In the information age of rapid news sharing, iSchool professor Bei Yu has started a project about misinformation to better understand scientific research. Yu, whose research area is in applied Natural Language Processing (NLP), is taking an NLP approach to analyze exaggerated claims in science communications. Specifically, Yu and her team is building a computational model on the language of certainty, gathering and comparing science claims from prior research. Their findings will help the public understand how people describe similar topics in different ways. For instance, how news stories and social media communication might be different from what scientists are saying. “I think [our findings] can be training materials for science education, so the public can get more familiar with scientists’ language as a primary source,” Yu says.

Ultimately, her goal is to monitor the quality of science communications in society with this automated tool. Yu feels that original findings from scientists aren’t directly presented to the public, whether the findings are hard to understand or difficult to locate. Now that anyone can report science claims online, she finds that there’s more public information. For Yu, the challenge is to develop better methods of accessing accurate information.

Inspired by her past work, Yu wants to apply her experiences in opinion mining and sentiment analysis to this project. “I have this personal need of wanting some tool to help me do literature review, so the tool is something I want to have myself,” she says.

Although Yu’s research analyzes science communications, she is currently focused on the health domain. She works on medical literature mostly due to the availability of data from the National Institutes of Health (NIH). After reading biomedical research, Yu came across fake stories and misinformation about the health industry. She says that she was distraught after seeing friends and family find false health claims in social media.

“I can’t help but to speak up about it,” she says. “But I also felt that one person’s voice can’t influence much people, so I became interested in how misinformation in science gets circulated and what NLP can do to help.”

For this project, Yu is working with Jieke Zhang, a recent master’s student graduate, and PhD student Yingya Li, who is also working with Yu on CORA, a citation opinion retrieval and analysis tool.

Yu, who says she’s more of a “system development” person than a writer or activist, wants to influence the general public by developing this machine intelligence tool that monitors the quality of science communications. The purpose of her work, she says, is to warn others or remind them about the validity of health claims.