How optimal transport can tackle gender biases in multi-class neural-network classifiers for job recommendations?

Open in new window