Datasests

• For the ELLIPSE and ASAP++ datasets we used in the paper, see .\dataset.

• Chatgpt_essays.txt in .\dataset\ELLIPSE is the essays of chatgpt_data.csv, which can be simply pasted as the input of ChatGPT

Rerequisites

Before starting the training, please ensure that you have prepared the necessary pretrained model.

• Download deberta-v3-base from HuggingFace https://huggingface.co/microsoft/deberta-v3-base and put the downloaded model in the ./pretrained_models/ directory.

Reproducing Results

To reproduce the results from the paper:

• Download the checkpoints from here ell_ckpt and place them in the ./ckpt directory. Run the following command to start inference on the ELLIPSE dataset:

bash run.sh -t ell -i false

Citation

If you find this code useful, please consider citing our paper:

@inproceedings{chen2024multi,
  title={A Multi-task Automated Assessment System for Essay Scoring},
  author={Chen, Shigeng and Lan, Yunshi and Yuan, Zheng},
  booktitle={International Conference on Artificial Intelligence in Education},
  pages={276--283},
  year={2024},
  organization={Springer},
  url={https://doi.org/10.1007/978-3-031-64299-9_22} 
}