for resource-intensive applications have introduced significant challenges in computation, storage, and network efficiency. Although cloud computing provides partial relief, its centralized nature leads to unacceptable latency for delay-sensitive applications. Multi-access Edge Computing (MEC), especially with the advent of 5G, has emerged as a compelling solution by relocating computation closer to data sources, thereby reducing latency and improving responsiveness in applications such as smart agriculture, autonomous vehicles, augmented reality, and telemedicine. However, efficient workload offloading in MEC environments remains complex due to system heterogeneity, varying application requirements, and limited edge resources. This paper proposes a novel neural network-based approach to computation offloading in MEC, integrating workload allocation and resource management while accounting for application delay sensitivity, processing capacity, and communication constraints. The proposed model enables driving offloading decisions, adapting to fluctuating system states without relying on complex mathematical formulations. Simulation results demonstrate that the approach significantly reduces service time and enhances resource utilization, ensuring responsiveness for modern IoT applications. This research underscores MEC’s potential to meet the rising computational and latency demands of next-generation IoT infrastructure.

DOI: 10.36244/ICJ.2026.1.2
Download 

 

 

Hussein Tuama, and Sándor Imre
Enhancing Quantum State Transmission Fidelity through Quantum Orthogonal Frequency Division Multiple Access 
In this paper, we propose quantum orthogonal frequency division multiple access (Q-OFDMA), a novel quantum communication scheme designed to overcome the fidelity limitations imposed by noise in multi-user quantum networks. Inspired by its classical counterpart, Q-OFDMA employs the quantum Fourier transform (QFT) and its inverse (IQFT) to encode and decode information across quantum channels. We evaluate our model under both a depolarization channel and a generalized noise model that interpolates between depolarizing and phase-damping noises. The simulation results conducted on Qiskit platform demonstrate that Q-OFDMA outperforms the reference model, achieving superior average fidelity across varying qubit counts and noise levels.

DOI: 10.36244/ICJ.2026.1.3
Download 

 

 

Márton Pál Lipcsey-Magyar, Attila Ármin Madarász, and Adrian Pekar
Beyond JA4+: Flow Statistics vs. TLS Fingerprinting for Encrypted Malware Detection 
The deployment of Encrypted Client Hello (ECH) challenges TLS fingerprinting, a widely used approach for encrypted malware detection, by encrypting the handshake fields these methods rely on. This paper presents a systematic evaluation of flow-based statistical features as a handshakeindependent alternative to fingerprinting. Through validation against the official JA4+ implementation, we establish limitations in fingerprinting approaches for this corpus: only 64.9% of malware families possess unique signatures, placing an inherent ceiling on achievable recall in our evaluation. We evaluate flow-level features—packet counts, timing patterns, and size distributions—across 27 experimental configurations on a dataset of 16,542 flows spanning 101 families (59 malware and 42 benign applications). Random Forest classifiers using combined flow statistics and sequential packet length features achieve 98.11% F1-score for binary malware detection with 97.22% recall, substantially exceeding fingerprinting’s theoretical recall bound of 64.9%. For fine-grained family identification, we obtain 54.81% macro F1 across 101 classes and 48.71% macro F1 for malwareonly attribution, demonstrating that flow-based methods retain meaningful discriminative power where fingerprinting abstains. Across all tasks, Random Forest consistently outperforms neural networks and k-NN, with performance gaps widening in complex multiclass scenarios. These findings highlight flow-based classification as a practical and reproducible approach that can help maintain network security visibility as ECH deployment progresses, showing that behavioral traffic patterns are expected to provide durable signals for detection even as handshake fields become encrypted.

DOI: 10.36244/ICJ.2026.1.4
Download 

 

 

Nour Ammar and László T. Kóczy
Fuzzy Linguistic Signatures  
Fuzzy Linguistic Signatures (FLS) extend the concept of Fuzzy Signatures (FSigs) by introducing linguistic variables as qualitative descriptors within a hierarchical fuzzy structure. Although fuzzy signatures have been successfully applied in various domains, their reliance on numerical membership degrees limits their ability to model subjective or linguistically defined information. This paper establishes a formal mathematical frame-work for FLS by defining a family of fuzzy linguistic signatures equipped with suitable linguistic aggregation operators and a partial ordering relation among linguistic values. Furthermore, meet-and-join operators are introduced to demonstrate that FLS satisfies the properties of a lattice as an algebraic structure. Consequently, fuzzy linguistic signatures provide an expressive representational framework capable of handling qualitative, human-like reasoning. 

