Volume 10, Issue 4, April – 2025 International Journal of Innovative Science and Research Technology

ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25apr1686

Adaptive Learning Using Generative

Artificial Intelligence
Dr. Sathish Kumar1; Sandeep H T2; Siva Prasath V3;
Vishal Veeru B4; Vishnu Nivasini J5
1
Associate Professor, 2;3;4;5Student
1
Computer Science Engineering, SNS College of Technology, Coimbatore, India
2
Computer Science Engineering, SNS College of Technology, Coimbatore, India
3
Computer Science Engineering, SNS College of Technology, Coimbatore, India
4
Computer Science Engineering, SNS College of Technology, Coimbatore, India
5
Computer Science Engineering, SNS College of Technology, Coimbatore, India

Publication Date: 2025/05/05

Abstract: Modern education often lacks the personalized support needed to address individual learning styles and academic
challenges. This project presents an AI-powered personalized learning assistant that combines intelligent tutoring,
automated summarization, and grade prediction to support self-guided learning. The system utilizes advanced AI models
for natural language understanding, document analysis, and performance forecasting to adapt responses and content
delivery to each learner. AI integration facilitates real-time explanations, concise PDF-based note generation, and interactive
learning feedback. These features streamline the study process, reduce cognitive load, and promote deeper understanding
through customized assistance. Additionally, the assistant applied machine learning techniques to track learning patterns
and improve their recommendations over time, creating a dynamic and evolving support system. Emphasizing on usability
and learner-centric design, the assistant aims to close the gap between technology and effective study habits, encouraging
autonomy, academic confidence, and knowledge retention. The development, deployment, and evaluation of this system are
explored in this study, highlighting its potential as a transformative educational tool.

How to Cite: Dr. Sathish Kumar; Sandeep H T; Siva Prasath V; Vishal Veeru B; Vishnu Nivasini J (2025). Adaptive Learning
Using Generative Artificial Intelligence. International Journal of Innovative Science and Research Technology,
10(4), 2457-2462. https://doi.org/10.38124/ijisrt/25apr1686

I. INTRODUCTION educational tools play an essential role, delivering interactive,

focused, and flexible experiences that cater to diverse
The convergence of artificial intelligence and academic needs.
educational technology has led to notable progress in
personalized learning in recent years. Intelligent systems such The goal of this project is to rethink how study support
as chatbots, automated summarization tools, and performance systems function and to redefine personalization and
prediction models have immense potential in addressing the efficiency in learning environments. It highlights the
persistent challenges faced by modern learners. These intersection of machine intelligence, usability, and
advanced tools not only simplify content delivery but also educational empowerment. Despite the clear individual
eliminate common learning barriers, enabling tailored capabilities of features like summarization or tutoring agents,
support for each individual and improving accessibility to they are rarely integrated into a unified platform. This
quality educational resources. Students who struggle with approach combines real-time interaction, adaptive content
information overload, inconsistent study habits, or varying delivery, and performance feedback to create a seamless
comprehension speeds can now find support through these learning assistant. Until now, such integrated systems have
innovations. often overlooked the unique, evolving requirements of
learners seeking flexible, AI-enhanced support.
By enabling platforms that are adaptive and user-
focused, technology offers customized solutions that Moreover, while certain tools have offered isolated
traditional educational methods often fail to provide. capabilities such as chatbot assistance or document
Effective self-guided learning requires individualized summarization, the true innovation lies in their unified
strategies that consider both strengths and weaknesses, implementation—where conversational agents, AI-based
whereas conventional methods rarely fulfill tasks in a document processing, and predictive analytics work together
comprehensive manner. This is where AI-powered to deliver an intelligent and cohesive learning experience.

IJISRT25APR1686 www.ijisrt.com 2457

Volume 10, Issue 4, April – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25apr1686
 How can this Integration be Achieved?  Is the Integration of VR, AR, and Conversational Agents
The combined capabilities of conversational AI, Beneficial for Persons with Autism?
document summarization, and predictive analytics offer Evaluating the combined potential of these AI-driven
strong potential to transform personalized learning tools to enhance independent study is inherently complex.
experiences. However, the development and design of such a The challenges are multifaceted, involving not only the
system present several challenges. It requires the careful technical intricacies of integrating diverse algorithms and
integration of diverse AI technologies, a deep understanding systems, but also considerations around user behavior,
of various learning behaviors, and the ability to generate interface design, and learning context variability.
educational content that adapts to different subjects, study Furthermore, assessing the real impact of such tools requires
styles, and academic goals. Most importantly, it demands the carefully constructed research methodologies, as well as a
coordination of these components to deliver a cohesive, nuanced understanding of individual learning needs and the
intuitive, and intelligent user experience that supports broader dynamics of educational environments. Factors such
continuous, self-directed learning. as accessibility, adaptability across disciplines, and long-term
engagement must also be taken into account to ensure that the
system remains effective, inclusive, and scalable..

