Microsoft Is Training AI Gliders to Fly Themselves

• Updated on: 2017-08-17
• Read original article here

Click search or press enter
The Download
What's up in emerging technology
Today
Microsoft Is Training AI Gliders to Fly Themselves
Last week in the Nevada desert, Microsoft researchers threw a glider into the air, then left it to its own devices hoping that it would fly for hours. But they’re not as foolhardy as they sound: the New York Times reports that their Styrofoam aircraft is loaded with AI to study data from onboard sensors, predict conditions, and seek out thermals to keep it aloft. The long-term goal: to have the aircraft fly for days to track weather or deliver Internet to rural areas. That's much like the aim of Alphabet’s Project Loon, which  also uses AI to help control its stratospheric balloons . But the Loon balloons can only use inflation to move up or down: Microsoft’s glider has far more decisions (and turns) to make in order to stay aloft.
Source:
an hour ago
Need a Polio Vaccine? Get a Plant to Grow It For You
Researchers from the U.K.'s John Innes Center have hijacked tobacco plants to have them produce a polio vaccine. Writing in Nature Communications , the team explains that they inserted genes from the polio virus into soil bacteria that's been engineered… Read more
Researchers from the U.K.'s John Innes Center have hijacked tobacco plants to have them produce a polio vaccine. Writing in Nature Communications , the team explains that they inserted genes from the polio virus into soil bacteria that's been engineered to infect plants. That causes the plants to produce something that looks like the polio virus, but is stripped of its harmful attributes, making it perfect for training people's immune systems to fight the disease without causing infection. The vaccine, once extracted, protects mice from polio. The team tells the BBC that it could create vaccines for viruses like Zika, too.
Image credit:
Yesterday
Inside the Self-Driving Bromance of Travis Kalanick and Anthony Levandowski
Uber and Waymo's nasty legal dispute has taken a strange turn, as the release of a series of text messages between the former Uber CEO and his star self-driving-car engineer charts the blossoming of a close connection between the two.
For those unfamiliar,… Read more
Uber and Waymo's nasty legal dispute has taken a strange turn, as the release of a series of text messages between the former Uber CEO and his star self-driving-car engineer charts the blossoming of a close connection between the two.
For those unfamiliar, Waymo has alleged that Levandowski stole secrets related to its lidar system and brought them with him to Uber. The lawsuit has become a blockbuster, because depending on the ruling, it could severely damage Uber's plans— which have already taken several hits —to roll out self-driving taxis across its massive ride-hailing network. As part of the suit, Waymo thought texts between the two might shed light on whether Kalanick, and by extension Uber, was indeed after the trade secrets.
It offers up some interesting gems, including the fact that Levandowski and Kalanick both paid a lot of attention to Tesla:
9/14/2016 Levandowski: Tesla crash in January … implies Elon is lying about millions of miles without incident. We should have LDP on Tesla just to catch all the crashes that are going on.
9/22/2016: We’ve got to start calling Elon on his shit. I'm not on social media but let's start "faketesla" and start give physics lessons about stupid shit Elon says like [saying his cars don’t need lidar]
Levandowski also spent time schooling Kalanick in the intricacies of self-driving technology, and Kalanick was more than happy to hand out management advice in return. The hoped-for bombshell that would prove Uber illegally used Waymo's proprietary technology isn't there, but you can still check out the choice cuts over at IEEE Spectrum .
Source:
18 hours ago
Here Come the Mind-Blowing Smartphone Sensors That Could Bring AR to the Masses
Over the course of the next year, get ready for phones to get some awesome new imaging capabilities. We're talking more than just a tweak to the camera, or stereoscopic vision. Think full-on depth sensing and mapping. 
Today, Qualcomm released a video… Read more
Over the course of the next year, get ready for phones to get some awesome new imaging capabilities. We're talking more than just a tweak to the camera, or stereoscopic vision. Think full-on depth sensing and mapping. 
