Introduction:
• Overview:
In the modern world, e-learning has become a crucial educational tool. E-learning gives students
several benefits, such as accessibility to content at any time and place, as well as at their own
pace. While classroom availability of knowledge may be subject to several considerations, such as
space and resources, e-learning's availability of information is generally fast and easily accessible.
E-learning systems are also based on interactive multimedia content with self-directed learner
immersion. All these support a paradigm shift in education for greater results and personalization,
not just convenience, in today's rapidly changing world [1] [2]. During the COVID-19 pandemic,
virtual classrooms served as an important reminder to enhance scalable and user-friendly e-
learning systems [3]. In order to be effective, e-learning should not just be about sharing
disconnected information; it needs to be able to allow the student to be autonomous as well as
have a sound pedagogical base and an integrated framework of ongoing feedback. As defined by
the present digital age, there are many smarter education-supporting technologies, aside from just
content management, which systematically boost learner success as well as educational quality.
References:
The majority of the large language models (LLMs) currently in use in e-learning strategies include
recommendation engines, adaptive assessment systems, and personalised feedback mechanisms
that use data-driven algorithms and natural language processing (NLP) to customise instructional
content [7][8]. In order to improve engagement and learning outcomes, these LLM-based systems
usually use learner data analysis to produce personalised recommendations, formative tests, and
real-time feedback. But even with their sophisticated features, these models frequently struggle to
meet the various and ever-changing needs of individual students. A one-size-fits-all methodology
that impedes true personalisation is the result of many current approaches' reliance on static or
rule-based adaptation strategies, which are inflexible enough to ignore learners' backgrounds, prior
knowledge, or distinct learning styles [4][5]. Furthermore, the majority of existing models are limited
in their capacity to provide truly customised instructional paths due to their inability to accurately
model complex learner profiles over time. This disparity emphasises how urgently more advanced,
flexible LLM frameworks are needed in order to support individualised learning at scale and react
dynamically to changing learner characteristics [6].
A prominent gap in current e-learning methods, particularly those utilising Moodle for personalised
assessment and learning path generation, emerges consistently across multiple studies [12] [13]
[17]. While existing systems incorporate foundational adaptive mechanisms, such as learning style
detection and rule-based content adaptation, they often lack dynamic, real-time responsiveness to
individual student data [16][18][19]. In order to fully support self-regulated learning and sustained
engagement, many existing solutions rely on static models for learner profiling [5][14].
Furthermore, integrating these technologies seamlessly within scalable, open-source Moodle
environments [9][18] is still a major challenge, despite advancements in AI and algorithmic
approaches for quiz generation and feedback [15][18]. The effectiveness and broad adoption of
these customised e-learning systems are further limited by the heterogeneity of educational
contexts, scalability concerns, and the lack of empirical validation of long-term effects [10][11]. In
order to promote deeper learning outcomes and learner autonomy, it is imperative to create
increasingly complex, adaptable, and flexible frameworks that can dynamically model learner
profiles and offer personalised pathways in real-time.
By utilising Large Language Models (LLMs), the suggested solution fills in the gaps in personalised
adaptive learning and offers a novel way to improve Moodle-based e-learning. In contrast to
conventional Moodle plugins that depend on static, rule-based adaptation, this approach gives
students the ability to actively design unique quizzes that are geared towards particular weeks or
subjects of interest, allowing for knowledge checks and self-evaluation whenever it is most
convenient for them. A deeper understanding is facilitated by the instant, thorough feedback that
students receive after finishing these tests, which includes scores and explanations of both right
and wrong answers. Most importantly, the system then makes use of LLMs to create customised
learning plans that are based on individual performance data. These plans identify the student's
areas of weakness and suggest learning activities that are in line with their needs. This learning
plan is incorporated within the student's grade report, making it simple to access, download
individual basis. This effectively promotes independent, self-directed learning. This method
provides a flexible, real-time, data-driven solution that adjusts to each learner's individual strengths
and weaknesses, thereby directly addressing the gaps found in previous research, including limited
real-time personalisation, inadequate feedback depth, and lack of dynamic learning plan
generation. By closely matching real-world educational needs and optimising the potential for
better student outcomes, the integration of LLMs not only improves the accuracy, depth, and
contextual relevance of feedback but also fosters a more positive and engaging learning
environment.