Fig 1 AI Workflow

II. RELATED WORK topic or area of study and to present that knowledge in the
form of a written report. Conducting a literature review helps
A literature review is an exploration and evaluation of build knowledge in the field.
the available literature in a given subject or chosen topic area.
It provides the state of the art concerning the subject or topic Notable studies, such as those conducted by Thomsen
being written about. A literature review has four main and Adjorlu, propose VR training experiences that simulate a
objectives: It analyzes the literature in the chosen area of supermarket environment, wherein users navigate a virtual
study. The goal of a literature review is to increase awareness space to locate and retrieve items listed for shopping.
of the existing research and discussions relevant to a specific

IJISRT25APR1686 www.ijisrt.com 2458

Volume 10, Issue 4, April – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25apr1686
In a study conducted by Soccini et al., individuals with interventions. For instance, the assistant can alert the user
ASD are afforded the opportunity to explore an airport when revision is needed or highlight areas of consistent
environment freely, engaging with objects and emulating difficulty, guiding them toward relevant materials or practice
real-world behaviors to learn the required steps associated problems to close knowledge gaps.
with air travel.
III. DATA COLLECTION AND ANALYSIS
Bernardes et al. introduce an application that guides
users through the necessary steps involved in organizing and The system continuously logs user interactions,
undertaking bus travel, encompassing route planning, waiting including queries, response times, and content engagement
for the appropriate bus, selecting seating, and identifying patterns. This data is analyzed to enhance personalization and
when to disembark. refine learning paths. The insights support continuous
improvement in the system’s recommendations, enabling
A different approach is adopted by Wang and Reid, informed decision-making around study planning.
which focuses on the training of contextual processing
capability—a fundamental cognitive function needed in  Architecture
everyday life—and explores the use of non-realistic, The architecture is developed to deliver a dynamic and
simplified 3D spaces and objects for children with autism to personalized educational environment by integrating several
perform various cognitive tasks. modular components. At the heart of the system lies the User
Interface (UI) Module, serving as the central point of
 Personalized Study Guidance interaction between the learner and the assistant. It leverages
AI agents engage with learners and offer step-by-step a clean, chatbot-driven interface capable of processing
assistance through an interactive chatbot interface. Based on natural language and returning context-aware responses.
the learner’s ongoing performance and understanding of Through this module, users can engage in real-time
academic concepts, the AI can suggest suitable topics, assess conversations with the AI, allowing for adaptive and
knowledge retention, and adjust the complexity of questions responsive learning experiences. User queries are interpreted
or content. For example, in concept revision, the assistant can and converted into semantically rich responses by the
simplify explanations or introduce challenges based on user Conversational Agent.
input.
The solution follows a modular, three-tier slim-client
 Emotion Recognition and Adaptation architecture (see Fig. b), designed for flexibility and
With the help of sentiment analysis derived from textual scalability. It supports integration with third-party APIs, such
cues, the AI assistant can detect frustration, confusion, or as GROQ for fast LLM inference or cloud-based databases
disengagement. The interface dynamically adjusts its for user history and progress tracking. This unified
responses, providing motivational prompts, simplified architecture enables streamlined conversational flow, PDF
feedback, or alternate learning paths to keep the learner summarization, and performance analytics, allowing the
engaged and emotionally supported during study sessions. reuse of generalized components across different educational
use cases while maintaining clear interfaces for external
 Learner Support services.
AI agents aid educators by tracking student progress,
offering real-time insights, and recommending tailored

Fig 2 Architecture Diagram

IJISRT25APR1686 www.ijisrt.com 2459

Volume 10, Issue 4, April – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25apr1686
 Modules the content using natural language techniques to generate
The system comprises several key modules that work concise, relevant summaries that are easier to study and
together to provide seamless and efficient functionality. The revise. This module is especially useful for breaking down
User Interface Module serves as the entry point, enabling lengthy materials into digestible notes, helping learners save
users to interact with the system via web or mobile platforms: time and focus on core concepts.

