How Drones Can Help Nepal Recover From the Earthquake
Nepal has suffered its worst earthquake in 80 years. Across the affected region, which stretches into India and Tibet, the death toll may top 10,000. The poor Asian nation’s limited infrastructure has been stretched to the breaking point by overwhelming need, as first responders attempt to rescue those trapped in remote areas and assess the damage
Drones will play a practical part in the national and international relief response to this enormous natural disaster. Nepal is currently suffering from a shortage of available manned helicopters, and drone first responders hope that their camera-bearing robots will be able to fill in some of the gaps, leaving precious helicopters for rescue missions.
Unmanned aerial vehicles, or UAVs, are valuable in several ways. A drone video of the devastation in Kathmandu demonstrates their considerable worth as a storytelling tool. With crumbled Buddhist temples and crowds of frightened Nepalese victims, Kishor Rana’s drone video, uploaded to YouTube on Monday, is a powerful visual record of the destruction the earthquake has wrought.
Drones will also be used for more comprehensive data collection and mapping of the destruction. Why use drones for disaster response instead of ground surveys, manned aircraft, or satellite imagery? All of these more traditional tactics remain important, but UAVs play a complementary role. Ground-based field surveys take a long time to carry out and are especially difficult to manage in the rough terrain of the Himalayas, while quick-moving UAVs can cover as much as 5 to 10 square kilometers in a matter of hours, at a very high resolution.
Weather is another factor: Small UAVs can be deployed below cloud cover that hampers satellite imagery visibility, a particular problem in Nepal, where overcast conditions have made gathering post-disaster satellite visuals difficult over the past few days.
The drones headed to Nepal are coming from around the world. For instance, India has announced it’s sending UAVs to the disaster area, and the Canadian relief charity GlobalMedic has already begun flying UAVs in the Nepalese disaster area, using machinery supplied by Ontario-based Aeryon Labs.
More international teams of drone pilots are headed to Nepal, where they will conduct aerial surveys of areas identified as a high priority by the U.N. Office for the Coordination of Humanitarian Affairs and UNICEF. The teams have been organized via the Humanitarian UAV Network organization, also known as UAViators. Team Rubicon, which is staffed by U.S. military veterans, is already on the ground in Nepal with drones in tow. As of Tuesday, teams from the U.S. drone startup SkyCatch and UAV services provider Halo Drop were en route to Nepal as well.
The drone pilots plan to create high-resolution nadir imagery (shot from directly above the ground) and oblique imagery (shot at a 45-degree angle from the ground) once they arrive. The resulting aerial imagery will be processed to create detailed 3-D point clouds—large collections of points in a three-dimensional coordinate system, which can then be used to create 3-D models of the terrain below. The 3-D models, which look a lot like video game maps, give responders more detailed information about the destruction than traditional 2-D maps can furnish, and today’s inexpensive UAV technology is particularly adept at shooting the imagery required to make them.
The year-old UAViators network, an initiative of the Qatar Computing Research Institute, has become a central organizing point for drone pilots who want to use their technological skill to help established humanitarian aid organizations engage in damage assessment and rescue efforts in the wake of disaster. Most recently, UAViators was activated after Cyclone Pam in Vanuatu. Drone pilots in the island nation collected high-resolution imagery of the cyclone’s destruction, which was then published on mapping website Mapbox, and shared with the Humanitarian Open Street Map Team and the MicroMappers crowdsourcing tool.
Online volunteers used Web tools to map the path of destruction and to assess the damage, including rating the destruction to buildings on a three-tiered scale, and updating Open Street Map maps based on the recorded destruction. Now, volunteers from around the world are using the same tactics to map the destruction in Nepal, and UAV imagery should be added to their repository of source data in the next few days.
While drone imagery may be extremely helpful in disaster assessment situations, it’s worth emphasizing that so-called cowboy drone pilots are less than welcome at the scene of major disasters. Organization and coordination are both key to carrying out useful post-disaster drone missions, in tandem with aid agencies that know the situation. There are many special considerations involved with safely using a UAV in disaster zones with many manned helicopters in the air and in high-altitude, difficult terrain such as that found in Nepal—considerations that the Center for Robot-Assisted Search and Rescue at Texas A&M has handily outlined.
