Article Per Year (5 Year)
p-Index From 2021 - 2026
0.961
P-Index
This Author published in this journals
All Journal Journal of Education and Learning (EduLearn) Aktual : Jurnal Pengabdian Kepada Masyarakat International Journal of Advances in Artificial Intelligence and Machine Learning Journal of Science, Technology, and Social Innovation
Mageswaran Sanmugam
Universiti Sains Malaysia
Humanities Computer Science & IT Education Engineering Environmental Science Library & Information Science Public Health Social Sciences Other

Published : 6 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 6 Documents
Search

 1 
Knowledge Distillation for Enhancing Interpretability and Efficiency in Complex Machine Learning Models Jaesik Jeong; Kit Ling Chan; Mageswaran Sanmugam
International Journal of Advances in Artificial Intelligence and Machine Learning Vol. 3 No. 1 (2026): International Journal of Advances in Artificial Intelligence and Machine Learni
Publisher : CV Media Inti Teknologi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.58723/ijaaiml.v3i1.649

Abstract

Background: Complex machine learning (ML) systems often require substantial computational resources, making them difficult to deploy in real-world environments constrained by hardware limitations, interpretability requirements, and regulatory standards. While knowledge distillation (KD) has traditionally been viewed as a model compression technique, its broader implications for efficiency, interpretability, and regulatory compliance remain underexplored.Aims: This study aims to reconceptualize knowledge distillation beyond model compression by framing it as a dual strategy for efficiency and interpretability enhancement. The paper proposes a structured distillation protocol that integrates predictive performance assessment, computational profiling, and feature attribution alignment within a unified experimental design.Methods: The proposed distillation protocol employs a temperature-scaled objective function combining supervised cross-entropy loss and Kullback Leibler divergence to facilitate relational knowledge transfer from teacher to student models. Experiments were conducted across multiple benchmark datasets. Evaluation consisted of three components: (1) predictive performance measurement, (2) computational efficiency profiling including parameter counts and inference latency, and (3) interpretability analysis using feature attribution similarity and perturbation stability metrics. Statistical analyses were performed to assess performance differences.Result: Across benchmark datasets, distilled student models achieved teacher-level accuracy ranging between 95% and 98%. Parameter counts and inference latency were reduced by more than 60%. Interpretability analyses showed improved explanation consistency, smoother decision structures, and higher feature attribution alignment. Statistical testing confirmed that efficiency and interpretability gains were obtained without significant performance degradation.Conclusion: The findings support the reconceptualization of knowledge distillation as a dual optimization strategy that enhances both operational efficiency and interpretability while preserving predictive strength. Rather than serving solely as a compression mechanism, KD functions as a scalable and adaptive framework for deployment-ready AI systems that balance performance, computational constraints, and explanation stability.
Human-AI Collaboration in Scientific Writing Training: A Quantitative Evaluation of Learning Effectiveness and Academic Integrity Eko Risdianto; Nanik Setyowati; Mohammad Qaiz Rezvani; M. Abdul Jamal; Mageswaran Sanmugam; M. Esad Kuloğlu
Aktual: Jurnal Pengabdian Kepada Masyarakat Vol. 4 No. 2 (2026): Aktual: Jurnal Pengabdian Kepada Masyarakat
Publisher : CV Media Inti Teknologi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.58723/aktual.v4i2.728

Abstract

Background: The rapid advancement of Artificial Intelligence (AI) has significantly transformed academic writing practices, offering both opportunities and ethical challenges, particularly related to plagiarism and over-reliance on automated tools.Objectives: This study aims to evaluate the effectiveness of AI-assisted scientific writing training in enhancing participants’ writing skills, academic integrity awareness, and writing efficiency through a Human-AI Collaboration approach.Methods: A quantitative descriptive method was employed using a structured questionnaire consisting of 20 Likert-scale items administered to 85 participants. The instrument measured five dimensions: relevance and conceptual understanding, instructional quality, writing skill improvement, academic integrity awareness, and perceived impact. Data were analyzed using descriptive statistics, including mean scores and percentage distributions.Results: The findings reveal that the training achieved a very high level of effectiveness, with an overall mean score of 3.59. All variables were categorized as very high, with academic integrity awareness obtaining the highest mean score (3.70), indicating strong improvement in ethical understanding. Writing skill improvement showed relatively lower scores, suggesting the need for continuous practice.Conclusion: AI-assisted scientific writing training based on the Human-AI Collaboration framework is highly effective in improving scientific writing competence and promoting ethical awareness. This approach provides a balanced model integrating technical skills and academic integrity in AI-supported writing practices.
Mapping the landscape of a decade: a bibliometric review of mobile assistive technology research for dyslexic children Mariam Mohamad; Noratikah Abdullah; Mageswaran Sanmugam
Journal of Education and Learning (EduLearn) Vol 20, No 1: February 2026
Publisher : Intelektual Pustaka Media Utama

