AI-based Q-learning model for personalized learning strategies for students with disabilities

Theyazn H.H Aldhyani; Samina Amin; Mossab Saud Alholiby; M. Irfan Uddin; Theyazn H.H Aldhyani; Samina Amin; Mossab Saud Alholiby; M. Irfan Uddin

doi:10.3934/steme.2026025

STEM Education

2026, Volume 6, Issue 4: 606-631. doi: 10.3934/steme.2026025

Previous Article Next Article

Research article Topical Sections

AI-based Q-learning model for personalized learning strategies for students with disabilities

1.
Applied College in Abqaiq, King Faisal University, P.O. Box 400, Al-Ahsa 31982, Saudi Arabia; taldhyani@kfu.edu.sa
2.
Institute of Computing, Kohat University of Science and Technology, Pakistan; samina@kust.edu.pk
3.
Educational Leadership Department, Education College, King Faisal University, P.O. Box 400, Al-Ahsa 31982, Saudi Arabia; Malholiby@kfu.edu.sa
4.
Department of Computer Science, University of Swabi, Swabi 23540, KP, Pakistan; irfanuddin@uoswabi.edu.pk

Academic Editor: Jun Shen

Received: 02 March 2026 Revised: 08 May 2026 Accepted: 13 May 2026 Published: 10 June 2026

Students with disabilities find learning challenging in traditional learning environments due to the lack of instructional strategies that are not personalized or adaptive. Various studies have focused on identifying learning difficulties such as dyslexia, dysgraphia, and dyscalculia, which require a multi-step screening process under the supervision of psychologists. Identifying these difficulties is challenging but essential, as it impacts a student's learning and academic success. Everyone's comprehension ability depends on several factors, including the experience and knowledge they bring to the learning environment. With the evolution of technology and the advancement of e-learning platforms, adaptive e-learning has bridged the gap between students' needs and educational institutions' extra classes, enabling students to select targeted courses aligned with their interests. Since the onset of COVID-19, universities have recognized the necessity of online learning and have continued to use these platforms for student assessment. Educational institutions seek innovative strategies to enhance personalized learning (PL) for students with disabilities. The use of technology in schools has created new opportunities for PL, especially for students with disabilities. Traditional learning systems frequently neglect the diverse focuses and distinct requirements of students with visual, cognitive, motor, or auditory impairments. Because of this lack of flexibility, students with disabilities are less likely to be engaged and to complete tasks. Identifying and using effective learning strategies that work for each student remains a significant challenge in education. To address this issue, this paper proposes a reinforcement learning–based PL system for students with disabilities (PLS-SD) using Q-learning. We suggest PL actions, including audio instructions, augmented reality, and text-based resources. A reward system based on real-world outcomes, i.e., how well students complete their work and how engaged they are, helps them learn. Experimental results demonstrate that the proposed approach effectively identifies optimal actions for different learner states, with immersive and adaptive strategies, such as augmented reality and interactive content, consistently achieving higher rewards. The model shows stable learning behavior across training episodes and successfully adapts its policy to maximize learner engagement and task completion. As indicated by comparing the results with currently advanced models, the proposed method outperforms approaches that are considered traditional by providing context-aware and adaptable recommendations. These findings highlight the potential of reinforcement learning to support scalable and personalized educational solutions for diverse learners.
- reinforcement learning,
- Q learning,
- artificial intelligence,
- neural network,
- modeling
Citation: Theyazn H.H Aldhyani, Samina Amin, Mossab Saud Alholiby, M. Irfan Uddin. AI-based Q-learning model for personalized learning strategies for students with disabilities[J]. STEM Education, 2026, 6(4): 606-631. doi: 10.3934/steme.2026025

Related Papers:

