Man Buys Two Metric Tons of LEGO Bricks; Sorts Them Via Machine Learning

iStock // Ekaterina Minaeva
iStock // Ekaterina Minaeva

Jacques Mattheij made a small, but awesome, mistake. He went on eBay one evening and bid on a bunch of bulk LEGO brick auctions, then went to sleep. Upon waking, he discovered that he was the high bidder on many, and was now the proud owner of two tons of LEGO bricks. (This is about 4400 pounds.) He wrote, "[L]esson 1: if you win almost all bids you are bidding too high."

Mattheij had noticed that bulk, unsorted bricks sell for something like €10/kilogram, whereas sets are roughly €40/kg and rare parts go for up to €100/kg. Much of the value of the bricks is in their sorting. If he could reduce the entropy of these bins of unsorted bricks, he could make a tidy profit. While many people do this work by hand, the problem is enormous—just the kind of challenge for a computer. Mattheij writes:

There are 38000+ shapes and there are 100+ possible shades of color (you can roughly tell how old someone is by asking them what lego colors they remember from their youth).

In the following months, Mattheij built a proof-of-concept sorting system using, of course, LEGO. He broke the problem down into a series of sub-problems (including "feeding LEGO reliably from a hopper is surprisingly hard," one of those facts of nature that will stymie even the best system design). After tinkering with the prototype at length, he expanded the system to a surprisingly complex system of conveyer belts (powered by a home treadmill), various pieces of cabinetry, and "copious quantities of crazy glue."

Here's a video showing the current system running at low speed:

The key part of the system was running the bricks past a camera paired with a computer running a neural net-based image classifier. That allows the computer (when sufficiently trained on brick images) to recognize bricks and thus categorize them by color, shape, or other parameters. Remember that as bricks pass by, they can be in any orientation, can be dirty, can even be stuck to other pieces. So having a flexible software system is key to recognizing—in a fraction of a second—what a given brick is, in order to sort it out. When a match is found, a jet of compressed air pops the piece off the conveyer belt and into a waiting bin.

After much experimentation, Mattheij rewrote the software (several times in fact) to accomplish a variety of basic tasks. At its core, the system takes images from a webcam and feeds them to a neural network to do the classification. Of course, the neural net needs to be "trained" by showing it lots of images, and telling it what those images represent. Mattheij's breakthrough was allowing the machine to effectively train itself, with guidance: Running pieces through allows the system to take its own photos, make a guess, and build on that guess. As long as Mattheij corrects the incorrect guesses, he ends up with a decent (and self-reinforcing) corpus of training data. As the machine continues running, it can rack up more training, allowing it to recognize a broad variety of pieces on the fly.

Here's another video, focusing on how the pieces move on conveyer belts (running at slow speed so puny humans can follow). You can also see the air jets in action:

In an email interview, Mattheij told Mental Floss that the system currently sorts LEGO bricks into more than 50 categories. It can also be run in a color-sorting mode to bin the parts across 12 color groups. (Thus at present you'd likely do a two-pass sort on the bricks: once for shape, then a separate pass for color.) He continues to refine the system, with a focus on making its recognition abilities faster. At some point down the line, he plans to make the software portion open source. You're on your own as far as building conveyer belts, bins, and so forth.

Check out Mattheij's writeup in two parts for more information. It starts with an overview of the story, followed up with a deep dive on the software. He's also tweeting about the project (among other things). And if you look around a bit, you'll find bulk LEGO brick auctions online—it's definitely a thing!

12 Facts About Netflix, Recommended For You

kasinv/iStock Editorial via Getty Images Plus
kasinv/iStock Editorial via Getty Images Plus

Netflix has become the world’s intravenous line for filmed entertainment. And like any media empire, it has a few stories of its own to tell. Take a look at some lesser-known, non-buffering facts about the streaming giant.

1. Early Netflix subscribers got a lot of Chinese pornography.

Addict man at computer laptop watching porn internet addiction concept
OcusFocus/iStock via Getty Images

In 1998, Netflix was still in the business of selling as well as renting DVDs. To try and offer consumers something new, co-founder Marc Randolph decided to offer footage of President Bill Clinton’s Grand Jury testimony about his involvement with Monica Lewinsky. But according to the book Netflixed, the duplicating house had a mix-up: out of the 1000 customers who ordered Clinton's interview, a few hundred received discs full of hardcore Chinese pornography.

2. Netflix was originally called Kibble.

Choosing a name for the company was a drawn-out process. Directpix.com, Replay.com, and other names were considered; so was Luna.com, which was the name of Randolph’s dog. When the company was being incorporated, he named it Kibble.com until they could decide on something permanent.

3. Netflix executives used to make house calls.

From the beginning, Netflix has been preoccupied with seeing how users interact with its software in order to select titles. In the late 1990s, subscribers near the company’s location in Los Gatos, California were reached via telephone and asked a series of questions. Then staffers would ask if they could stop by to watch them use the site. Surprisingly, most agreed. Netflix brought them coffee, a small investment for gaining valuable information about their usage.

4. Netflix got Dennis Quaid to sing.

For a 2006-2007 publicity tour, Netflix decided to screen films in thematically-correct locations: For example, Field of Dreams was shown in the “real” Iowa cornfield-slash-baseball diamond featured in the movie. But the company also wanted actors to make appearances. Their approach: offer to let those with bands perform for the crowds. Kevin Costner, Bruce Willis, Dennis Quaid, and Kevin Bacon all agreed to the barter deal. Quaid and his band, The Sharks, played in New Orleans before a screening of his film The Big Easy.