DOI: 10.36244/ICJ.2026.1.5
Download 

 

 

Doaa Mostafa, Sally S. Ismail, and Mostafa Aref 
A Hybrid Syntactic–Statistical–Semantic Framework for Detecting AI-Generated Text Across Domains 
Recent advances in large language models (LLMs) have enabled highly human-like text generation, raising concerns related to misinformation, authorship verification, and academic integrity. Current approaches for detecting LLMgenerated text suffer from several limitations, including limited robustness to linguistic diversity, sensitivity to text length variations and paraphrasing, weak domain generalization, and high computational cost. To address these challenges, this paper proposes a hybrid framework for detecting LLM-generated text that integrates syntactic and statistical features with deep semantic representations learned using GloVe embeddings, Convolutional Neural Networks (CNNs), and Bidirectional Long Short-Term Memory (BiLSTM) networks. By combining linguistic cues with contextual semantics, the proposed model captures both structural and semantic patterns to distinguish human-written text from LLM-generated content. Experiments conducted on the ChatGPT Research Abstracts and ElectAI datasets demonstrate strong cross-domain generalization and robustness to text length variations and paraphrasing. The proposed framework achieves an accuracy of 98.63%, an F1-score of 98.66%, and a minimum false positive rate (FPR) of 0.01. These results indicate the effectiveness, stability, and reliability of the framework for detecting LLM- generated text.

DOI: 10.36244/ICJ.2026.1.6
Download 

 

 

György Wersényi and Victor Koech 
Content Credentials: Trust Issues, Technical Solutions and Future Perspectives Using Encrypted Metadata in Image Processing 
Emerging technologies offer validation and authentication solutions in the field of audiovisual content creation. Visible or invisible watermarking, embedded metadata, and digital signatures can be used to maintain the validity and creditability of still images and video data. The Coalition for Content Provenance and Authenticity (C2PA) was established to create an open source framework and to provide technical solutions for image capture, processing, delivery, and verification. The leading market players in hardware and software development set the goal of applying encrypted metadata information to guarantee the authenticity of the data. Currently, only a few devices and applications are available and have been implemented based on this technology. This paper gives an introductory overview of the recent state, highlighting advantages, drawbacks, available implementations, and future perspectives on research directions.

DOI: 10.36244/ICJ.2026.1.7
Download 

 

 

Sang-Quang Nguyen, and Chi-Bao Le
Rate-Splitting Multiple Access for Satellite Short-Packet Communications: Finite Blocklength Modeling and Reliability Analysis 
Short-packet transmission is becoming crucial for satellite services that cannot rely on long codewords to hit the required latency and reliability. This study investigates ratesplitting multiple access (RSMA) in that context and builds a finite-blocklength (FBL) model for a downlink satellite-terrestrial link affected by Shadowed-Rician fading. We obtain closed-form approximations for the block error rate (BLER) of both the common and private streams, explicitly incorporating imperfect successive interference cancellation (ipSIC) at the receivers. Compared with power-domain non-orthogonal multiple access (NOMA), RSMA exhibits more stable BLER the common stream helps dampen residual interference due to ipSIC and the short-packet effect–so RSMA generally needs less transmit power to attain the same error targets. Numerical results validate the analysis and demonstrate consistent RSMA advantages across a wide range of transmit powers, blocklengths, shadowing severities, and antenna configurations. The results suggest that RSMA is a very promising option for future satellite systems that need to provide reliable, low-latency, and short-packet communications in view of realistic SIC imperfections.