Simply arriving at the scene and launching a flying machine into the air is unlikely to be helpful and could even be dangerous—creating more work and more risk for first responders who are already pushed to the limit of their capacities. It could also hurt the reputation of well-organized groups using drones for rescue efforts.
So if you and your drone want to help, get in touch with the Humanitarian UAV Network and read the Network’s Code of Conduct, says UAViators founder Patrick Meier. But don’t just book a ticket to Kathmandu.
New Traffic-Enforcement Tech Peers Into Your Car and Counts Passengers
The go-to surveillance tech list includes things like automatic license plate readers and Stingrays (aka IMSI catchers). Now the Electronic Frontier Foundation is reporting that there's another device to add. The Automated Vehicle Occupancy Detection systems uses two video cameras and some analytics algorithms to surveil cars in carpool lanes and ticket them if they don't have enough passengers.
These sensor arrays were developed by Xerox and are currently part of a San Diego Association of Governments carpool lane enforcement pilot program on I-15. CBS 8 evaluated SANDAG documents that explain the system, including its 99-percent-accurate passenger counts. If the cameras detect a car that is wrongly driving in a carpool lane, they photograph the vehicle's interior and license plate and send the data to California Highway Patrol.
CBS 8 reports that the pilot system isn't storing license plate numbers or identifying information about individuals. In November, Xerox said in a statement that Automated Vehicle Occupancy Detection is part of its "advanced solutions" for "revolutionizing the movement of people and goods worldwide."
EFF says you're going to have to "add Automated Vehicle Occupancy/Passenger Detection to your vocabulary." It's being used to enforce carpool lanes now, but since it can see and count passengers and then tie them to particular cars at certain times, you can see how it could be used for other things.
Google Is Testing a Creative New Way to Thwart Patent Trolls
Even if you don't actually know what patent trolls are, you probably know they are a problem. These companies or individuals make money by purchasing intellectual property and then collecting fees from enforcing the patent rights (even though they have no specific interest in a given industry or idea). And since this practice wreaks havoc on startups and big companies alike, Google is trying to find a way to shut the trolls down.
Patent Purchase Promotion is basically an effort to streamline the process of selling patents to Google so there isn't much incentive to sell to trolls instead. As the company puts it, when you sell a patent to trolls, "bad things happen, like lawsuits, lots of wasted effort, and generally bad karma." Indeed.
Google doesn't mention whether it would ever use a patent from the Purchase Promotion in litigation, but the company did promise in 2013 that it would only wield patents defensively. From May 8 through May 22, Google will open the Purchase Promotion for submissions and will then evaluate the submissions and choose the patents it wants to buy by June 26. The company, which is calling the initiative "an experiment," says it hopes that all transactions will be finished by late August.
In a blog post, Google's deputy general counsel for patents, Allen Lo, wrote:
Hopefully this will translate into better experiences for sellers, and remove the complications of working with entities such as patent trolls. There’s some fine print that you absolutely want to make sure you fully understand before participating, and we encourage participants to speak with an attorney. ... Throughout this process, Google reserves the right to not transact for any reason.
As Gizmodo points out, though, the more patents Google owns, the more tempting it will be for the company to become a patent troll itself. Google says it's just trying to help, but even if the company means it now, it could become part of the problem later.
Is Crowdfunded Science Crap?
A crowdfunding campaign for a brain imaging study closed Monday after raising almost $80,000 toward a unique goal: the first functional magnetic resonance images of the brain on LSD. The Beckley Foundation, a U.K.–based charitable trust that promotes research and awareness of psychoactive drugs, will use the money to scan volunteers who’ve dropped acid. Such are the sacrifices people will make for science.
Now, it’s little surprise scientists studying the effects of illicit drugs must sometimes find unconventional benefactors—or that thousands of people would invest in seeing the brains of volunteers tripping balls. But in recent years, crowdfunding has grown increasingly popular among researchers in nearly every field. Successful campaigns have explored drought tolerance in Spanish and Indian oak species, attempted to explain jokes with math, and worked to discover exoplanets in the far reaches of space. The first crowdfunded experiments popped up on traditional platforms like Kickstarter and Indiegogo; now sites like Petridish, Experiment, and Walacea cater specifically to scientific fundraising.