Abstract

Students with disabilities find learning challenging in traditional learning environments due to the lack of instructional strategies that are not personalized or adaptive. Various studies have focused on identifying learning difficulties such as dyslexia, dysgraphia, and dyscalculia, which require a multi-step screening process under the supervision of psychologists. Identifying these difficulties is challenging but essential, as it impacts a student's learning and academic success. Everyone's comprehension ability depends on several factors, including the experience and knowledge they bring to the learning environment. With the evolution of technology and the advancement of e-learning platforms, adaptive e-learning has bridged the gap between students' needs and educational institutions' extra classes, enabling students to select targeted courses aligned with their interests. Since the onset of COVID-19, universities have recognized the necessity of online learning and have continued to use these platforms for student assessment. Educational institutions seek innovative strategies to enhance personalized learning (PL) for students with disabilities. The use of technology in schools has created new opportunities for PL, especially for students with disabilities. Traditional learning systems frequently neglect the diverse focuses and distinct requirements of students with visual, cognitive, motor, or auditory impairments. Because of this lack of flexibility, students with disabilities are less likely to be engaged and to complete tasks. Identifying and using effective learning strategies that work for each student remains a significant challenge in education. To address this issue, this paper proposes a reinforcement learning–based PL system for students with disabilities (PLS-SD) using Q-learning. We suggest PL actions, including audio instructions, augmented reality, and text-based resources. A reward system based on real-world outcomes, i.e., how well students complete their work and how engaged they are, helps them learn. Experimental results demonstrate that the proposed approach effectively identifies optimal actions for different learner states, with immersive and adaptive strategies, such as augmented reality and interactive content, consistently achieving higher rewards. The model shows stable learning behavior across training episodes and successfully adapts its policy to maximize learner engagement and task completion. As indicated by comparing the results with currently advanced models, the proposed method outperforms approaches that are considered traditional by providing context-aware and adaptable recommendations. These findings highlight the potential of reinforcement learning to support scalable and personalized educational solutions for diverse learners.

