Adversarial Scrubber (AdS)

This repository presents the implementation of EMNLP 2021 paper: "Adversarial Scrubbing of Demographic Information for Text Classification"

Somnath Basu Roy Chowdhury, Sayan Ghosh, Yiyuan Li, Junier Oliva, Shashank Srivastava and Snigdha Chaturvedi.

Pre-requisities

python 3.8

# create a clean conda env
conda create -n adv-scrubber python==3.8 
source activate adv-scrubber

pip install -r requirements.txt

Preparing the data

The dialogue datasets from the paper Multi-dimensional gender bias classification. We use the huggingface library (code present in src/prepare_data.py) to retrieve the datasets and use the same split.

DIAL and PAN16 datasets were generated using this open-source project.

The Biographies corpus was retrieved from this project.

Please cite the respective paper if you're using the any of the above data.

Demo for the project is provided for Wizard dataset.
Run the following command to prepare the data in the format required in our project.

cd src/
python prepare_data.py --dataset wizard  --save_path ../data/wizard/ 

Running AdS

Execute the following commands to run our AdS.

cd src/
python train.py \
--dataset wizard \
--MODEL bert-base-uncased \
--max_len 32 \
--embedding_size 768 \
--device cuda:0 \
--model_save_path ../model/wizard/ \
--results_save_path ../results/wizard/ \
--save_path data/wizard \
--epochs 3 \
--lambda_1 10 \
--lambda_2 0

Evaluating AdS

Evaluate the representations formed by the model by running the following command

cd src/
python evaluate.py \
    --dataset wizard

Citation

@inproceedings{basu-roy-chowdhury-etal-2021-adversarial,
    title = "Adversarial Scrubbing of Demographic Information for Text Classification",
    author = "Basu Roy Chowdhury, Somnath  and
      Ghosh, Sayan  and
      Li, Yiyuan  and
      Oliva, Junier  and
      Srivastava, Shashank  and
      Chaturvedi, Snigdha",
    booktitle = "Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing",
    month = nov,
    year = "2021",
    address = "Online and Punta Cana, Dominican Republic",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2021.emnlp-main.43",
    doi = "10.18653/v1/2021.emnlp-main.43",
    pages = "550--562",
}