ALPHA, according to retired Col. Gene Lee, is “the most aggressive, responsive, dynamic and credible AI I’ve seen to date.” Lee, according to the University of Cincinnati magazine that originally publicized the research, has yet to defeat ALPHA in simulated aerial combat.
+ ALSO ON NETWORK WORLD:Ultimate guide to Raspberry Pi OSes +
ALPHA, the brainchild of Nick Ernest, works via what’s called a Genetic Fuzzy Tree system, allowing the artificial intelligence to make complicated decisions based on inputs from many sources without using an excess of computing power.
The idea is to have the system making decisions more like humans do, rather than have it directly comparing variables and purely numeric data. A wide receiver in a football game, said Ernest, doesn’t think about the cornerback’s precise 40-yard-dash time or number of interceptions this year.
“That receiver – rather than standing still on the line of scrimmage before the play trying to remember all of the different specific statistics and what they mean individually and combined to how he should change his performance – would just consider the cornerback as ‘really good,’” he said.
ALPHA’s current fighter-ace-besting iteration is the product of automated development and “training” – different variations of the system are matched against each other and the most successful ones make up the code base for the next version. So, in other words, it teaches itself and it’s only going to get better. If it can learn to ride a motorcycle and play shirtless beach volleyball, the fighter pilot may become a thing of the past.
If you’ve ever had a hankering to get to behind-the-scenes grips with Amazon’s Alexa voice assistant service, here’s your chance: Hackster.io is running a contest this month to see who can combine the Raspberry Pi with Alexa in the most interesting way.
The Internet of Voice challenge (full details here) demands the use of the best voice user interface practices, full documentation of all code and hardware design, and a lot of creativity.
For those of you who don’t want to play the video: Raspberry Pi founder Eben Upton explains that there are two categories for the contest, which ends Aug. 1. One set of prizes will be given to the project that makes the best use of the Alexa Skills Kit (IoT control and so on), and another awarded to the better-known Alexa Voice Service. Get hacking!
With all due respect to thin client inventor Jeff McNaught, who gently rubbished the concept of Raspberry Pi-based thin clients in the last installment of this column, Viewsonic’s undercutting his argument a little bit by going to market with a thin client built around a Raspberry Pi 3 for around $100.
OK, it’s not quite the $50 price point that Dell’s McNaught mooted last time around, but that’s still pretty cheap for a thin client – prices for new, modern thin clients on Newegg start around $75, so, depending on features, $100 could be pretty competitive.