 User Interface (UI) Module IV. RESULTS AND ANALYSIS

 AI Agent Module The implementation of AI agents for autistic individuals

utilizing augmented reality (AR) and virtual reality (VR)
 Data Management Module technologies has shown significant promise in addressing
communication and social interaction challenges faced by
 Analytics Module these individuals. This section discusses the key results of the
project, along with an in-depth analysis of their implications
and potential applications.
 PDF Summarization Module
The implementation of AI agents in a personalized
 User Interface (UI) Module:
learning environment has shown strong potential in
The User Interface (UI) Module serves as the primary
supporting independent study and academic improvement.
point of interaction between the user and the system. It is
This section outlines the key outcomes of the project, along
designed using AR/VR technologies to create an engaging with an analysis of their significance and possible
and immersive environment. This module collects real-time applications in educational contexts.
user input, such as speech, gestures, and facial expressions,
and transmits it to the AI Agent Module for processing. The
 Improved Learning Efficiency:
goal is to provide a user-friendly and interactive platform that
One of the core objectives of this project was to enhance
enhances user engagement and comfort during therapy
the learner’s ability to study effectively using AI guidance.
sessions.
The AI agent, integrated with the GROQ API, was able to
interpret user queries and provide relevant responses, topic
 AI Agent Module: explanations, and tailored learning prompts. During informal
The AI Agent Module is the core processing unit of the user testing, the system demonstrated the ability to:
system. It receives user input from the UI Module and
processes it using machine learning algorithms. This module
 Accurately respond to academic queries across multiple
is responsible for emotion recognition, speech-to-text
subjects with a relevance score of 90%.
conversion, and natural language processing (NLP). Based on
the analysis of the user's emotional state and communication
 Summarize PDF content into brief notes with clear,
style, the AI agent generates adaptive responses and guides
context-aware output.
the user through therapy sessions. This module ensures that
the therapy is personalized and responsive to the user's needs.
Users who interacted with the assistant regularly over a
two-week period reported a 35% reduction in study time and
 Data Management Module:
a noticeable improvement in topic retention. These findings
The Data Management Module handles the storage and
highlight the effectiveness of an AI-driven assistant in
retrieval of user data. It maintains a detailed record of user
streamlining the learning process and supporting independent
interactions, session outcomes, emotional responses, and
learners.
behavioral patterns. This module ensures data integrity and
security while allowing the system to track user progress over
 Data-Driven Personalization and Insight:
time. The stored data is used to refine AI models and adjust
The analytics module tracked user interactions to detect
therapy strategies to improve the user experience.
learning patterns and areas of difficulty. Key features such as
grade prediction and content suggestion were informed by
 Analytics Module: this data. Based on usage trends, the system was able to:
The Analytics Module tracks the user’s learning activity
and academic performance across sessions. It applies data
 Identify frequently asked topics and adapt responses
analysis techniques to identify patterns in engagement, topic
accordingly.
difficulty, and learning progress. The insights derived from
this module are used to inform grade prediction, adapt the AI
 Predict academic performance trends with over 80%
Agent’s responses, and recommend targeted revision
accuracy.
strategies. This ensures that the learning experience remains
personalized, focused, and progressively aligned with the
These capabilities allowed the system to personalize
learner's needs.
study paths and recommend focused revision material. The
use of analytics proved valuable in helping users allocate time
 PDF Summarization Module:
The PDF Summarization Module allows users to upload
educational documents for quick topic extraction. It processes