DOI: 10.36244/ICJ.2026.1.8
Download 

 

 

John Baghous, and Mohamed Khaled Chahine
Evaluating Data Transmission Performance in 5G mmWave Networks using Multi-Layer Transmission and MIMO Technology 
The transition from 4G to 5G networks was necessitated by fundamental limitations in spectral efficiency and data capacity inherent to the existing framework. Fifthgeneration (5G) systems address these constraints by capitalizing on key enabling technologies, such as mmWave spectrum, multiple-input multiple- output (MIMO), massive MIMO (mMIMO), beamforming (BF) or Precoding. This paper investigates a multi-layer transmission scheme employing MIMO with Precoding (SVD) as a cooperative technique to enhance downlink (DL) data transmission performance in an enhanced mobile broadband (eMBB) scenario. The study operates within the standard 5G mmWave frequency band (FR2 at 40 GHz). We differentiate between key performance metrics: the user-experienced data rate (or throughput) Measured at the Receiver (Rx) and the peak theoretical data rate (or Bit Rate) Measured at the Transmitter (Tx). Simulation results, conducted using MATLAB, demonstrate that the proposed approach significantly improves both the achievable throughput and spectral efficiency within a fixed bandwidth. Throughput is evaluated in absolute terms (Mbps) and as a normalized percentage of the peak theoretical data rate (Bit Rate). The core of this study examines the impact of the number of spatial data streamng layers on a 5G-NR system performance. While increasing the transmission layers enhances the potential peak data rate at the transmitter, it concurrently elevates the bit error rate (BER) at the receiver, ultimately degrading the net throughput. This underscores the necessity for advanced receiver-side technologies, such as MIMO processing, to counteract high-path loss and other impairments prevalent at mmWave frequencies. The results confirm that augmenting the number of antennas in the MIMO configuration effectively mitigates this limitation. It improves the overall throughput and reduces the received BER by enhancing spatial diversity and signal recovery capabilities.

DOI: 10.36244/ICJ.2026.1.9
Download 

 

 

Angel D, and Dr. Robin Rohit Vincent
A Review of Security Challenges and Intrusion Detection Mechanisms to Mitigate Sub-Optimization Attacks in RPL-Based 6LoWPAN IoT Networks 
The Internet of Things (IoT) has transformed device connectivity with the smooth interfacing for realtime data exchange across multiple applications, from smart homes to industrial automation. Nonetheless, as networks under IoT, especially those using the routing protocol for low-power and lossy networks (RPL), continue in their expansion, the security penetration becomes much more evident. One of the major security constraints is suboptimization attacks-they negatively affect network performance, scalability, and data integrity. These attacks impede the very efficiency of the IoT systems, thereby making it so challenging for the systems to be secured and maintained successfully. Traditional IDS and cryptographic solutions are seldom fit-for-purpose in dynamic IoT environments, which opens up the need for the ability to provide scalable and energy-aware security solutions. This review investigates and surveys existing IDS, cryptographic solutions, and machine learning techniques targeting and working against such threats. It puts forth an integrated solution where an adaptive IDS is combined with scalable, energy-efficient, real-time anomaly detection to make IoT networks more resilient to sub-optimization attacks. According to this study, dynamic, contextaware safety measures are essential, as they are capeble of adressing the new challenges arising from IoT environments.

DOI: 10.36244/ICJ.2026.1.10
Download 

 

 

Aphilak Lonklang, and János Botzheim
Path Planning Transformer Supervised by Improved RRT* with Reduced Random Map Size for Mobile Robots 
The Improved Rapidly-exploring Random Tree with Reduced Random Map Size (IRRT*-RRMS) algorithm was previ- ously developed to find collision-free paths for mobile robot path planning. Given the excellent performance of Transformer Neural Networks with sequential data, we propose an encoder-decoder transformer model combining a Vision Transformer (ViT) as the encoder and a time-series forecasting module as the decoder to learn the path planning algorithm. The novelty of this paper lies in developing a model supervised by a dataset generated from the IRRT*-RRMS algorithm and using this trained model for the path planning task. The trained model efficiently predicts intermediate points between the desired starting and goal points. The performance was validated on a real robot, demonstrating that the trained model required less computation time compared to the IRRT*-RRMS algorithm.

DOI: 10.36244/ICJ.2026.1.11
Download