Today, Qualcomm released a video demonstrating the abilities of its next-generation Snapdragon mobile processor and "active depth sensing" sensor package, in which thousands of dots of infrared light map objects in three dimensions in real time (check out the inset in the video above). Qualcomm avoids calling it lidar, but the effect is the similar: any device with active depth-sensing on board gains the ability to gauge the shape of objects and their distance to within about 0.1mm, the company claims. The Verge notes  that Android phones featuring the new Snapdragon chips and sensor could be available in 2018.
That might be a little bit behind the timeline Apple is planning for its first phone equipped with laser range-finding. In July, it was widely reported that such hardware may be unveiled at Apple's annual event in the fall. We thought that made a lot of sense , as Apple is clearly pushing the development of augmented reality apps for iOS, and super-accurate depth sensing would be a key technology for helping AR take off on mobile devices.
Source:
Get The Download delivered to your inbox every day.
The Download
Amazon’s Latest Delivery Estimate: Two Minutes
Jeff Bezos loves speed. First it was next-day delivery with Prime. Then it was same-day. Amazon Now upped the ante with in-the-hour provision of, well, whatever you need inside an hour. And AmazonFresh Pickup has groceries ready for you within 15 minutes.… Read more
Jeff Bezos loves speed. First it was next-day delivery with Prime. Then it was same-day. Amazon Now upped the ante with in-the-hour provision of, well, whatever you need inside an hour. And AmazonFresh Pickup has groceries ready for you within 15 minutes.
Amazon’s latest trick is getting an order to you in two minutes flat. It's the e-commerce giant’s new Instant Pickup offering , which is being rolled out at five college campuses, including University of California, Berkeley, with more claimed to be on the way.
The service isn’t quite as impressive at it may seem at first blush. Customers order from a list of several hundred popular items via the Amazon app—think headphones, candy, and, naturally, Echo Dot smart speakers. Employees then scurry around a storage room, pop the items into a locker within a couple of minutes, and send the customer a bar code that can be used to unlock a door and retrieve the loot. Definitely fast, but not the kind of full-on choice that you might have hoped for.
By this point, of course, it’s clear that Amazon wants to take over physical retail as well as digital, with its recent acquisition of Whole Foods , real-life bookstores , and a range of retail experiments taking place in Seattle . With Instant Pickup, though, it seems to be taking a weirdly labor-intensive aim at ... vending machines. And Amazon's director of student programs, Ripley MacDonald, tells Reuters as much , explaining that he sees the system being perfect for ordering a can of soda, and even going so far as saying that the firm considered, but ultimately rejected, full automation of the service.
Source:
a day ago
Falling Robots Are Funny, but That’s How They’ll Learn to Take Your Job
If at first you don’t succeed, try again—and, if you’re a robot, again and again and again and again and again and again. Because it’s worth remembering that unlike many humans, automatons will keep at a task until they do achieve success.
This GIF of… Read more
If at first you don’t succeed, try again—and, if you’re a robot, again and again and again and again and again and again. Because it’s worth remembering that unlike many humans, automatons will keep at a task until they do achieve success.
This GIF of Boston Dynamics’s Atlas robot taking a tumble, sliced from a TED talk published Monday , has gone viral. Presumably, that's because when humans aren’t fretting about how they’ll steal our jobs , we sure do seem to enjoy laughing at robots falling over .
But robots are tenacious: failures don’t demoralize them the way they do humans. So you can bet that Atlas carried on trying to nail the task of moving a box for hours, and then shared its learnings with all its buddies so that none of them make the same mistake in the future. (And at any rate, the robot that steals your job is unlikely to be a humanoid .)
So laugh all you want, but automation is inevitable. It might not be as fast as some people promise, and there may be plenty of slips along the way, but it is happening .
