Best AI and ML Applications of the Week—April 05, 2018
Now on to other applications that have garnered attention since last Thursday:
“Wow, I can get information at the same time as my hearing peers!” –Joseph Adjei
College life for the deaf and hard of hearing just got a little easier at Rochester Institute of Technology, the world’s largest mainstream program for deaf and hard-of-hearing people. Microsoft is partnering with the university to pilot their AI communication technology in the classroom. The tech takes raw language, with all the glorious ums and ahs, and converts it into punctuated text that the university then projects up on a wall behind the teacher. This allows deaf students to access information at the exact same time as their hearing peers. Since the tech isn’t yet perfect and at times it mistranslates some key concepts, it works best when used in conjunction with an ASL translator. Watch it at work here:
Look at It from this Perspective . . .
One of the major concerns with AI, as described by Google’s Francois Chollet in his recent Medium article, is the creation and use by interested parties of detailed psychological profiles to subtly manipulate social media users in various ways, especially politically. While Francois offers his own insightful thoughts on solutions to this problem, another startup, Knowhere, is using AI to rewrite news stories from three distinct perspectives, liberal, impartial, and conservative, so users can read one story from all three angles to (hopefully) gain a broader perspective. The founder hopes it will help people breakout of their online echo chambers. Time will tell if this approach to journalism works.
With a lack of resources and a lack of personnel, the World Anti-Doping Agency (WADA) often hasn’t been able to peg guilty athletes for doping. Better-funded athletes can avoid getting caught for cheating by staying ahead of latest testing methods and regulations. But WADA is bringing out some big guns—a system that will more quickly sift all the electronic records and flag potential cheaters, helping the organization hone in on the right athletes to target for additional testing. The aim is to roll it out over the next few years.