International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 13 Issue: 02 | Feb 2026
p-ISSN: 2395-0072
www.irjet.net
INSTRUCTED – A LIGHT RAG POWERED PLATFORM ENHANCING LEARNING THROUGH TECHNOLOGY Malipatel Ranjith Reddy 1, K.Sai Snehith 2 , K.Koti Reddy 3, M.Abhigna Reddy 4 1234Department of Information Technology, TKR College of Engineering and Technology, Telangana, India
---------------------------------------------------------------------***--------------------------------------------------------------------generic, incomplete, or misaligned with prescribed text
Abstract - The rapid adoption of artificial intelligence in
books and academic syllabi. This limitation negatively affects students’ conceptual clarity and exam preparation. Additionally, teachers continue to face a substantial workload in manually organizing syllabus content, preparing assignments, and creating a question paper, which reduces the time available for effective teaching and student engagement.
education has significantly transformed digital learning platforms; however, most existing AI-based systems generate generic or unreliable responses that are not strictly aligned with academic syllabi or prescribed textbooks. This often leads to conceptual ambiguity among students and increases the manual workload for teachers in syllabus planning, assignment preparation, and evaluation. This paper proposes InstructEd, a Light RetrievalAugmented Generation (Light RAG) powered educational platform designed to deliver accurate, curriculum-aligned, and textbook-grounded academic assistance. The proposed system integrates Optical Character Recognition (OCR), optimized text chunking, MiniLM-based semantic embeddings, and FAISS vector similarity search to retrieve relevant content from authenticated learning materials before generating responses. By grounding every answer in retrieved textbook content, the system minimizes hallucinations while ensuring high accuracy and low response latency. The platform supports role-based access for teachers and students, enabling efficient content management and exam-oriented learning support. Experimental results demonstrate improved answer reliability, reduced teacher workload through automation, and enhanced learning effectiveness, indicating that Light RAG-based educational platforms provide a scalable and reliable solution for modern syllabus-driven learning environments.
To address these challenges, this paper proposes InstructEd, a Light RAG-based educational platform that ensures all responses are grounded in authenticated academic materials. By combining lightweight retrieval mechanisms with efficient generation models, the system delivers accurate, syllabus-aligned answers while reducing manual academic workload. The platform supports both students and teachers through structured content management and automated academic workflows.
1.1 Light RAG–Based Learning Platform
Light Retrieval-Augmented Generation (Light RAG) refers to an AI-driven learning model in which student queries are answered by retrieving information directly from authenticated academic sources rather than relying on generic language model knowledge. In educational environments, this approach enables students to receive accurate, syllabus-aligned explanations derived strictly from prescribed textbooks and learning materials.
Key Words: Light RAG, Retrieval-Augmented Generation, FAISS, MiniLM Embeddings, OCR, AI in Education
In Light RAG-based educational systems, retrieval mechanisms act as the core component that selects relevant textbook sections before response generation. These systems use semantic embeddings and vector similarity search to match student questions with the most appropriate academic content. By grounding responses in retrieved material, Light RAG ensures contextual accuracy, reduces hallucinations, and improves trust in AI-assisted learning.
1. INTRODUCTION The education sector is undergoing a significant transformation due to the rapid adoption of artificial intelligence and digital learning technologies. AI-driven platforms are increasingly used for online tutoring, content delivery, and academic assistance, enabling students to access learning resources beyond traditional classrooms. While these advancements improve accessibility, they also introduce challenges related to content accuracy, syllabus alignment, and academic reliability. Conventional AI-based educational systems often rely on generalized language models or unrestricted internet data. Such systems frequently generate responses that are
© 2026, IRJET
|
Impact Factor value: 8.315
Textbook-Grounded
In the proposed InstructEd platform, academic content is processed using Optical Character Recognition (OCR), structured chunking, and MiniLM-based embeddings. The retrieved content is then used to generate answers aligned with curriculum requirements. This approach improves learning reliability, enhances exam preparation,
|
ISO 9001:2008 Certified Journal
|
Page 46