Siddharth, Rishi
In this repo, we provide a full pipeline to train, combine, and analyze various bias mitigation methods (Bolukbasi et al., Manzini et al., Maudslay et al.). We also provide several metrics for comparison:
Word Embedding Association Test (WEAT) - Caliskan et al. (2017) Relative Norm Distance (RND) - Garg et al. (2018), which has two variations (cosine and euclidean) Mean Average Cosine (MAC) - Manzini et al. (2019)
python3 data-loader.py
debias.ipynb
pythonw evalbias.py
- Gensim, NumPy, SciPy