5. Netflix has made a science out of spoilers.

Because so much of Netflix’s high-profile content can be “binged” in a single weekend, the company commissioned cultural anthropologist Grant McCracken to examine how spoilers affect a person’s viewing habits. McCracken identified classifications of spoiler-prone people by whether they ruin a plot twist intentionally or hold it over others. (Some people are “Coded Spoilers,” too self-aware to let anything slip. These people are your friends.) His verdict? Some people enjoy the power they get from having knowledge of spoilers. But if a show is good enough, knowing about key scenes won't dissuade viewers from watching.

6. Netflix staffers think you decide on a movie in two minutes.

Apple iPad displaying Netflix app, Black with Reflection
bmcent1/iStock Editorial via Getty Images

Netflix spends more than $150 million on improving their recommendation system every year, trying to arrange selections based on what they think you might like. That kind of personalized menu is necessary: The company estimates that users spend only two minutes browsing for a title before choosing one or opting for another diversion entirely.

7. Netflix staffers also think you might be kind of a liar. 

You can stop trying to impress Netflix with the streaming version of keeping Ulysses on your coffee table. In a 2013 WIRED interview, Carlos Gomez-Uribe—the company's vice president of product innovation from 2010 to 2016—noted that viewers often report viewing documentaries or esoteric foreign movies. “But in practice,” he said, “that doesn’t happen very much.”

8. the first "netflix original" was an abstract test footage short.

In order to test frame rates and how their streaming service handles different kinds of content, Netflix produced 11 minutes of test footage in 2011 that can be viewed by typing “example show” in their search engine. Cut together (as seen above), the shorts become a very strange, very abstract art film, with an unidentified man juggling and reciting Shakespeare. (But not, sadly, juggling while reciting Shakespeare.)

9. Netflix binge-watching might correlate with depression. 

A 2015 study by the University of Texas found that respondents who claimed to binge Netflix shows were more likely to suffer from depression, lack of self-control, or loneliness. The good news? The sample group was small—only 316 people—and the university’s definition of “binge-watching” was as low as two episodes. Amateurs. 

10. There’s a secret Netflix menu.

Netflix website showing on screen laptop with macbook pro at cafe
wutwhanfoto/iStock Editorial via Getty Images Plus

No, not that kind of secret menu. Pressing Shift + Alt + a left mouse click brings up a troubleshooting menu that allows you to adjust the bit rate of a stream so it doesn’t buffer. (On a Mac, it's Shift + Option + click.) The picture quality won’t be as good, but it’s better than a pixelated Demogorgon.

11. There was once a glitch in the Netflix matrix. 

In 2014, Netflix’s content descriptions became odd amalgamations of two different titles to create one completely nonsensical listing. The summaries were quickly fixed, but not before someone took several screen shots of the mishaps.

12. You'll soon be able to stream Netflix in a Tesla.


Justin Sullivan, Getty Images

In July 2019, Tesla founder Elon Musk informed Tesla owners they would soon be able to stream both Netflix and YouTube in their cars, an attractive option for anyone looking to keep passengers occupied. But there's a catch: The services only work when the cars are parked. The feature will be available in newer-model cars at a date to be determined.

Surprise! Microsoft Admits That People Are Listening In on Your Cortana and Skype Conversations

mabe123/iStock via Getty Images
mabe123/iStock via Getty Images

In a turn of events that, at this point, probably shocks no one, Microsoft has confessed that human contractors have been listening in on some people's private conversations via Skype and Cortana.

Vice first broke the news on the privacy breach earlier this month, after a contractor passed along documents, screenshots, and actual audio files of some conversations. “The fact that I can even share some of this with you shows how lax things are in terms of protecting user data,” the contractor, whose name was withheld (for obvious reasons), told Motherboard. Unlike Apple’s recent Siri snafu, these conversations didn’t include potential criminal activity, but they did catch intimate exchanges about weight loss, love, and relationship problems.

Also unlike Apple: Microsoft is not suspending its practices. Instead, the tech monolith has updated its privacy policy to clarify that humans might, in fact, be eavesdropping on you.

“We realized, based on questions raised recently, that we could do a better job of clarifying that humans sometimes review this content,” a Microsoft representative told Vice. Before, the Skype website had mentioned that your content could be analyzed in order to improve the technology, but it never explicitly stated that humans would be listening to it.

Microsoft only records Skype conversations that use its translation features, in order to “help the translation and speech technology learn and grow,” according to the Skype FAQ section. If you’re not using translation features, your sweet nothings are reportedly as private as you want them to be. The updated FAQ section also now states that “Microsoft employees and vendors” may be transcribing the translated audio, and the procedures are “designed to protect users’ privacy, including taking steps to de-identify data, requiring non-disclosure agreements with vendors and their employees,” and more.

But Cortana’s data gathering isn’t limited to translation. According to its support page, Microsoft can collect your voice data literally any time you “use your voice to say something to Cortana or invoke skills.” If that worries you, we recommend spending some time adjusting the settings on your Microsoft Privacy Dashboard.

[h/t Vice]

SECTIONS

arrow
LIVE SMARTER