References

[1]	Alsolami, A. S., The Effectiveness of Using Artificial Intelligence in Improving Academic Skills of School-Aged Students with Mild Intellectual Disabilities in Saudi Arabia. Research in Developmental Disabilities, 2025, 156: 104884. https://doi.org/10.1016/j.ridd.2024.104884 doi: 10.1016/j.ridd.2024.104884
[2]	Fu, L., The Role of STEM Teachers' Emotional Intelligence and Psychological Well-Being in Predicting Their Artificial Intelligence Literacy. Acta Psychologica, 2025, 253: 104708. https://doi.org/10.1016/j.actpsy.2025.104708 doi: 10.1016/j.actpsy.2025.104708
[3]	Zhang, L., Basham, J. D. and Yang, S., Understanding the Implementation of Personalized Learning: A Research Synthesis. Educational Research Review, 2020, 31: 100339. https://doi.org/10.1016/j.edurev.2020.100339 doi: 10.1016/j.edurev.2020.100339
[4]	Hocine, N. and Sehaba, K., A Systematic Review of Online Personalized Systems for the Autonomous Learning of People with Cognitive Disabilities. Human–Computer Interaction, 2024, 39 (3–4): 174–205.
[5]	Nganji, J. T. and Brayshaw, M., Disability-Aware Adaptive and Personalised Learning for Students with Multiple Disabilities. The International Journal of Information and Learning Technology, 2017, 34(4): 307–21. https://doi.org/10.1108/IJILT-08-2016-0027 doi: 10.1108/IJILT-08-2016-0027
[6]	Normadhi, N. B. A., Shuib, L., Nasir, H. N. M., Bimba, A., Idris, N. and Balakrishnan, V., Identification of Personal Traits in Adaptive Learning Environment: Systematic Literature Review. Computers & Education, 2019, 130: 168–90. https://doi.org/10.1016/j.compedu.2018.11.005 doi: 10.1016/j.compedu.2018.11.005
[7]	Strielkowski, W., Grebennikova, V., Lisovskiy, A., Rakhimova, G. and Vasileva, T., AI‐driven Adaptive Learning for Sustainable Educational Transformation. Sustainable Development, 2025, 33(2): 1921‒1947.
[8]	Zhang, F., Feng, X. and Wang, Y., Personalized Process–Type Learning Path Recommendation Based on Process Mining and Deep Knowledge Tracing. Knowledge-Based Systems, 2024, 303: 112431. https://doi.org/10.1016/j.knosys.2024.112431 doi: 10.1016/j.knosys.2024.112431
[9]	Pliakos, K., Joo, S. H., Park, J. Y., Cornillie, F., Vens, C. and Van den Noortgate, W., Integrating Machine Learning into Item Response Theory for Addressing the Cold Start Problem in Adaptive Learning Systems. Computers & Education, 2019, 137: 91–103. https://doi.org/10.1016/j.compedu.2019.04.009 doi: 10.1016/j.compedu.2019.04.009
[10]	Gligorea, I., Cioca, M., Oancea, R., Gorski, A. T., Gorski, H. and Tudorache, P., Adaptive Learning Using Artificial Intelligence in E-Learning: A Literature Review. Education Sciences, 2023, 13(12): 1216.
[11]	Hung, Y. H., Chang, R. I. and Lin, C. F., Hybrid Learning Style Identification and Developing Adaptive Problem-Solving Learning Activities. Computers in Human Behavior, 2016, 55: 552–61. https://doi.org/10.1016/j.chb.2015.07.004 doi: 10.1016/j.chb.2015.07.004
[12]	Hills, K., Andersen, K. and Davidson, S., Personalized Learning and Teaching Approaches to Meet Diverse Needs: A Prototype Tertiary Education Program. Reimagining Christian Education: Cultivating Transformative Approaches, 2018, 233–57. Singapore: Springer Singapore. https://doi.org/10.1007/978-981-13-0851-2_16
[13]	Amin, S., Uddin, M. I., Alarood, A. A., Mashwani, W. K., Alzahrani, A. and Alzahrani, A. O., Smart E-Learning Framework For Personalized Adaptive Learning and Sequential Path Recommendations Using Reinforcement Learning. IEEE Access, 2013, 11: 89769–90. https://doi.org/10.1109/ACCESS.2023.3305584 doi: 10.1109/ACCESS.2023.3305584
[14]	Amin, S., Uddin, M. I., Alarood, A. A., Mashwani, W. K., Alzahrani, A. O. and Alzahrani, H. A., An Adaptable and Personalized Framework for Top-N Course Recommendations in Online Learning. Scientific Reports, 2024, 14(1): 10382. https://doi.org/10.1038/s41598-024-56497-1 doi: 10.1038/s41598-024-56497-1
[15]	Essa, S. G., Celik, T. and Human-Hendricks, N. E., Personalized Adaptive Learning Technologies Based on Machine Learning Techniques to Identify Learning Styles: A Systematic Literature Review. IEEE Access, 2023, 11: 48392–409.
[16]	Isabona, J., Imoize, A. L. and Kim, Y., Machine Learning-Based Boosted Regression Ensemble Combined with Hyperparameter Tuning for Optimal Adaptive Learning. Sensors, 2022, 22(10): 3776.
[17]	Sutton, R. S. and Barto, A. G., Reinforcement Learning : An Introduction, 2nd ed. The MIT Press, 2018.
[18]	Müller, H., Berg, L. and Kudenko, D., Using Incomplete and Incorrect Plans to Shape Reinforcement Learning in Long-Sequence Sparse-Reward Tasks. Neural Computing and Applications, 2025, 37(23): 18851‒66. https://doi.org/10.1007/s00521-024-10615-2 doi: 10.1007/s00521-024-10615-2