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.11591/edulearn.v20i1.22020

Abstract

Dyslexia, a prevalent neurodevelopmental disorder affecting reading and writing skills, poses significant challenges to educational attainment. In recent years, mobile assistive technologies have emerged as promising tools to support dyslexic children in their journey. Despite the increasing focus on mobile assistive technology for dyslexic children, a comprehensive overview of the research landscape is lacking. The field is characterized by a proliferation of studies, diverse methodologies, and an expanding knowledge base. This bibliometric review leverages advanced analytical tools, with a primary focus on the VOSviewer software, to conduct a comprehensive analysis of the 53 literatures spanning from 2014 to 2024. A carefully curated dataset, comprising research articles sourced from reputable databases, forms the basis of the analysis. Anticipated outcomes include visually rich maps depicting the keyword co-occurrence patterns within the realm of mobile assistive technology for dyslexic children. We expect to identify key articles shaping the field, prominent clusters of research, and evolving trends. This bibliometric review aspires to contribute a panoramic view of the last decade's research landscape in mobile assistive technology for dyslexic children. The anticipated insights hold the potential to guide future research directions, technological innovations, and educational interventions, ultimately enhancing the support available to dyslexic children through mobile assistive technologies.
Enhancing primary English writing with authentic learning in mobile cloud Chin Da Bun Tiang; Mariam Mohamad; Mageswaran Sanmugam
Journal of Education and Learning (EduLearn) Vol 20, No 2: May 2026
Publisher : Intelektual Pustaka Media Utama

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.11591/edulearn.v20i2.22019

Abstract

This qualitative study examines the English writing experiences of 12 primary schoolchildren (PSC) from a vernacular school in Malaysia, employing Herrington and Kervin’s principles of authentic learning as its theoretical framework. Data were collected through in-depth one-on-one interviews, e-diaries, and unstructured non-participant observations, and analyzed using thematic analysis. The findings reveal that the participants effectively employed authentic learning strategies (AuLStra) in their writing activities, which included online reflective writing, collaborative online writing, and peer feedback. The use of AuLStra facilitated a collaborative environment where participants engaged in authentic writing tasks, benefiting from teacher scaffolding and peer assistance. Participants noted that reflecting on meaningful personal topics and engaging in creative writing through online collaboration significantly enhanced their writing fluency. The study highlights the impact of authentic learning practices on writing development and offers insights into the pedagogical and theoretical implications of integrating such strategies in primary education.
Development of mobile application-based digital drama for Arabic language to improve students’ vocabulary Nurul Azni Mhd Alkasirah; Mariam Mohamad; Mageswaran Sanmugam; Girija Ramdas
Journal of Education and Learning (EduLearn) Vol 20, No 2: May 2026
Publisher : Intelektual Pustaka Media Utama

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.11591/edulearn.v20i2.23655

Abstract

The Arabic language is one of the foreign languages offered in Malaysian education. In Arabic, learning vocabulary is particularly important because it forms the foundation of language acquisition. However, past research showed that a large number of Malaysian students struggle with understanding and memorizing Arabic vocabulary because they are unfamiliar with it. Thus, this research focuses on developing an appropriate mobile application that will support Malaysian religious secondary school students in improving their mastery of Arabic vocabulary. This study attempts to create a mobile application-based digital drama by adopting the analyze, design, develop, implement, and evaluate (ADDIE) model. The ADDIE model was integrated as a structured approach for producing educational material that is both enjoyable and beneficial to students. Six subject matter experts were invited to evaluate the effectiveness of the mobile application-based digital drama. Based on the validation results and feedback obtained, it can be concluded that mobile application-based digital drama was valid and effective with high scores for format and graphic suitability to be utilized as an instructional tool for learning Arabic vocabulary among Malaysian religious secondary school students.
Artificial Intelligence for Social Innovation: A Systematic Literature Review of Applications, Societal Impact, Ethical Challenges, and Future Research Directions Melani Maulia Apriana; Syifahul Husna; Arsela Eko Listiono; Mageswaran Sanmugam; Muhammad Qais Rezvani
Journal of Science, Technology, and Social Innovation Vol. 2 No. 1 (2026): Journal of Science, Technology, and Social Innovation
Publisher : CV Media Inti Teknologi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.58723/jstsi.v2i1.174

Abstract

Purpose of the study: In this article, the development of research on the application of Artificial Intelligence (AI) in supporting social innovation over the period 2005-2026 is systematically mapped, covering applications, societal impact, ethical challenges, and future research directions.Methodology: This study combines a Systematic Literature Review (SLR) following the PRISMA 2020 guidelines with bibliometric analysis. Data were collected from the Scopus database using four keyword combinations related to Artificial Intelligence, social innovation, social sustainability, and societal impact, yielding 101 initial records; after duplicate removal with OpenRefine, 90 documents were analyzed with Bibliometrix/Biblioshiny and VOSviewer, and 39 of them were further processed with Elicit AI.Main Findings: Publications increased sharply, peaking in 2025 with 32 articles. Systematic literature reviews (25.6%) and conceptual studies (12.8%) dominate the literature, while quantitative/experimental studies, mixed methods, and model/simulation development each account for 12.8%. The reviewed studies address AI applications in smart city governance, community economic empowerment, education, health, and environmental sustainability, and highlight ethical challenges such as algorithmic bias, the digital divide, and the lack of participatory governance frameworks.Novelty/Originality of this study: This study integrates bibliometric analysis with a systematic literature review to comprehensively map the relationship between Artificial Intelligence and social innovation, offering an integrated overview of publication trends, thematic clusters, research gaps, and future directions that can serve researchers, policymakers, and practitioners designing evidence-based, responsible AI-driven social innovation strategies.
Co-Authors Arsela Eko Listiono Chin Da Bun Tiang Eko Risdianto Girija Ramdas Jaesik Jeong Kit Ling Chan M. Abdul Jamal M. Esad Kuloğlu Mariam Mohamad Melani Maulia Apriana Mohammad Qaiz Rezvani Muhammad Qais Rezvani Nanik Setyowati Noratikah Abdullah Nurul Azni Mhd Alkasirah Syifahul Husna