International Journal of Computer Science, Algorithms and Programming Languages

Adversarial Attacks and Defenses in Machine Learning-Based Intrusion Detection Systems

2026-06-03T11:45:41+00:00

Machine learning–based intrusion detection systems (IDS) have become essential for identifying sophisticated cyber threats in modern network environments. By learning patterns of normal and malicious behaviour, these systems can detect previously unknown attacks and adapt to evolving threat landscapes. However, their reliance on data-driven models makes them vulnerable to adversarial manipulation. Adversarial attacks exploit weaknesses in learning algorithms by crafting inputs that evade detection, corrupt training data, or manipulate model behaviour. Such attacks, including evasion, poisoning, and backdoor insertion, can significantly degrade detection accuracy and compromise system reliability. This study examines the nature of adversarial threats targeting machine learning–based IDS, analyses their operational mechanisms, and evaluates their impact on detection performance. It further explores defensive strategies such as adversarial training, robust feature engineering, ensemble modelling, secure data pipelines, and adaptive monitoring frameworks. The paper highlights the trade-offs between robustness, computational cost, and detection efficiency, and identifies key challenges in developing resilient IDS capable of operating in dynamic and adversarial environments. By synthesizing current research and emerging defense paradigms, this work provides a comprehensive foundation for designing secure and trustworthy machine learning–driven intrusion detection systems.

SocietySync ERP - A Normalized Relational Database Architecture for Residential Management

2026-02-27T08:35:09+00:00

Many residential apartments from today are still using paper/pencil systems, instead of an automated system for recording new tenants, maintaining financial records, tracking customer complaints, and communicating with tenants. As expected, utilizing these systems creates both inefficiencies and inaccuracies within the systems. SocietySyncERP was designed and developed to assist in eliminating these inefficiencies and inaccuracies with database automation. SocietySyncERP uses a PostgreSQL database that consists of 11 interrelated tables and three end-user types: Administrators, Property Owners, and Tenants. The database design for SocietySyncERP has been normalized to the third normal form (3NF), supports the requirements for ACID Transactions, has included the use of bcrypt for password hashing, and has implemented measures to detect and prevent SQL injection attacks to date. Currently, the SocietySyncERP database has approximately twenty-five features, such as billing management, complaint tracking, visitor logs, polling, etc. The SocietySyncERP database has been tested by 44 users and has demonstrated acceptable performance on queries and data integrity, with query times averaging between 2 and 20 milliseconds. SocietySyncERP adheres to a standard three-tier architectural design to provide the system with the ability to scale to meet future needs. The database theories of relationship algebra, normalization, and complexity theory form the foundation for SocietySyncERP databases and represent a direct relationship between the conceptual theories of database design taught in the classroom and the actual needs of contemporary apartment communities.

Artificial Intelligence Techniques for CO₂ Emission Prediction and Sustainable Optimization in Geopolymer Concrete: A Comprehensive Review

2026-06-09T05:24:56+00:00

The construction industry is one of the largest contributors to global carbon dioxide (CO₂) emissions because of its dependence on Ordinary Portland Cement (OPC). Cement production contributes nearly 7–8% of global anthropogenic CO₂ emissions due to limestone calcination, fossil fuel combustion, and high-temperature clinker manufacturing. Geopolymer concrete has emerged as a sustainable alternative because it uses aluminosilicate-rich industrial and agricultural by-products such as fly ash, ground granulated blast furnace slag, metakaolin, rice husk ash, silica fume, and sugarcane bagasse ash. These materials reduce reliance on OPC while supporting waste reutilization and circular economy principles. Recently, Artificial Intelligence (AI) techniques have been widely applied to geopolymer concrete research for predicting compressive strength, optimizing mix design, estimating CO₂ emissions, improving lifecycle assessment, and supporting sustainable construction decision-making. This review paper discusses AI-based methods including Artificial Neural Networks, Support Vector Machines, Random Forest, Gradient Boosting, XGBoost, Deep Learning, and optimization algorithms for geopolymer concrete. The paper also reviews AI-based CO₂ emission prediction, evaluation metrics, lifecycle assessment integration, optimization approaches, limitations, and future research directions. The findings indicate that AI-assisted geopolymer concrete systems can reduce experimental effort, improve prediction accuracy, minimize environmental impact, and accelerate the development of low-carbon construction materials.

Design of an AI-Driven Personalized English Learning System Using Machine Learning Algorithms

2026-05-28T08:46:13+00:00

This research explores the design and development of an AI-Driven Personalized English Learning System (AIP-ELS) that applies Artificial Intelligence (AI) and Machine Learning (ML) techniques to improve the effectiveness of English language education. Conventional English learning methods generally provide the same instructional materials and activities to all learners, regardless of their individual abilities, learning speed, interests, or performance levels. As a result, many learners experience difficulties in maintaining engagement and achieving consistent progress. To overcome these limitations, the proposed system introduces a personalized learning environment that adapts educational content and learning strategies according to each learner’s needs. The system combines Natural Language Processing (NLP), learner behavior analysis, and intelligent recommendation algorithms to monitor learner performance and provide customized lessons, exercises, assessments, and feedback. By analyzing data such as vocabulary usage, grammar accuracy, response time, and speaking performance, the platform continuously updates individualized learning pathways to support efficient skill development. The adaptive framework also encourages active participation and motivation by delivering content that matches the learner’s proficiency level and learning preferences. An experimental evaluation was conducted to compare the proposed system with traditional e-learning approaches. The results show that learners using the AI-driven platform achieved noticeable improvements in vocabulary acquisition, grammatical accuracy, speaking proficiency, and overall engagement. In addition, users reported higher satisfaction due to the personalized learning experience and real-time feedback mechanisms. The study demonstrates that AI-based personalization can significantly enhance English language learning and contribute to the development of scalable, intelligent and data-driven educational systems for diverse learning communities.