We wake up this morning to the news that Google’s deep-search neural network project called AlphaGo has beaten the second ranked world Go master (who happens to be a human being). This is the first of five matches between the two adversaries that will play out this week.
On one hand, this is a sign of maturing technology. It has been almost twenty years since Deep Blue beat Gary Kasparov, the reigning chess world champion at the time. Although there are still four games to play against Lee Sedol, it was recently reported that AlphaGo beat European Go champion Fan Hui in five games straight. Go is generally considered a more difficult game for machine minds to play than chess. This is because Go has a much larger pool of possible moves at any given time.
Does This Matter?
Okay, the news part of this event has been covered: machine beats man. Does it matter? Will this affect your life and how? We want to hear what you think in the comments below. But I’m going to keep going with some of my thoughts on the topic.
Let’s look first at what AlphaGo did to win. At its core, the game of Go is won by figuring out where your opponent will likely make a low-percentage move and then capitalizing on that choice. Know Your Enemy has been a tenet of strategy for a few millennia now and it holds true in the digital age. In addition to the rules of the game, AlphaGo was fed a healthy diet of 30 million positions from expert games. This builds behavior recognition into the system. Not just what moves can be made, but what moves are most likely to be made.
DeepMind, the company behind AlphaGo which was acquired by Google in 2014, has published a paper in Nature about their approach. They were even nice enough to let us read without dealing with a paywall. The secret sauce is the learning process which at its core tries to mimic how living entities learn: observe repetitively while assigning values to outcomes. This is key as it leads past “intellect”, to “intelligence” (the “I” in AI that everyone seems to be waiting for). But this is a bastardized version of “intelligence”. AlphaGo is able to recognize and predict behavior, then make choices that lead to a desired outcome. This is more than intellect as it does value the purpose of an opponent’s decisions. But it falls short of intelligence as AlphaGo doesn’t consciously understand the purpose it has detected. In my mind this is exactly what we need. Truly successful machine learning will be able to make sense out of sometimes irrational input.
The paper from Nature doesn’t go into details about Go, but it explains the approach of the learning system applied to Atari 2600. The algorithm was given 210×160 color video at 60Hz as an input and then told it could use a joystick with one button. From there it taught itself to play 49 games. It was not told the purpose or the rules of the games, but it was given examples of scores from human performance and rewarded for its own quality performances. The chart above shows that it learned to play 29 of them at or above human skill levels.
A few summers ago, Google and IEEE announced a one million dollar prize to build the most efficient and compact DC to AC inverter. It was called the Little Box Challenge, with the goal of a 2kW inverter with a power density greater than 50 Watts per cubic inch.
To put this goal into perspective, the DC inverter that would plug into a cigarette lighter in your car has a power density of about 1 or 2 Watts per cubic inch. Very expensive inverters meant for solar installations have a power density of about 5 Watts per cubic inch. This competition aimed to build an inverter with ten times the power density of what is available today.
Now, the results are in, and the results are extremely surprising. The best entry didn’t just meet the goal of 50 W/in³, it blew the goal out of the water.
The winning entry (PDF) comes from CE+T Power, and comes in a package with a volume of 13.77 in³. That’s a power density of 143 W/in³ for a unit you can hold in the palm of your hand. The biggest innovations come from the use of GaN transistors and an incredible thermal management solution.
You can find the funniest things in public government documents. There’s always ample evidence your local congress critter is working against the interests of their constituency, nation, and industry controlled by the commission they’re chairperson of. Rarely, though, do you find something surprising, and rarer still does it portend some sort of experiments conducted by Google at a spaceport in New Mexico.
In a publication released last week, Google asked the FCC to treat some information relating to radio experiments as confidential. These experiments involve highly directional and therefore high power transmissions at 2.5 GHz, 5.8GHz, 24GHz, 71-76GHz, and 81-86GHz. These experiments will take place at Spaceport America, a 12,000 foot runway in the middle of New Mexico occasionally used by SpaceX, Virgin Galactic, and now Google.
For the most part, this document only tells the FCC that Google won’t be causing harmful interference in their radio experiments. There few other details, save for what bands and transmitters Google will be using and an experimental radio license call sign (WI9XZE) that doesn’t show up in the FCC database.
Of the few details listed in the documents, one thing does pop out as exceptionally odd: a 70-80 GHz transmitter with an effective radiated power (ERP) 96,411 W. That’s close enough to 100 kilowatts to call it as such. This is the maximum effective radiated power of the highest power FM stations in the US, but radio stations are omnidirectional, whereas Google is using very high gain antennas with a beam width of less than half a degree. The actual power output of this transmitter is a mere half watt.
The best guess for what Google is doing out in the New Mexico desert is Project Skybender, a project to use millimeter waves to bring faster Internet to everyone. There aren’t many details, but there is a lot of speculation ranging from application in low Earth orbit to something with Google Loon.
