AI Finds 40,000 Scientists Without Wikipedia Entries


Wikipedia's books. / Photo by: VGrigas via Wikimedia Commons


AI startup Primer has recently developed an artificial intelligence tool called Quicksilver that has spotted 40,000 scientists that have no Wikipedia pages, reports magazine Popular Science.

Named after a historical novel by sci-fi author Neal Stephenson, Quicksilver was trained using 30,000 Wikipedia pages of prominent scientists. Then, the Primer team fed the system with 200,000 names, together with their employment information. These names belong to listed authors of scientific papers that focus on biomedical research and computer science. 

Primer’s director of science John Bohannon said, “We just fed it in, and we went home, and it ran,” referring to their input to the AI system. He added that the system has generated 40,000 new people overnight. 

Bohannon added that the process was not easy because the technology has to do a disambiguation. It is a part that makes sure that the system is not mixing up two people with the same name. It also has to read news sources in order to learn more about the scientists. Then, it builds the model of the person. 

The technology has identified the scientists, whose coverage of work is the same as those prominent scientists who have Wikipedia pages.

After sifting out the most cited figures, the AI provides a draft article about the scientists and their work. An example of this would be the AI-written article about a scientist named Teresa Woodruff. While Woodruff doesn’t have a Wikipedia entry for now, the AI system recognized her because she was one of the “Most Influential Persons” in 2013 by Time magazine. She “specializes in gynecology and obstetrics,” reads the AI-written content about Woodruff.