Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Hey AI Can You Grade My Essay?: Automatic Essay Grading

Maisha Maliha, Vishal Pramanik

TL;DR

This paper tackles holistic automatic essay grading (AEG) and the limitations of single-network systems. It introduces a Collaborative Deep Learning Network (CDLN) that fuses CNN, Recursive Neural Network (RvNN), and LSTM components with a dense output head, leveraging Word2Vec embeddings and transfer learning to jointly address grammatical/structural features and overall essay ideas. Empirical results on ASAP AES show CDLN achieving an average accuracy of $0.8550$, with PCC $0.7545$ and QWK $0.7036$, outperforming baselines including SVM and BERT, and demonstrating robustness to paraphrasing. The work highlights the value of collaborative, multi-network architectures for holistic scoring and suggests future directions involving more pretrained models and expanded domain adaptation for scalable educational assessment.

Abstract

Automatic essay grading (AEG) has attracted the the attention of the NLP community because of its applications to several educational applications, such as scoring essays, short answers, etc. AEG systems can save significant time and money when grading essays. In the existing works, the essays are graded where a single network is responsible for the whole process, which may be ineffective because a single network may not be able to learn all the features of a human-written essay. In this work, we have introduced a new model that outperforms the state-of-the-art models in the field of AEG. We have used the concept of collaborative and transfer learning, where one network will be responsible for checking the grammatical and structural features of the sentences of an essay while another network is responsible for scoring the overall idea present in the essay. These learnings are transferred to another network to score the essay. We also compared the performances of the different models mentioned in our work, and our proposed model has shown the highest accuracy of 85.50%.

Hey AI Can You Grade My Essay?: Automatic Essay Grading

TL;DR

This paper tackles holistic automatic essay grading (AEG) and the limitations of single-network systems. It introduces a Collaborative Deep Learning Network (CDLN) that fuses CNN, Recursive Neural Network (RvNN), and LSTM components with a dense output head, leveraging Word2Vec embeddings and transfer learning to jointly address grammatical/structural features and overall essay ideas. Empirical results on ASAP AES show CDLN achieving an average accuracy of , with PCC and QWK , outperforming baselines including SVM and BERT, and demonstrating robustness to paraphrasing. The work highlights the value of collaborative, multi-network architectures for holistic scoring and suggests future directions involving more pretrained models and expanded domain adaptation for scalable educational assessment.

Abstract

Automatic essay grading (AEG) has attracted the the attention of the NLP community because of its applications to several educational applications, such as scoring essays, short answers, etc. AEG systems can save significant time and money when grading essays. In the existing works, the essays are graded where a single network is responsible for the whole process, which may be ineffective because a single network may not be able to learn all the features of a human-written essay. In this work, we have introduced a new model that outperforms the state-of-the-art models in the field of AEG. We have used the concept of collaborative and transfer learning, where one network will be responsible for checking the grammatical and structural features of the sentences of an essay while another network is responsible for scoring the overall idea present in the essay. These learnings are transferred to another network to score the essay. We also compared the performances of the different models mentioned in our work, and our proposed model has shown the highest accuracy of 85.50%.

Paper Structure

This paper contains 23 sections, 8 equations, 3 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Difficulties in Automatic Essay Grading
  3. Related Works
  4. Classical Machine-Learning based Systems
  5. Deep-Learning based Systems
  6. Background and Technologies
  7. Recurrent Neural Networks
  8. Convolutional Neural Network
  9. Bidirectional Encoder Representations from Transformers (BERT)
  10. Dataset
  11. Techniques and Implementation
  12. SVM
  13. BERT
  14. Collaborative Deep Learning Network (CDLN)
  15. Convolutional Neural Network
  16. ...and 8 more sections

Figures (3)

  • Figure 1: The above images show how the RvNN works for a sentence. The word embeddings are fed to a neural network to get the phrase embedding, which is then passes into network to get the sentence embedding output vector. This vectors stores the grammatical and structural properties of the sentence
  • Figure 2: The architecture of our Collaborative Deep Learning Network
  • Figure 3: The image compares the average grade of the original essays and the essays after paraphrasing