International Journal of Education and Management Engineering (IJEME)
IJEME Vol. 16, No. 4, 8 Aug. 2026
Cover page and Table of Contents: PDF (size: 793KB)
Edge-First Adaptive Learning with Lightweight RL and LNN for STEM Education in Low- Resource Kenyan Schools
PDF (793KB), PP.71-88
Views: 0 Downloads: 0
Author(s)
Index Terms
Adaptive Learning, Artificial Intelligence, STEM Education, Resource-Constraint Schools, Reinforcement Learning, Educational Technology, Kenya
Abstract
While artificial intelligence (AI) holds significant promise for personalizing education, research on AI applications in resource-constrained settings remains limited. This paper describes a proposed AI-based adaptive learning system that aims to personalize STEM education in a low-resource school environment in Kenya. The system addresses key challenges including irregular internet connectivity, limited hardware infrastructure, and teachers’ lack of prior experience with AI technologies. To address these issues, an adapted reinforcement learning algorithm will personalize the content shown to students, and a modified liquid neural network is used for the prediction of student success, while not being computationally expensive. As compared to traditional adaptive systems, this adaptive learning platform supports edge computing and offline updates in order to operate in a consistently low connectivity environment. Furthermore, the platform improves the educational experience by dynamically adjusting content difficulty, format, and pacing to align with individual students’ learning styles, prior knowledge, and attention levels. An 8-month case study was done with 6 Kenyan low-resource schools and 6 comparison schools located in the city. In this case, we report a 31% increase in the level of understanding of students’ key STEM subjects, 27% reduction in student drop rate from STEM topics, and 43% increase in teacher efficiency over traditional methods. Additionally, the system achieved 89% prediction accuracy with a memory footprint of only 1.9 MB, making it feasible for deployment on budget Android devices. This study presents evidence that the utilization of AI for personalized adaptive learning technologies in order to minimize the disparities in the provision of STEM education in low-resource settings worldwide is possible.
Cite This Paper
David Shiala Ongoma, “Edge-First Adaptive Learning with Lightweight RL and LNN for STEM Education in Low-Resource Kenyan Schools”, International Journal of Education and Management Engineering (IJEME), Vol.16, No.4, pp. 71-88, 2026. DOI:10.5815/ijeme.2026.04.06
Reference
[1]UNESCO, "Global Education Monitoring Report: Technology in Education," UNESCO Publishing, Paris, 2023. Available at: https://www.unesco.org/gem-report/en/technology
[2]World Bank, "Facing Forward: Schooling for Learning in Africa," World Bank Group, Washington DC, 2024. https://www.worldbank.org/en/region/afr/publication/facing-forward-schooling-for-learning-in-africa
[3]Kenya National Bureau of Statistics, "Economic Survey 2024," Government Printer, Nairobi, 2024. PDF: https://www.knbs.or.ke/wp-content/uploads/2024/05/2024-Economic-Survey.pdf.
[4]J.A. Banks, “Networking as a Strategy for Improvement of Science and Mathematics Teaching and Learning in Kenyan Secondary Schools,” Journal of Education and Practice, Jul. 2019. DOI: 10.7176/jep/10-21-16
[5]Kenya Institute of Curriculum Development, "Competency-Based Curriculum Framework," KICD, Nairobi, 2022. Available: https://kicd.ac.ke
[6]J.A. Banks, “Digital Divide, Digital Learning, and Equal Access,” Encyclopedia of Diversity in Education, 2012, DOI: 10.4135/9781452218533.n201.
[7]P. Amolloh Odundo and C. Richard Oyier, “Influence Of Policy Framework On Budgeting For Science Instructional Resources In Kenyan Secondary Schools,” Archives of Business Research, vol. 5, no. 4, Apr. 2017, doi: 10.14738/abr.54.2944.
[8]Dr. Priyanka Datta, “Artificial Intelligence and Education,” Artificial Intelligence in Education: Revolutionizing Learning And Teaching, Oct. 2020, doi: 10.25215/9358094575.32.
[9]Y. Gao, “A Meta-Analysis of Adaptive Learning Systems on Learning Achievement (Poster 16),” Proceedings of the 2024 AERA Annual Meeting, 2024, doi: 10.3102/2105897.
[10]Carnegie Learning, "MATHia Research and Efficacy Studies," 2023. Available: https://www.carnegielearning.com
[11]D. I. Okonkwo and E. A. Ade-Ibijola, "Chatbots Applications in Education: A Systematic Review," Comput. Educ. Artif. Intell., vol. 2, 2021. DOI: 10.1016/j.caeai.2021.100033
[12]Communications Authority of Kenya, "Sector Statistics Report Q4 2024," CAK, Nairobi, 2025. Available: https://www.ca.go.ke
[13]J. Wishart, “Current and recent mobile learning teaching and learning initiatives in secondary schools,” Mobile Learning in Schools, pp. 36–49, Oct. 2017, doi: 10.4324/9781315536774-4.