In the U.S., most scientific funding comes from the government, distributed in grants awarded by an assortment of federal science, health, and defense agencies. So it’s a bit disconcerting that some scientists find it necessary to fund their research the same way dudebros raise money for a potato salad. Does that migration suggest the current grant system is broken? If it is, how can we ensure that funding goes to legitimate science working toward meaningful discoveries?
On its own, the fact that scientists are seeking new sources of funding isn’t so weird. In the view of David Kaiser, a science historian at MIT, crowdfunding is simply the latest “pendulum swing” in how scientists and research institutions fund their work. Once upon a time, research at MIT and other universities was funded primarily by student tuition and private philanthropists. In 1919, however, with philanthropic investment drying up, MIT launched an ambitious plan that allowed local companies to sponsor specific labs and projects.
Critics complained the university had allowed corporate interests to dig their claws into scientific endeavors and befoul intellectual autonomy. (Sound familiar?) But once WWII began, the U.S. government became a force for funding, giving huge wartime grants to research groups nationwide. Federal patronage continued expanding in the decades after the war.
Seventy years later, that trend has reversed: As the federal budget shrinks, government investment in scientific research has reached new lows. The conventional models for federal grants, explains University of Iowa immunologist Gail Bishop, “were designed to work such that 25 to 30 percent of studies were funded. Now it’s around 10 percent.”
That’s part of the reason scientists like Bishop, who successfully funded a study of new nanoparticles to fight cancer cells on Experiment, have embraced crowdfunding. The grant-awarding process can feel arbitrary, says Bishop, and “the selection of those studies skews towards projects that take fewer risks.” In addition, most grants require scientists to provide proof-of-concept studies for their research, which makes it extremely difficult for more experimental projects to get off the ground. “It used to be that the grant money was there to do the experiment,” says Bishop. “Now the cynical joke is that you need to do the experiment first in order to be awarded a grant.”
Crowdfunding allows small-scale research to bypass the time and trouble required to draft and defend a grant proposal, and find money to cover the startup costs of a project. While most biomedical grants leave nine long months between submission and decision, Bishop was pleasantly surprised to see her campaign progress in real time toward its meager $1,500 goal.
But like all shortcuts, crowdfunding has its downsides. Government-funded institutions might be bogged down by tight regulations, but those rules can act as crucial checks on studies that might stand on shaky evidence or harm participants. (The Beckley Foundation’s LSD study is being run at Imperial College, so the researchers must follow institutional guidelines that are meant to protect participants.) Government grants also typically require a transparent accounting of how money is used, while money raised through crowdfunding doesn’t have those stipulations attached. There’s always the possibility—however rare—that a scientist or researcher may use funds in irresponsible or unethical ways.
The “crowd” part of crowdfunding also presents a significant problem. When scientists have to shill their ideas on social media like every other entrepreneur, there’s nothing to stop sexy, sensational campaigns from overshadowing more important or legitimate studies. “You want to go viral, but that’s not so easy to do,” says Aaron Seitz, a psychologist at UC–Riverside who used a crowdfunded campaign to pay for his study on auditory dysfunction in veterans of war. Hundreds of thousands of dollars have been given to Immunity Project, a company seeking to develop a vaccine for HIV, but whose scientific credentials are debatable at best. That’s money that could have instead gone toward campaigns to study autism, or measure toxic chemicals in local bodies of water.
That’s the advantage of agency-funded research: At least in theory, there’s a body of intelligent, scientifically informed individuals making the call about what research needs to get done. You can bet that the majority of funders of the Beckley LSD study didn’t do it because they believe in the scientific pre-eminence of the researchers behind it: They did it because they like the idea of sticking high people in an fMRI (and who wouldn’t?).
But that doesn’t mean that the science getting crowdfunded is crap. The researchers for the LSD study, for example, have been investigating the effects of psychedelics for years now, and their results have been published in several prestigious journals. The majority of people looking for funding on these platforms are credentialed scientists whose work will ultimately be subjected to peer review. As long as those checks and balances are still in place, any junk science that makes its way onto a crowdfunding platform still won’t get published. And the amount of money awarded to poorly designed studies will remain quite low. Common funding goals rank in the single-digit thousands—only a fraction of the size of a typical federal grant.