[19]	Dutt, S., Ahuja, N. J. and Kumar, M., An Intelligent Tutoring System Architecture Based on Fuzzy Neural Network (FNN) for Special Education of Learning Disabled Learners. Education and Information Technologies, 2022, 27(2): 2613–33.
[20]	Minoofam, S. A. H., Bastanfard, A. and Keyvanpour, M. R., RALF: An Adaptive Reinforcement Learning Framework for Teaching Dyslexic Students. Multimedia Tools and Applications, 2022, 81(5): 6389–6412.
[21]	Modak, M. M., Gharpure, P. and M, S., Adaptive Learning and Correlative Assessment of Differential Usage Patterns for Students With-or-without Learning Disabilities via Learning Analytics. ACM Transactions on Asian and Low-Resource Language Information Processing, 2023, 22(12): 1–25.
[22]	Khabbaz, A. H., Pouyan, A., Fateh, M. and Abolghasemi, V., An Adaptive Learning Game for Autistic Children Using Reinforcement Learning and Fuzzy Logic. Journal of AI and Data Mining, 2019, 7(2): 321–29.
[23]	Ma, D., Zhu, H., Liao, S., Chen, Y., Liu, J., Tian, F., et al., Learning Path Recommendation with Multi-Behavior User Modeling and Cascading Deep Q Networks. Knowledge-Based Systems, 2024, 294: 111743. https://doi.org/10.1016/j.knosys.2024.111743 doi: 10.1016/j.knosys.2024.111743
[24]	Liu, Z., Hou, J., Ning, D., Zhou, C., Liang, G. and Zhang, F., Improving Deep Q Network Based on Marketing Psychology for AUV Path Planning in Unknown Marine Environments. IEEE Internet of Things Journal, 2024, 12(5): 5476‒5487. https://doi.org/10.1109/JIOT.2024.3487129 doi: 10.1109/JIOT.2024.3487129
[25]	Shawky, D. and Badawi, A., Towards a Personalized Learning Experience Using Reinforcement Learning. Machine Learning Paradigms: Theory and Application, 2018, 169–87. Cham: Springer International Publishing. https://doi.org/10.1007/978-3-030-02357-7_8
[26]	Islam, M. Z., Ali, R., Haider, A., Islam, M. Z. and Kim, H. S., PAKES: A Reinforcement Learning-Based Personalized Adaptability Knowledge Extraction Strategy for Adaptive Learning Systems. IEEE Access, 2021, 9: 155123–37. https://doi.org/10.1109/ACCESS.2021.3128578 doi: 10.1109/ACCESS.2021.3128578
[27]	Yuhana, U. L., Djunaidy, A. and Purnomo, M. H., Enhancing Students Performance through Dynamic Personalized Learning Path Using Ant Colony and Item Response Theory (ACOIRT). Computers and Education: Artificial Intelligence, 2024, 7: 100280. https://doi.org/10.1016/j.caeai.2024.100280 doi: 10.1016/j.caeai.2024.100280
[28]	Sajja, R., Sermet, Y., Cikmaz, M., Cwiertny, D. and Demir, I., Artificial Intelligence-Enabled Intelligent Assistant for Personalized and Adaptive Learning in Higher Education. Information, 2024, 15(10): 596.
[29]	Demartini, C. G., Sciascia, L., Bosso, A. and Manuri, F., Artificial Intelligence Bringing Improvements to Adaptive Learning in Education: A Case Study. Sustainability, 2024, 16(3): 1347.
[30]	Khoso, A. K., Honggang, W. and Darazi, M. A., Trust and attitude towards AI as pathways to creativity: a TAM Model study of EFL students' digital literacy and AI acceptance. Humanit Soc Sci Commun, 2026, 13: 69. https://doi.org/10.1057/s41599-025-06362-x doi: 10.1057/s41599-025-06362-x
[31]	Khoso, A. K., Honggang, W., Tahir, S. R., Nusrat, A. and Younas, M., Integrating Sustainable Development Goals (SDGs) and Generative AI to Enhance Language Digital Literacy and Creativity in EFL Learning Environments. J Vis Exp, 2026, (228): e69445. https://doi.org/10.3791/69445 doi: 10.3791/69445
[32]	Khoso, A. K., Honggang, W., Younas, M., Salam E. D., DA The Dual Forces of AI: How Generative AI and Perceived AI Dependency Influence Fear of Missing Out (FoMO) and EFL Students' Vocabulary Acquisition. J Vis Exp, 2025, 226: e69637. https://doi.org/10.3791/69637 doi: 10.3791/69637
[33]	Khoso, A. K., Honggang, W. and Darazi, M. A., Empowering creativity and engagement: The impact of generative artificial intelligence usage on Chines EFL students' language learning experience. Computers in Human Behavior Reports, 2025, 18: 100627. https://doi.org/10.1016/j.chbr.2025.100627 doi: 10.1016/j.chbr.2025.100627
[34]	Jang, B., Kim, M., Harerimana, G. and Kim, J. W., Q-Learning Algorithms: A Comprehensive Classification and Applications. IEEE Access, 2019, 7: 133653–67.
[35]	Fan, J., Wang, Z., Xie, Y. and Yang, Z., A Theoretical Analysis of Deep Q-Learning. In Learning for Dynamics and Control, 2020, 486–89. PMLR.
[36]	Spano, S., Cardarilli, G. C., Di Nunzio, L., Fazzolari, R., Giardino, D., Matta, M., et al., An Efficient Hardware Implementation of Reinforcement Learning: The q-Learning Algorithm. Ieee Access, 2019, 7: 186340–51.
[37]	Doshi, K., Reinforcement Learning Explained Visually, 2020. Available from: https://towardsdatascience.com/reinforcement-learning-explained-visually-part-4-q-learning-step-by-step-b65efb731d3e

