The Allen Institute for Artificial Intelligence, or AI2, launched a new search engine for scientists on Monday, the latest project to come out of Paul Allen’s growing Seattle-based nonprofit.
The idea behind Semantic Scholar is to give researchers more powerful tools to scour through the millions of academic papers available online.
Right now, most scientists use Google Scholar. AI2 CEO Oren Etzioni said that’s a great service — but it was built about 10 years ago.
“Our goal is to raise the bar,” Etzioni said. “It would be foolhardy to take on Google, actually. … It’s part of our mission of AI for the common good. We’re very interested in helping scientists, not because that’s a good way to make money — it ain’t. Because it’s a good way to make the world a better place for all of us.”
Etzioni added that the technology also has the potential to advance the way companies like Google use AI to power their own search engines.
Semantic Scholar is powered by set of sophisticated algorithms. Instead of matching keywords like most search engines, the free service looks inside the articles for more nuanced context. It uses natural language understanding and computer vision to figure out things like when a paper was published or at which conference it was presented. Semantic Scholar then lets researchers filters results to find exactly what they’re looking for.
Etzioni recently showed GeekWire how it works on his iPhone. Semantic Scholar pulled up hundreds of results when he searched for a particular scientist’s name. But then he was able to tell it he was only interested in recent articles, relating to a certain topic, that were presented at a certain conference. That narrowed the list down to just one result.
“What we’re doing is using AI to take that search to the next level. To go beyond keyword searches and allow you to cut through the clutter,” Etzioni said.
A small group of researchers has been quietly working on the search engine for about a year, but Etzioni said AI2 is now looking to double the size of the Semantic Scholar team and add about 15 workers across the entire organization.
Semantic Scholar is just one of several projects AI2’s 43 employees are working on as the group tries to push AI technology to greater heights. Other initiatives include building machines capable of taking the SATs and launching a AI-focused startup incubator.