Quantifying and Reducing Gender Stereotypes in Word Embeddings

Quantifying and Reducing Gender Stereotypes in Word Embeddings

Explore gender stereotypes in word embeddings and learn techniques to quantify and reduce bias in this hands-on tutorial from the FAT* 2018 conference. Dive into the basics of word embedding learning and applications, then gain practical experience writing programs to display and measure gender stereotypes in these widely-used natural language processing tools. Discover methods to mitigate bias and create fairer algorithmic decision-making processes. Work with iPython notebooks to explore real-world examples and complete exercises that reinforce concepts of fairness in machine learning and natural language processing.