“Crowdfunding won’t replace conventional means of funding,” says Kaiser. What it will do is offer a chance for smaller studies to take off when they would be denied the opportunity otherwise. Kaiser thinks small projects, especially theoretical research that perhaps doesn’t require as much in equipment and staff costs, will take advantage of crowdfunding more and more. That, and probably a few more high fMRI studies than you’d otherwise see the NIH fund.
Also in Wired:
After Nepal Earthquake, Facebook and Google Offer Safety Check-In Features
Google and Facebook both offer check-in tools for large-scale disasters so users can indicate whether they and others are safe. And people in Nepal have been using both in the wake of a 7.8-magnitude earthquake that hit the country early Saturday morning. With the death toll above 1,000, and rescuers frantically trying to free people trapped in rubble, those who are safe or abroad have been getting information from these services.
Google engineers began working on the Person Finder feature after the January 2010 earthquake in Haiti. Google.org (the company's charitable division) has since expanded its functionality, and deployed it during crises like the 2013 Boston Marathon bombing. Meanwhile, Facebook's Safety Check, which launched in October, sends notifications to users who may be in a crisis area asking them "Are you safe?" Users can post "I'm safe" or "I'm not in the area," and can also indicate that other people they are with are safe.
The usefulness and even importance of these types of services is obvious, but the major problem with them is that people can post about their safety only if they have connectivity and an Internet-enabled device. If someone's smartphone is smashed in a collapsed building, or a natural disaster knocks out telecom systems in an area, it may be impossible to send a digital update. This was an issue during large-scale floods in Pakistan during 2010, when Google attempted to deploy Person Finder, mostly unsuccessfully. UNICEF reported Saturday that cellphone networks and power were out in many parts of Nepal.
Anecdotally, a Nepalese business owner in Harlem, New York, told Slate that he had been receiving some updates through social media from friends and family in Kathmandu, but that many people he knew were in areas that had lost connection.
Customers Aren’t Happy With Their Data Security. Execs Are the Only Ones Surprised.
There's always something going on with corporate data breaches or new digital security initiatives. And those things look very different to the executives who run companies than they do to the customers who use their products and services. A new study from consulting firm Deloitte illustrates in a few stark (and kind of hilarious) stats just how varied the views are.
The survey polled 2,001 American consumers and 70 consumer product executives, and the consumers were generally less positive or enthusiastic than the executives about corporate security/privacy efforts. For example, 47 percent of execs thought that customers felt it was worth it to share their personal information with companies in exchange for perks like coupons and customized promotions. But 75 percent of consumers disagreed. For product reviews, 47 percent of executives thought consumers found sharing their data worthwhile. Only 18 percent of consumers agreed. Ruh roh.
Consumers don't always seem to know the best way to protect their digital identities and assets, but they know that they don't know. Only 28 percent of consumers surveyed said they thought they knew which companies would protect their personal data. But 83 percent did say that they're aware of retailer data breaches. Only about half said they would be forgiving if a company had a data breach.
The researchers wrote, "Our survey suggests that the field is wide open for consumer product companies to build a reputation for strong data privacy and security practices," which is really just a sneaky way of saying that not nearly enough companies are implementing best practices right now.
Probably the most revealing stat is that 77 percent of execs said they think their companies' data privacy policies are clear and easy to understand. Ha! About 73 percent of consumers said they want more straightforward data privacy policies.
Microsoft Word Spells the Names of Game of Thrones Characters Better Than You Can
Jon Snow may know nothing, but Microsoft Word knows more than you might expect.
When Slate’s Jonathan Fischer sat down to write his review of the latest Game of Thrones season, he encountered an unexpected difficulty: It’s very hard to spell the name Daenerys. Worried that he wouldn’t notice if he got it wrong, he resorted to a manual solution. “I did a lot of copying and pasting Daenerys from my browser,” he told me.
Maybe he shouldn’t have gone to the trouble. Surprisingly, Microsoft Office’s spell-checker recognizes the names of virtually every Game of Thrones character, major and minor alike. Historian Greg Jenner, who first called attention to this oddity on Twitter, shrugged it off with a simple, “who knew?”
