Paul allen’s ai research group unveils program that aims to shake up how we search scientific knowledge. give it a try. – the washington post

11437-none

At a morning meeting held every day at the Allen Institute for Artificial Intelligence in Seattle, engineers, researchers and other staff meet to update each other on their work. The meeting is led by senior software architect Oyvind Tafjord (far right, back to camera). The Allen Institute for Artificial Intelligence (nicknamed AI2) is a research organization funded by Microsoft co-founder and billionaire philanthropist Paul Allen.


(Stuart Isett/For The Washington Post)

When it comes to the body of scientific literature in existence, the world is suffering from information overload. There are now more than 100 million academic papers online and that number is estimated to be growing at a rate of nearly 5,000 new articles each day. According to one study, only half of those papers are being read by anyone other than the author and the editors of whatever journal published it.

Many researchers think the answers to a lot of fundamental questions about our lives, science, society and history are hidden in these articles and the connections they make, but that humans just don’t have the brain power to make sense of it.

A group of computer scientists funded by Microsoft co-founder and billionaire philanthropist Paul Allen is trying to change that through artificial intelligence.

On Monday, they took a first step toward that goal by unveiling a new AI program that will help you search the vast universe of scholarly knowledge in a new way. Known as Semantic Scholar and developed at the Allen Institute for Artificial Intelligence or AI2 in Seattle, it’s a publicly available “smart” search service for journal articles.

“There are no Renaissance men or women anymore,” Oren Etzioni, the head of AI2, said in an interview. “It’s impossible to keep up with everything we are discovering these days.”

For now, the program’s knowledge is limited to about 3 million computer science papers, but the team is quickly adding more areas of expertise. Neuroscience research will be available sometime in 2016, and other medical topics will soon follow. Etzioni explained that medical science is the priority for the project because “the medical example is so visceral.”

“When you see a specialist you know he or she is not up on your condition simply because they can’t be,” he said. “Even doctors in the emergency room are feverishly looking up stuff on their phone to get the latest information.”

Etzioni said the difference between Semantic Scholar and services, like Google Scholar or PubMed, is its ability to quickly highlight the most important papers and to tell you what the connections are between different papers and that is something that the researchers will continue to build on.

But the program’s ambitions go far beyond search. Etzioni said the ultimate goal is to use it as a foundation for a kind of artificial intelligence scientific assistant which can suggest new hypothesis or new directions for research.

“It’s going to help connect the dots and help scientists potentially find answers to something like cancer,” he said.

How soon will this happen in the future? “I think that the first scientific assistants will emerge in the next 10 years, and they are going to keep getting better and better,” Etzioni said. “We’re not talking about beyond the horizon. We’re going to see this very soon.”

You can try the new AI2 Semantic Scholar program here or at semanticscholar. org. There’s also a mobile version for both iOS and Android.

Read More:

Cutting sugar from kids’ diets appears to have a beneficial effect in just 10 days

Sleep study on modern-day hunter-gatherers dispels notion that we’re wired to need 8 hours

WHO says hot dogs, bacon cause cancer. Does this mean we should all be vegetarians?

There are roughly 480 other things the WHO says might cause cancer.

SPECIAL REPORT: Billionaire Paul Allen’s quest to build an artificial brain

For more health news, you can sign up for our weekly newsletter here .

Post Contributor Badge

This commenter is a Washington Post contributor. Post contributors aren’t staff, but may write articles or columns. In some cases, contributors are sources or experts quoted in a story.

More about badges | Request a badge

Comments our editors find particularly useful or relevant are displayed in Top Comments, as are comments by users with these badges: . Replies to those posts appear here, as well as posts by staff writers.

All comments are posted in the All Comments tab.

Comments our editors find particularly useful or relevant are displayed in Top Comments, as are comments by users with these badges: . Replies to those posts appear here, as well as posts by staff writers.