1. Daas, M., & Ali, W. (2024). The Importance of E-Learning in Improving the Quality of Higher
Education (pp. 224–234). IGI Global. https://doi.& Saorg/10.4018/979-8-3693-3932-9.ch014
2. Sharma, H., Hashmi, I., Gupta, I., Maheshwari, L., & Jethwani, K. (2024). The Role of E-
learning. International Journal For Multidisciplinary Research, 6(6).
https://doi.org/10.36948/ijfmr.2024.v06i06.33333
3. Patil, D. U., & Chandankhede, S. A. (2022). E-Learning in Education. International Journal of
Advanced Research in Science, Communication and Technology, 32–35.
https://doi.org/10.48175/ijarsct-7407
4. Perišić, et al. (2018). Limitations of static learning style detection in adaptive e-learning
systems. Journal of Educational Computing Research, 56(4), 453-472.
5. Karagiannis & Satratzemi (2018). Challenges in dynamic adaptation of e-learning content
based on learner profiles. IEEE Transactions on Learning Technologies, 11(2), 245-257.
6. Ristić, et al. (2023). Limitations of current adaptive e-learning systems and future research
directions. International Journal of Artificial Intelligence in Education, 33(1), 1-20.
7. Wang, et al. (2022). NLP-based personalized feedback systems in e-learning. Computers &
Education, 181, 104446.
8. Yusupova, et al. (2024). AI-enhanced adaptive assessment in modern LMS environments.
Educational Technology Research and Development, 72(2), 341-359.
9. Ambrosio, J. (2019). AI techniques for adaptive quiz generation in Moodle. Journal of
Educational Technology, 35(2), 101–115.
10. Andhika, R., Wulandari, D., & Supriyadi, S. (2024). Adaptive learning systems in Moodle:
Review and future directions. International Journal of E-Learning, 19(1), 45–64.
11. Gamage, T., Perera, H., & Silva, D. (2022). The evolution of Moodle as a platform for
personalized learning. Computers & Education, 175, 104329.
12. Krahn, M., Perišić, D., & Satratzemi, M. (2023). Integration of learning styles in Moodle for
personalized education. Educational Technology & Society, 26(3), 48–62.
13. Limongelli, S., Sbarra, M., & Lorusso, A. (2011). Adaptive learning frameworks for Moodle
environments. Proceedings of the IEEE International Conference on Advanced Learning
Technologies, 291–295.
14. Perišić, D., Krahn, M., & Satratzemi, M. (2018). Challenges in implementing personalized
learning in Moodle: A review. Journal of Computer Assisted Learning, 34(5), 529–540.
15. Pardosi, D., Aziz, M., & Wicaksono, A. (2024). AI-driven feedback mechanisms in Moodle
quizzes: Enhancements and evaluation. International Journal of Artificial Intelligence in Education,
34(1), 78–94.
16. Ristić, M., Ristić, G., & Kovačević, D. (2023). Visual analytics for monitoring personalized
learning progress in Moodle. Journal of Learning Analytics, 10(1), 19–34.
17. Sayapina, E., Krahn, M., & Satratzemi, M. (2024). Detecting learning styles for adaptive
learning in Moodle: Methods and effectiveness. IEEE Transactions on Learning Technologies,
17(2), 112–124.
18. Wang, S., Zhang, L., & Yusupova, A. (2023). AI algorithms for adaptive quiz generation and
feedback in Moodle. Computers & Education, 189, 104528.
19. Yusupova, A., Wang, S., & Zhang, L. (2024). Personalized learning paths based on quiz
outcomes in Moodle: Design and evaluation. Journal of Educational Computing Research, 62(6),
752–774.