Microsoft Word's spellcheck seems to know all the names of Game of Thrones characters. Huh, who knew?— Greg Jenner (@greg_jenner) April 21, 2015
Things may be stranger than Jenner realized. This is an honor that the suite of programs seemingly affords to no other popular franchise. While neither Shae nor her spurned lover Tyrion Lannister elicit complaint from Word, Lord of the Rings generates all kinds of trouble: Red lines of judgment run beneath the names of five of the Fellowship of the Ring’s nine members—Samwise Gamgee, Legolas, Gimli, Boromir, Meriadoc Brandybuck.*
Much the same is true for important characters from Star Wars, The Hunger Games, and even the Harry Potter franchise. As with Tolkien’s Lord of the Rings, Office’s spell check recognizes the names of some prominent characters from each while ignoring others. Like old Ben Kenobi, it has heard the name Obi Wan, but not that of his old foe, Count Dooku. Perhaps Microsoft—like most endowed with good taste—simply doesn’t acknowledge that the Star Wars prequels exist. It’s more difficult, however, to explain why it seems friendly with Severus Snape, but not with Harry Potter’s actual friend, Rubeus Hagrid.
There are exceptions to the Game of Thrones rule: Characters who appear only in the novels still tend to get red-lined. Likewise, some place names (Meereen, Dorne, Yunkai, etc.) fail to appear in Microsoft’s dictionary. And for some reason the mysterious assassin Jaqen H’ghar goes unacknowledged, though that may be because “a man is not Jaqen H’ghar,” as he puts it. Nevertheless, the results are striking: You should be able to write your next episode recap with confidence. But prepare for trouble if, for some reason, you’re planning an essay on Lloyd Alexander’s Chronicles of Prydain.
There’s one person who probably won’t benefit from this discovery: Game of Thrones creator George R.R. Martin. In a 2014 interview with Conan O’Brien, Martin explained that he still uses the late-1980s program Word Star 4.0 when he’s writing. “I hate spell-check,” he told O’Brien at the time. Perhaps Word’s surprising familiarity with Martin’s work is a quiet rebuke to his disdain.
Update, April 24, 1:52 p.m.: A Microsoft representative just sent the following statement:
Glad you noticed that we are not just about common words. We regularly update the spellers to keep them fresh, including additions from the latest, most frequent names from movies, books, and TV shows. To do this, we research what people are talking about, what’s trending in the business world, current affairs, and other popular domains. We can’t add everything that comes up, so we reference different sources and determine which words to include. One of the 2014 lexical updates included the addition of characters from the Game of Thrones. Names relating to the TV show surfaced through several data sources which qualified them to be added. Up until 2014 we updated the English speller quarterly with 12,000 words added last year. Since January 2015 we’ve been updating the English speller on a monthly basis, and are on track to add an additional 32,000 words in 2015.
*Correction, April 24: This article originally misstated that Tyrion Lannister’s lover was named Osha Shae. Her name is simply Shae. Osha is a different character. A graphic in the post also misspelled the name of Chronicles of Prydain character Eilonwy. The graphic has been updated.
Do Robots Urinate? And Other Questions Raised by This Google Maps Prank.
There are just so many questions about this Android robot urinating on an Apple logo in Google Maps. Is it accurate to depict a robot peeing? If you want to support Google over Apple, isn't this a weird way to do it, since you're just getting runoff urine all over a Google product? We could go all day.
In a statement, Google explains:
The vast majority of users who edit our maps provide great contributions, such as mapping places that have never been mapped before, or adding new business openings or address changes. We’re sorry for this inappropriate user-created content; we’re working to remove it quickly. We also learn from these issues, and we’re constantly improving how we detect, prevent and handle bad edits.
The Guardian points out that a few miles east of the Android/Apple situation someone has added a "Google review policy is crap :(" message. So, yeah, it seems like someone figured out how to game Google's Map Maker feature and then added these images to make a statement. The whole thing is reminiscent of a similar, but less visually impactful, prank from last week in which users got a snowboard shop called "Edward's Snow Den" verified on Google Maps and then changed the address to make it look like Edward Snowden was inside the White House.
Google says it learns from these situations, but the takeaway seems to be that the company needs a better approval system for Map Maker and verified locations. Otherwise things like this will probably, you know, keep happening.
Why Don’t Computers Care More About Our Happiness?
Your computer doesn't care if you’re smiling. Why would it?
Your computer isn’t a person. It’s a computer! A tool. A machine. Computers are logical. They’re rational. They don’t get tired, or sad, or frustrated. They don’t get distracted. They’re oblivious to human happiness. They weren’t programmed to do emotion.
