Unlocking the Potential of Machine Learning, Deep Learning and Nlp in Automated Grading of Essays: A Comprehensive Review

Machine Learning mainly deals with the algorithms and models that can learn from training data and make predictions on unknown data without being explicitly programmed. Deep Learning is that subset of Machine Learning that deals with vast amount of data to create complex models. Natural Language Processing (NLP) works on language and speech data and understands how computers and human language interact while encompassing algorithms and models that allow computers to comprehend, decipher, understand, and also generate natural language. The concept of automatically grading essays, also known as Automated Essay Scoring (AES) or automated essay grading (AEG), involves assessing and evaluating essays or written responses by deploying various algorithms without human assistance. AEG has proved to be a significant milestone in the journey of evolution of artificial intelligence in educational arena, addressing the urgency for scalable, consistent, and timely assessment of written paragraphs, essays and compositions. This comprehensive study examines the evolution, methodologies, and recent scenario of automated essay grading systems, with special mention of ML, DL and NLP approaches. This research unveils the historical development from traditional rule-based systems through feature-engineering moving towards ML and DL models to contemporary transformer-based architectures and large language models emphasizing on the milestones of this process. This study systematically summarizes traditional ML techniques including support vector machines, random forests, and ensemble methods, alongside deep learning architectures such as Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM) networks, Convolutional Neural Networks (CNN), and attention-based transformers including BERT, GPT, and RoBERTa. This research review critically assesses all real-world applications and importance of Automated Essay Grading including its multi-faceted features. The results of this analysis s

View article page

Download PDF