Integrated Framework for Bloom’s based Question and Short Answer Generation using Transformer

This research presents an integrated Natural Language Processing (NLP) framework that combines Automatic Question Generation (AQG) based on Bloom’s Taxonomy with a Context-Aware Short Answer Generation (SAG) system. The AQG module utilizes transformer models, specifically T5, along with NLP techniques like dependency parsing and semantic role labeling to generate questions across various cognitive levels, aiming to reduce educators’ workload while enhancing assessment quality. The SAG component employs models such as RoBERTa and Long former to deliver concise, accurate responses using advanced techniques like tokenization, named entity recognition, and contextual summarization. Evaluation through BLEU scores and human feedback shows high accuracy for lower Bloom levels (70-80%) and moderate accuracy for higher levels (30-40%). The combined system supports applications in education, customer service, healthcare, and e-commerce, offering scalable and intelligent solutions for automated learning and assessment.