Your computer cares about important, useful things. It cares about your environment, dimming its screen when you turn off the lights. It cares about connecting you to friends and family, lighting up with a buzz whenever they want to get in touch. Increasingly, it cares about context: your next appointment, your next flight, the traffic in your city.
A smile? That’s just obviously not a computer’s concern.
All of which is to say: Smile Suggest, a Chrome extension that uses an open source computer vision library to automatically bookmark and share websites that make you smile—like, actually smile, with your mouth—is obviously a lark. Martin McAllister, the British copywriter who created it, called it “daft” no fewer than three times in our brief conversation, just so there was no confusion. There isn’t. It’s a joke! We don’t control our computers by smiling at them.
Still, McAllister did hazard what I thought was an astute observation about his “daft little B-side.” He began cautiously: “If I can read into it this deeply, Smile Suggest is a slight, flippant way of making a deeper point. When we like or share something, it’s not totally genuine. It’s something we’re putting our name on. A smile is something you do without thinking. You don’t have those layers of cynical thought. It’s just what you like and what’s funny to you.”
To any modern computer user, the idea of having every site that elicits a smile beamed automatically to Facebook is mortifying. But McAllister’s little extension got me wondering: Could a smile be a useful signal for a computer? Might we be able to do something interesting with such a genuine, unfiltered bit of input? Probably. I would like to review every YouTube video that made me laugh in 2012. I’d be delighted if my computer pointed me to a Gchat conversation, long forgotten, that made me crack up in college.
Granted, in a world of presumed total surveillance, it’s upsetting to imagine our computers having access to something as intimate as our unmediated emotions. That’s our last stand against the bureaucrats and the brands, the unquantifiable inner sanctum of self.
But supposing some alternate arrangement in which we could actually trust our devices and the people making them, emotion could be a profoundly powerful principle to design around. The Apple Watch will buzz you with a reminder to stand up if you’ve been sitting too long, perhaps the first time a mainstream consumer electronic device has tried to spur a healthy behavior change by default, right out of the box. Is it inconceivable that someday our gadgets would care about our emotional health in the same way? Smile Suggest is unthreateningly, unambiguously a lark—but what if it didn’t have to be?
A smile is the most basic unit of human happiness, the joy response embedded in our genes, more universal than any language or culture or custom. Shouldn’t that obviously be a computer’s concern?
Your computer already cares about a bunch of dumb things. It cares about your environment, dimming its screen when you turn off the lights, even though its blue light is screwing up your circadian rhythms. It cares about connecting you to the companies whose apps you’ve installed, lighting up with a buzz whenever they want to engage you. Increasingly, it cares about context: your next appointment, your next flight, the traffic in your city—even though, as a high school sophomore in Baltimore, you don’t use a calendar app, aren’t flying anywhere, and are still a decade away from your first rush-hour commute.
Your computer isn’t a person, but as psychological studies have shown, you often can’t help but treat it like one. Nope, your computer is just a dumb tool, a lowly machine. Computers were designed to be logical, rational, like we once thought humans to be, before we knew better. Computers don’t know to help when you’re tired, or sad, or frustrated. They don’t steer you away from distractions. They’re oblivious to your happiness. We don’t design them to do emotion.
Your computer doesn’t care if you’re smiling. It can’t hurt to occasionally ask: Why doesn’t it?
Also on Wired:
Swiss Postal Service Will Start Using Delivery Drones in Pilot Program This Summer
Swiss Post, Switzerland's postal service, wants to make drone mail delivery happen. For real. The service is partnering with delivery drone company Matternet to run a pilot this summer in which ONE drones will carry pieces of mail weighing up to 2.2 pounds for up to 12 miles.
Matternet ONE drones autonomously follow routes plotted by the company's cloud software. Light packages and other mail will be eligible for delivery through the project. Matternet said in a statement, “The primary aim of this pilot project is a Proof of Concept to clarify the legal framework, consider local conditions and explore the technical and business capabilities of the drones.”
Matternet has tested its drones in Haiti, but the Swiss Post collaboration is an opportunity to work out the logistics and complications of doing drone delivery in a different range of conditions. TechCrunch reports that Matternet has flown more drone hours than any other company. “We are extremely excited to bring Matternet ONE to Switzerland,” said Andreas Raptopoulos, Matternet's CEO.