Quick. What’s the difference in conductivity between silver and copper? Today, that’s easy to find out. You just ask Google (maybe even out loud if you have a phone handy). But it wasn’t that long ago that you needed another option. Before the Internet age, a big part of being “that guy” (or gal) was knowing where to go to find things. You had to be a master of the library’s reference section, know what might be in an encyclopedia or an almanac.
However if you were a hardcore math, science, or engineering geek you probably had, at least, one edition of CRC handbooks. Today, we usually think of CRC as cyclic redundancy check, but back then it was the Chemical Rubber Company.
The Chemical Rubber Company dates back to 1903 when brothers Arthur, Leo, and Emanuel Friedman were selling rubber lab aprons in Cleveland, Ohio (Arthur, apparently, had been in a similar business from 1900). In 1913, the brothers offered a short (116-page) booklet called the Rubber Handbook free with the purchase of a dozen aprons.
The Wall Street Journal reported that [Eric Schmidt] of Google and now Alphabet Inc, promoted the idea of an inexpensive version of the Raspberry Pi to the Raspberry Pi foundation’s [Eben Upton]. Apparently [Upton] accepted this recommendation despite existing plans to make a more expensive, more powerful version of the Pi. The outcome is the Raspberry Pi Zero that sells, in some places, for $5.00 and was given away for free on the cover of the MagPi magazine.
From the WSJ article:
“He [Schmidt] said it was very hard to compete with cheap. He made a very compelling case. It was a life-changing conversation,” Mr. Upton said, adding that he went back to the lab and scrapped all the engineering plans for more expensive versions of future Pi computers. “The idea was to make a more powerful thing at the same price, and then make a cheaper thing with the same power.”
Plans were scrapped. The more powerful Pi 2 was released at the price point of existing Pis, and now we have the Zero.
The Raspberry Pi Foundation is a registered educational charity in the UK. The purpose of this Foundation according to their About Us page is to, ‘advance the education of adults and children, particularly in the field of computers, computer science, and related subjects.’
Why is the Raspberry Pi Foundation so concerned about computer education? From the 1990s onward, fewer and fewer A Level students in the UK applying to study Computer Science had previous experience as hobbyist programmers. An applicant in the 2000s usually might have only done a little web design.
Why then does the Raspberry Pi Zero exist? [Upton] also told Cnet, “We really hope this is going to get those last few people in the door and involved in computer programming.”
Very good, but how well does the Zero support this goal or address their concerns?
In February, Google and Mattel introduced their Hello Barbie Internet-connected toy. This Barbie has an internal microphone, a WiFi connection to Google’s voice recognition services, and a speaker to carry on a “conversation” with the targeted child.
Like the folks at Somerset Recon, we’d say that this is an Internet of Things (IoT) device that’s just begging for a teardown, and we’re totally looking forward to their next installment when they pore through the firmware.
On the hardware front, Barbie looks exactly like what you’d expect on the inside. A Marvell 88MW300 WiFi SoC talks to a 24-bit (!) audio codec chip, and runs code from a 16Mbit flash ROM. There’s some battery management, and what totally looks like a JTAG port. There’s not much else, because all the brains are “in the cloud” as you kids say these days.
From day to day we alternate between the promise of IoT and being anti-IoT curmudgeons, so it should come as no surprise that we’re of two minds about Hello Barbie. First, there’s the creepy-factor of having your child’s every word overheard by a faceless corporation with “evil” in their mission statement (see what we did there?). Next, we’re not sure that it’s OK to record everything your child says to a toy and listen to it later, even if you are the parent. Hackaday’s [Sarah Petkus] summarized this neatly in this article.
But mostly, we’re curious about how well the thing actually works and what it will do with naughty words. And who will take on the task of reviving the Barbie Liberation Organization? Now we totally want to go out and buy one of these things.
There are a surprising number of Raspberry Pis being used in industrial equipment. This means the Arduino is left behind, but no longer. There’s your PLCs that use Arduinos.
A few weeks ago, Google introduced a machine intelligence and computer vision technique that made the world look psychedelic. Now, this library is available. On another note, head mounted displays exist, and a sufficiently creative person could mash these two things together into a very, very cool project.
Welcome to Kickstarter! Kickstarter is an uphill battle. People will doubt you because you don’t have a ‘target audience’ or ‘the rights to this franchise’ or ‘any talent whatsoever’, but that’s what crowdfunding is for!
Several years ago, Apple shipped a few million 17″ iMacs with defective displays. They’re still useful computers, though, especially if you can find a replacement LCD. Apple, in all its wisdom, used a weird connector for this LCD. Here’s the adapter board, and this adapter will allow displays running up to 1920×1200.
[Jan] has earned a reputation of building some very cool synths out of single ARM chips. His previous build was a Drumulator and now he’s shrinkified it. He’s put four drum sounds, pitch CV, and audio out on an 8-pin DIP ARM.
YouTube gives you cadmium! [AvE], recently got 100,000 subscribers on his YouTube channel. Apparently, YouTube sends you a terrible belt buckle when you manage to do that. At least he did it without playing video games and screaming.