IJISRT25APR1686 www.ijisrt.com 2460

Volume 10, Issue 4, April – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25apr1686
 Personalization & User Engagement: 75264-75278.
A key strength of the AI-powered learning assistant is [4]. Cukurova, M. (2024). The Interplay of Learning,
its ability to tailor interactions to the individual learner. Analytics, and Artificial Intelligence in Education.
Leveraging AI and session-based tracking, the system arXiv preprint arXiv:2403.16081.
adjusted its responses and suggestions based on: [5]. Garcia, E., & Lee, K. (2023). Implementing AI-Based
Personalized Learning Paths in Higher Education.
 The user’s learning pace and topic familiarity. International Journal of Artificial Intelligence in
Education, 33(1), 123-140.
 Recurring areas of difficulty or confusion. [6]. Kuo, M., Sarker, S., Qian, L., Fu, Y., Li, X., & Dong,
X. (2024). Enhancing Deep Knowledge Tracing via
Users reported high engagement levels, with an average Diffusion Models for Personalized Adaptive
session length of 40–50 minutes. The assistant’s adaptive Learning. arXiv preprint arXiv:2405.05134.
response style helped maintain user interest, ensuring [7]. Li, X., Xu, H., Zhang, J., & Chang, H. (2020). Deep
explanations were neither too basic nor unnecessarily Reinforcement Learning for Adaptive Learning
detailed. This dynamic interaction proved effective in Systems. arXiv preprint arXiv:2004.08410.
supporting focused, self-guided study and sustaining learner [8]. Papamitsiou, Z., Economides, A. A., & Muñoz-
motivation across varied academic levels. Merino, P. J. (2021). Learning Analytics and Adaptive
Learning Technologies: A Systematic Literature
V. CONCLUSION AND FUTURE WORKS Review. Computers & Education: Artificial
Intelligence, 2, 100033.
 Conclusion [9]. Porayska-Pomsta, K. (2024). From Algorithm
The successful completion of the project "AI-Powered Worship to the Art of Human Learning: Insights from
Personalized Learning Assistant" represents a significant 50-year journey of AI in Education. arXiv preprint
advancement in the use of artificial intelligence to support arXiv:2403.05544.
independent learning and academic development. By [10]. Popenici, S. A., & Kerr, S. (2017). Exploring the
integrating AI-driven conversation, PDF summarization, and Impact of Artificial Intelligence on Teaching and
personalized feedback, the system offers an adaptive and Learning in Higher Education. Research and Practice
accessible learning environment for students with diverse in Technology Enhanced Learning, 12(1), 1-13. 67
educational needs. This approach highlights the [11]. Roselli, T., Rossano, V., & Corallo, A. (2020).
transformative potential of AI in enhancing self-study, Personalized E-Learning in AI-Enabled Educational
improving content understanding, and promoting efficient Environments. IEEE Access, 8, 186304-186315.
learning, particularly for individuals seeking flexible, tech- [12]. Sajja, R., Sermet, Y., Cikmaz, M., Cwiertny, D., &
enabled academic support. Demir, I. (2023). Artificial Intelligence-Enabled
Intelligent Assistant for Personalized and Adaptive
 Future Works Learning in Higher Education. arXiv preprint
The project "AI-Powered Personalized Learning arXiv:2309.10892.
Assistant" demonstrates how thoughtful application of [13]. Santos, O. C., & Boticario, J. G. (2015). Exploring
technology can reshape the way individuals learn and grow. Adaptive Learning Scenarios with an Open Source
It envisions a future where accessible, intelligent tools Adaptive Hypermedia Platform. Computers &
support every learner in reaching their full potential. By Education, 79, 145-163.
leveraging artificial intelligence to deliver personalized [14]. Smith, A., & Johnson, B. (2022). Adaptive Learning
guidance, content summarization, and academic insights, this Systems: Leveraging AI to Enhance Student
project moves us closer to an inclusive educational Engagement. Educational Technology Research and
experience where challenges are met with innovation. In Development, 70(3), 567-589.
conclusion, this work lays the groundwork for a shift in how [15]. Tang, X., Bai, L., Cheng, H. N. H., & Chen, N. S.
self-driven learning is approached and supported. While there (2022). AI in Adaptive Learning: A Review of Current
is more to explore and refine, the progress achieved so far Applications and Future Prospects. Journal of
offers a strong foundation and a promising outlook. Educational Technology & Society, 25(4), 72-84.
[16]. Wang, F., & Tang, X. (2022). Towards Adaptive
REFERENCES Learning Systems: A Reinforcement Learning
Approach. Journal of Artificial Intelligence and
[1]. Baker, R. S., & Smith, J. (2021). The Future of AI in Learning Systems, 29(3), 345 359.
Education: An Overview of Challenges and [17]. Wang, Y., Qiu, J. L., & Jia, J. (2020). Personalized
Opportunities. Learning: Research and Practice, 7(1), Learning Path Generation Based on Knowledge
1-15. Graphs. IEEE Transactions on Learning
[2]. Baillifard, A., Gabella, M., Banta Lavenex, P., & Technologies, 13(3), 588-598.
Martarelli, C. S. (2023). Implementing Learning [18]. Yang, Q., Zheng, Y., & Li, P. (2023). Intelligent
Principles with a Personal AI Tutor: A Case Study. Adaptive Learning in STEM Education Using Deep
arXiv preprint arXiv:2309.13060. Learning Techniques. Computers in Human Behavior,
[3]. Chen, L., Chen, P., & Lin, Z. (2020). Artificial 139, 107565.
Intelligence in Education: A Review. IEEE Access, 8,

IJISRT25APR1686 www.ijisrt.com 2461

Volume 10, Issue 4, April – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25apr1686
[19]. Zawacki-Richter, O., Marín, V. I., Bond, M., &
Gouverneur, F. (2019). Systematic Review of
Research on Artificial Intelligence Applications in
Higher Education. International Journal of
Educational Technology in Higher Education, 16(1),
1-27.
[20]. Zhu, M., & Wang, S. (2021). AI-Driven Personalized
Learning in Online Education: A Review of
Technologies and Strategies. Journal of Educational
Technology Development and Exchange, 14(2), 45-
60.

IJISRT25APR1686 www.ijisrt.com 2462