Overcoming the Social Biases in Natural Language Processing Systems
Briefly