[14]A. Ade-Ibijola and C. Okonkwo, "Artificial Intelligence in Africa: Emerging Challenges," in Social and Cultural Studies of Robots and AI, Springer, 2023, doi: 10.1007/978-3-031-08215-3_5.
[15]Tondeur et al., "Preparing Pre-Service Teachers to Integrate Technology in Education: A Synthesis of Qualitative Evidence," Comput. Educ., vol. 59, no. 1, pp. 134-144, 2012. DOI: 10.1016/j.compedu.2011.10.009
[16]UNESCO, "Cost-Effectiveness of Educational Technology in Low-Income Countries," IIEP Research Paper, Paris, 2024. Available: https://www.iiep.unesco.org
[17]Khan Academy, "Personalized Learning: Technical Architecture and Impact," Khan Academy Technical Report, 2023. Available: https://www.khanacademy.org
[18]J. R. Anderson, et al., "Cognitive Tutors: Lessons Learned," Journal of the Learning Sciences, vol. 4, no. 2, pp. 167-207, 1995. DOI: 10.1207/s15327809jls0402_2
[19]A. Mutua Nzioki, “Improved Learning Management System (i-LMS): A Flat Form for Content Creation and Sharing for Kenyan Secondary Schools,” International Journal of Computer Applications Technology and Research, vol. 4, no. 2, pp. 86–91, Feb. 2015, doi: 10.7753/ijcatr0402.1001
[20]DreamBox Learning, "Intelligent Adaptive Learning Engine White Paper," DreamBox, Bellevue, WA, 2024. Available: https://www.dreambox.com
[21]Y. Shi, E. Yang, and J. Wu, "Edge Computing: Vision and Challenges for Internet of Things Applications," IEEE Internet Things J., vol. 3, no. 5, pp. 637-646, 2016. DOI: 10.1109/JIOT.2016.2579198
[22]C. Piech, et al., "Deep Knowledge Tracing," Advances in Neural Information Processing Systems, vol. 28, pp. 505-513, 2015. DOI: 10.48550/arXiv.1506.05908
[23]A. Ayari, M. Chaabouni, and H. Ben Ghezala, “A Deep Learning Approach for Automatic Detection of Learner Engagement in Educational Context,” Proceedings of the 17th International Conference on Computer Supported Education, pp. 372–379, 2025, doi: 10.5220/0013283200003932.
[24]V. Mnih, et al., "Human-level Control through Deep Reinforcement Learning," Nature, vol. 518, pp. 529-533, 2015. DOI: 10.1038/nature14236
[25]B. McMahan, et al., "Communication-Efficient Learning of Deep Networks from Decentralized Data," Proceedings of AISTATS, 2017. DOI: 10.48550/arXiv.1602.05629
[26]Kenya Education Cloud, "E-Content Utilization Report 2024," KICD, Nairobi, 2024. Available: https://kec.ac.ke
[27]Eneza Education, "Impact Report: SMS-based Learning in Kenya," Eneza Education, Nairobi, 2023. Available: https://enezaeducation.com
[28]CenturyTech, "Adaptive Learning in Emerging Markets: Pilot Results," Century Tech, London, 2024. Available: https://www.century.tech
[29]A. K. Soma, "Enhancing Supply Chain Transparency and Integrity: A Permissioned Blockchain Framework," in 2025 International Conference on Emerging Systems and Intelligent Computing (ESIC), pp. 819-826, 2025. doi: 10.1109/esic64052.2025.10962720
[30]A. kumar Soma, "Hybrid RNN-GRU-LSTM Model for Accurate Detection of DDoS Attacks on IDS Dataset", Journal of Modern Technology, vol. 2, no. 01, pp. 283-291, 2025. doi: 10.71426/jmt.v2.i1.pp283-291
[31]P. K. Udayaprasad, et al., "Energy Efficient Optimized Routing Technique with Distributed SDN-AI to Large Scale I-IoT Networks," IEEE Access, vol. 12, pp. 2742-2759, 2024. doi: 10.1109/access.2023.3346679
[32]C. L. D. Santos, et al., "ML-RPL: Machine Learning-Based Routing Protocol for Wireless Smart Grid Networks," IEEE Access, vol. 11, pp. 57401-57414, 2023. DOI: 10.1109/ACCESS.2023.3283208
[33]M. A. Islam, R. Atat and M. Ismail, "Software-Defined Networking-Based Resilient Proactive Routing in Smart Grids Using Graph Neural Networks and Deep Q-Networks," IEEE Access, vol. 12, pp. 111169-111186, 2024. doi: 10.1109/access.2024.3438938