Understanding Stereotype Within Text: An Application to Perceptions of Computer Scientists

In spite of the growing excitement and participation in the technology sector, there has still been great inequality and bias against certain groups. Many researchers suggest that the “stereotype” that surrounds computer scientists - commonly expressed as geeky, awkward, male, genius, and the like - contributes to this imbalance of representation across many different social demographics. One way to better understand the relationship between stereotyping and bias is to quantify these stereotypes using natural language processing techniques, so that we can somewhat “measure” how much descriptors for computer scientists project an unfair ideal. Using contributions from over 1500 participants in the United States who described what, in their mind, a stereotypical/anti-stereotypical computer scientist represents, we explore various techniques of translating language into meaningful numerical data with the hopes that understanding the nature of stereotype will eventually lead to minimizing that stereotype within computer science.

Speaker: Francesca Vera, Stanford