Revolutionizing Academic Evaluation: Bloom's Taxonomy Meets Deep Learning and NLP

Teaching and learning are deeply interconnected, forming a dynamic, two-way process that can be transformed through innovative approaches. Learning happens through various channels, such as the environment, people, and personal interpretations, and is assessed through different methods. In Indian higher education, especially in the engineering domain, Bloom's taxonomy is widely used to evaluate students' cognitive abilities. With the advent of artificial intelligence, traditional assessment methods are gradually being replaced by automated and advanced techniques. This research explores the cutting-edge intersection of Bloom's Taxonomy, Deep Neural Network (DNN) and Natural Language Processing (NLP), pioneering a new wave in the evaluation specifically tailored for higher education students. This study discovers how Bloom's taxonomy is integrated with advanced AI models like Bidirectional Encoder Representations from Transformers (BERT) for contextual understanding and deep neural networks (DNN) such as Long Short-Term Memory (LSTM) and Bidirectional Gated Recurrent Unit (Bi-GRU) for sequential dependency analysis. The research focuses on AI-based models for academic question classification, presenting a hybrid methodology that bridges traditional educational frameworks and cutting-edge AI tools. By analyzing and comparing the performance of four DNN models, it highlights their potential to enhance competency-based assessments in higher education. This work envisions a future of personalized and data-driven educational support systems that advance learning evaluation.