Hey AI Can You Grade My Essay?: Automatic Essay Grading
Project Overview
The document explores the transformative role of generative AI in education, particularly focusing on Automatic Essay Grading (AEG). It highlights the development of a collaborative deep learning model that enhances essay assessment by evaluating both structural and semantic features, outperforming traditional grading systems. This innovative approach employs collaborative and transfer learning techniques, which not only improve grading accuracy but also significantly reduce the time and resources required for the grading process. The findings indicate that generative AI can facilitate a more efficient and reliable assessment landscape in education, ultimately benefiting educators and students alike by allowing for timely feedback and a more streamlined grading experience. The outcomes suggest that integrating generative AI into educational practices could reshape the way assessments are conducted, leading to enhanced learning experiences and academic performance.
Key Applications
Collaborative Deep Learning Network (CDLN)
Context: Educational context for grading essays written by students in grades 7 through 10.
Implementation: Implemented a model that combines multiple neural network architectures (CNN, RNN, LSTM) for collaborative learning.
Outcomes: Achieved an accuracy of 85.50% and improved performance over existing models.
Challenges: Holistic AEG can be difficult due to the need for human-graded essays for system evaluation.
Implementation Barriers
Technical Barrier
The challenge of holistic AEG where the system must evaluate the overall quality of an essay.
Proposed Solutions: Developing domain adaptation techniques and collaborative models to improve grading accuracy.
Project Team
Maisha Maliha
Researcher
Vishal Pramanik
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Maisha Maliha, Vishal Pramanik
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai