mennafateen / sss-vs-incml

sss-vs-incml

Introduction

This repository contains the code used for the experiments in the paper "In-Context Meta-Learning vs. Semantic Score-Based Similarity: A Comparative Study in Arabic Short Answer Grading" by Fateen et. al. (2023).

Requirements

To install requirements:

1. Create a conda environemnt with specified python version

conda create -n sss-vs-incml python=3.9

2.Install the required libraries by running

pip install -r requirements.txt

Code Organization

The code is organized as follows:

1. data/: contains the data used in the experiments.
2. sss/: contains the code for the SSS model. The files can be run in the order defined by the file names.
3. incml/: contains the code for the INCML model. The files can be run in the order defined by the file names.
4. prompt_classify.py: the code for the prompt classification model.

Data

The data used in the experiments was collected from the Arabic ASAG dataset We divided the data per prompt type and divided them into train/test sets. The data can be found in the data/csv folder.

Citation

If you use this code in your research, please cite the following paper:

@inproceedings{fateen2023incontext,
  title={In-Context Meta-Learning vs. Semantic Score-Based Similarity: A Comparative Study in Arabic Short Answer Grading},
  author={Fateen, Menna and Mine, Tsunenori},
  year={2023}
}