Image credit:
a day ago
China Is Establishing the World’s Biggest Carbon Market
When President Donald Trump withdrew America from the Paris pact , China emerged as the world’s unlikely climate leader . Now  ClimateWire reports that China is living up to that role, with plans to establish a nationwide carbon market by November that… Read more
When President Donald Trump withdrew America from the Paris pact , China emerged as the world’s unlikely climate leader . Now  ClimateWire reports that China is living up to that role, with plans to establish a nationwide carbon market by November that will be larger than any other in the world.
It’s not yet clear exactly how the market will work. ClimateWire speculates that it might require organizations to buy carbon allowances in order to cover their emissions, or it could be a less stringent regulation that just requires reporting of emissions. But in terms of scale, it will reportedly cover four times the quantity of emissions as a previous carbon market in China, making it the largest ever such scheme on the planet.
It’s not being universally celebrated, though. Critics say that it will cover only 25 percent of the nation’s industrial carbon dioxide and is unlikely to work as well as Western efforts at first—especially ambitious ones like the scheme laid out recently by California . Then there’s the fact that no such market has yet been able to sharply reduce emissions.
Even so, it’s a sign that China takes its de facto climate leadership seriously. And if it sets an example to other Asian nations, that’s definitely no bad thing.
Source:
2 days ago
Increasing Minimum Wage Puts More Jobs at Risk of Automation
When the minimum wage goes up, the robots come for people's jobs. That's the upshot of a paper published today on the National Bureau of Economic Research's website ( abstract , full PDF paywalled), which analyzed how changes to the minimum wage from 1980… Read more
When the minimum wage goes up, the robots come for people's jobs. That's the upshot of a paper published today on the National Bureau of Economic Research's website ( abstract , full PDF paywalled), which analyzed how changes to the minimum wage from 1980 to 2015 affected low-skill jobs in various sectors of the U.S. economy. 
Federal minimum wage is currently $7.25 an hour, the same level it's been at since 2009. But 30 states have laws on the books that mandate a higher wage—it's $11 in Washington State, for example, and Seattle recently voted to phase in a pay hike that would bring it to $15 by 2022. Such measures are designed to ensure that "minimum wage" is the same thing as a "living wage."
Interestingly, a study of Seattle's new law, released in June,  suggested that cuts to working hours meant people were actually losing as much as $125 a month.
The new analysis, by Grace Lorden of the London School of Economics and David Neumark at the University of California, Irvine, suggests that there's a similar negative effect among people who work minimum-wage jobs that machines can do. The researchers found that across all industries they measured, raising minimum wage by $1 equates to a decline in "automatable" jobs—things like packing boxes or operating a sewing machine—of 0.43 percent.
That may not sound like much, but we're talking about millions of jobs across the entire U.S. economy. And certain industries were affected far more than others—in manufacturing, an uptick of $1 in minimum wage drove employment in automatable jobs down a full percentage point.
Of course, we know that automation is already gobbling up jobs in the U.S. (see " Who Will Own the Robots? "). This latest study suggests that even wage policies designed to help America's workforce may instead be speeding up that process.
Source:
2 days ago
An AI Dreamed Up Street Scenes, and They’re Surprisingly Good
You're looking at pure fiction: this image was actually created by an AI , trained on the kinds of driver's-eye labeled images often supplied to self-driving cars. Usually, humans describe which parts of a picture are, say, cars or sidewalks, and the… Read more
You're looking at pure fiction: this image was actually created by an AI , trained on the kinds of driver's-eye labeled images often supplied to self-driving cars. Usually, humans describe which parts of a picture are, say, cars or sidewalks, and the labeled images are used to train neural networks to recognize what they're looking at. Instead, Qifeng Chen, from Stanford University and Intel, got a similar neural net to use those learnings to render new street scenes. It puts a road somewhere down the middle, trees down the side, cars on the road ... and the results are surprisingly good.
Chen tells New Scientist that the software could be used to great effect in video games, where it could create realistic virtual worlds on the fly. It's worth bearing in mind that the games industry is big business: Twitter famously invested $150 million in Magic Pony, which also uses AI to generate high-quality computer-game graphics , giving the startup a valuation of a cool $1 billion.
Source: