Skip to content

Active learning for Question Difficulty Estimation (QDE)

License

Notifications You must be signed in to change notification settings

arthur-thuy/qde-active-learning

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

27 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Active Learning to Guide Labeling Efforts for Question Difficulty Estimation

This repository contains the code to reproduce the results of the paper Active Learning to Guide Labeling Efforts for Question Difficulty Estimation.

The paper is available here as an arXiv preprint.

Workflow

Download the RACE (here) and RACE-c (here) dataset. Together, these datasets form the RACE++ dataset. Save the datasets in the data/raw/ folder.

Prepare the datasets by running:

python data_preparation.py
python create_small_dev.py race_pp --size 1000

Create the environment from the environment.yml file and activate it:

conda env create -f environment.yml
conda activate qdet_active

The configuration files are located in the folder src/config/race_pp/. Run the experiments with:

python main.py race_pp

Finally, inspect the results by running the analysis.ipynb notebook.

Cite as

If you use this code in your workflow or scientific publication, please cite the corresponding paper:

@article{thuy2024active,
  title={Active Learning to Guide Labeling Efforts for Question Difficulty Estimation},
  author={Thuy, Arthur and Loginova, Ekaterina and Benoit, Dries F},
  journal={arXiv preprint arXiv:2409.09258},
  year={2024},
  doi={10.48550/arXiv.2409.09258}
}

About

Active learning for Question Difficulty Estimation (QDE)

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published