With its new free Semantic Scholar search engine, the Allen Institute for Artificial Intelligence hopes to provide researchers a way to sift through papers and begin to understand what those papers mean, Nature News reports.
Semantic Scholar relies on machine-reading and vision techniques to tease out papers' topics. While it currently is limited to some 3 million open-access computer science publications, Etzioni says that they hope to expand the engine into other fields.
"We're trying to get deep into the papers and be fast and clean and usable," Oren Etzioni, chief executive officer of AI2, says.
Nature News notes that there are already search engines that trawl through the scientific literature, most notably PubMed and Google Scholar. But, Semantic Scholar has a few innovative features, it notes, as it can gauge which of a paper's citations were actually influential, rather than included as background material.
Still, both Google Scholar and Semantic Scholar have their limitations — for instance, Google Scholar sometimes mistakes page numbers for publication year, and Sematic Scholar is stymied by paywalls.
"But we feel the tide is turning. More and more stuff is available somewhere," Etzioni says.