Recently we wrote about deep learning and its potential to revolutionize virtually every industry. One startup thinks that it can apply it to scientific research itself putting technological progress on an even faster development path. Meet ARIS.AI, the virtual assistant designed to scan through the entire scientific literature and categorize and display the relevant studies in any area in a easy to navigate structured way.

As every post-grad student is very well aware of, modern scientific research involves long hours of scrolling through the thousands of papers published in scientific journals around the world every year, while looking for just a handful of studies relevant to a particular topic. Now imagine this task to be accomplished in seconds with every possible connection and journal being examined. This would not only save researchers a lot of time, but also shield them from the "dangers of tunnel vision" as the founders of ARIS.AI put it. The company's virtual assistant promises to do just that. The A.I. is based on Deep Learning algorithms combining text recognition and categorization and neural networks trying to link the key concepts in any text to similar studies in the scientific literature. As the name suggests the machine is self-educating and is currently practising on 2000 or so TED talks. Since it is still a 'young' scientist, ARIS.AI is actually looking for human teacher to validate the algorithms. So if you ever wanted to participate in the creation of SkyNet, now you can do just that by joining their training program here. If not you can still use their tool and analyze any TED talk here. Bellow is a short presentation from Our Future Health 2016:

ARIS.AI was founded in 2015 as result of the Global Solutions Program at Singularity University. For those of you, who are not aware, Singularity University is a Silicon Valley think tank founded in 2008 focusing on exponential technologies with the potential to impact billions of people. The Global Solutions Program is a 10-week educational program/workshop designed to educate, empower and prepare participants from across the globe to use exponential technology to solve global challenges. The 2016 GSP takes place from June 18 to August 28.

ARIS.AI is a promising project with great potential to greatly increase research efficiency. Needless to say there are also a lot of challenges. I.E. most of the scientific literature is still not open source. Respectively any long-term solution will have to involve difficult negotiations with publishing networks over IP rights. We will definitely keep an eye on the company and keep you informed.

#DeepLearning #AI #Startups #ExponentialTechnologies #ARISAI #TheETechBlog

Recent Posts