Date Posted: 8/22/2016

CS PhD student Liye Fu, IS professor Cristian Danescu-Niculescu-Mizil, and CS professor Lillian Lee received the Best Paper Award at the 2016 IJCAI "Natural language processing meets journalism" workshop for the paper "Tie-breaker: Using language models to quantify gender bias in sports journalism."

The work is motivated by suspicions that the questions posed by interviewers to female athletes differ from those posed to male athletes, a hypothesized tendency satirized by a CoverTheAthlete video.
The paper introduces a technique based on probabilistic language models to quantitatively measure differences in the kinds of questions asked. This enables investigation into whether questions to male athletes are more game-related than those to female athletes, in a way that (1) avoids requiring a subjectively determined list of game-related terms, and (2) runs deeper than merely finding individual words whose distributions in to-male vs. to-female questions differ. The experiments focus on tennis post-match interviews, since tennis is one of the few sports in which there is relative gender parity in the (significant) amounts of attention and prize money male and female players receive.

Also at the workshop, IS professor Mor Naaman gave an invited talk on "The past and future of systems for current events."

Coverage by the Pacific Standard
Workshop webpage