Author's biography Theyazn H.H Aldhyani: In 2017, he was awarded the Ph.D. degree in Computer Science and Information Technology from NMU University. His areas of research interest are Artificial Intelligence, Machine Learning, Soft Computing, Big Data, Healthcare information, deep learning, cybersecurity, and IoT. He is currently an associate professor in the Faculty of Computer Science and Information Technology at King Faisal University. He has published over 35 research papers in highly reputable journals published by MDPI, Springer, and IEEE. He is a Reviewer in MDPI, Springer, IEEE, and Elsevier; Samina Amin: A passionate researcher in Computer Science, specializing in Artificial Intelligence (AI) and Machine Learning (ML). She earned his Ph.D. in Computer Science in December 2024 from the Institute of Computing, Kohat University of Science & Technology, Pakistan, where she also completed my Master's degree in 2021. Her research primarily focused on leveraging reinforcement learning (RL) to enhance online learning through intelligent algorithms that recommend personalized course content; Mossab Saud Alholiby: Dr. Mossab Saud Alholiby: Associate Professor of Educational Leadership and Executive President of the Applied College at King Faisal University. He received his Ph.D. in Higher Education Management from the University of Glasgow and his Master's degree in Educational Administration from King Faisal University. He has held several academic and administrative leadership positions, including Assistant Vice President for Academic Affairs and Advisor to the Vice President for Academic Affairs at King Faisal University. His research interests include higher education leadership, quality assurance and academic accreditation, strategic planning, digital transformation, organizational effectiveness, and the use of artificial intelligence in improving administrative and educational processes. He has published a number of research papers in the fields of educational leadership, governance, quality in higher education, and organizational development, and has actively contributed to strategic and academic initiatives at both university and national levels; M. Irfan Uddin: With a solid educational foundation and over two decades of teaching and research experience at renowned academic institutions, he possesses a robust academic and research background across diverse domains of computer science. He is an active member of several esteemed scientific societies, including IEEE, ACM, HiPEAC, CSTA, IAENG, KSS, and Science-i. He has played a leading role in organizing numerous national and international seminars, workshops, and conferences. His research contributions include over 130 research articles published in JCR and Scopus/ISI-indexed journals, as well as national/international conferences, in addition to three authored books published by reputed publishers. He has also published two patents with the United States Patent and Trademark Office. He has served as (PI/Co-PI/Collaborator) in various nationally/internationally funded research projects. Additionally, he actively contributes as a reviewer, editorial board member, and technical program committee member for several prestigious journals and conferences

Reader Comments

Your name:*

Email:*
© 2026 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)