The system stores a database of potential ditch sites for safe emergency landings, and is able to choose the ideal site based on range, size, type of terrain, reliability, and time or day constraints. It's a much more advanced system than what is currently used in most commercial UAVs, which require a designated "home" point, to which the vehicle will attempt to return in the case of a hardware malfunction or drained battery. Current models are unable to safely ditch if, for example, the remaining battery charge is unable to return the drone to its home point, or if that home point is out of date. Once these remaining technological challenges are solved, Roy believes that Safe2Ditch, or similar systems, could become an FAA-mandated safety standard in UAV manufacturing.
A new competition heralds what is likely to become the future of cybersecurity and cyberwarfare, with offensive and defensive AI algorithms doing battle. "It's a brilliant idea to catalyze research into both fooling deep neural networks and designing deep neural networks that cannot be fooled," says Jeff Clune, an assistant professor at the University of Wyoming who studies the limits of machine learning. Machine learning, and deep learning in particular, is rapidly becoming an indispensable tool in many industries. "Adversarial machine learning is more difficult to study than conventional machine learning--it's hard to tell if your attack is strong or if your defense is actually weak," says Ian Goodfellow, a researcher at Google Brain, a division of Google dedicated to researching and applying machine learning, who organized the contest.
Both cities are replacing outdated phone booths with Wi-Fi kiosks that have embedded computing tablets, USB charging ports, keypads for making phone calls, and large screens that display relevant information to passersby. New York, which started installing its "LinkNYC" kiosks in 2016, currently has more than 900 activated across all five boroughs and plans to increase that number to 7,500. Eventually, information from Intersection's future sensors could be combined to create real-time data maps that might be useful for emerging technologies such as self-driving cars. Next, Intersection is looking to deploy its digital screens in airports, apartment buildings, and office complexes.
Apprehensions about automotive cybersecurity came to a head when a pair of white-hat hackers broke into a Jeep Cherokee in 2015, leading to the recall of 1.4 million vehicles by Chrysler Fiat to fix a software bug in the Uconnect infotainment system (see "Carmakers Accelerate Security Efforts after Hacking Stunts"). "Dealing with consumer safety, and not just with data security, requires different security methods to protect our cars, in contrast to technologies that protect servers and enterprise networks," says David Barzilai, executive chairman and cofounder of Karamba Security, a two-year-old startup based in Hod HaSharon, Israel, with an office in metropolitan Detroit. Harman International, maker of the Uconnect system in the hacked Jeep, acquired TowerSec, an Israeli cybersecurity firm, in early 2016. Glen De Vos, chief technical officer for the automotive parts maker Delphi Automotive, says that layers of security beyond what Karamba is proposing will become necessary as cars develop more connected properties, including autonomous driving, and therefore transmit more data wirelessly both to the cloud and to one another.
On the factory floor the day of my visit, GeNI engineer Leopoldo Ortiz looks over a two-tiered work station that Genesis-ICESA designed for his factory in Puebla, a city three hours by car from the Bajío region. Wages in Mexico are low compared with levels in the U.S., where the average auto worker makes more in an hour than a Mexican worker makes in a day. The automation drive has cut back on Bosal's local hiring needs, although strong demand for parts has still led the company to increase its Mexican workforce by 50 percent since 2014, says Luis Palomé, regional director for Bosal in Mexico. Not far away, at a more manual station, workers earn on average $13.50 a day--three times Mexico's minimum wage--to insert metal tubes into a machine that then bends the tubes.
"We have significant usage today, and it's accelerating," says Jeff Dean, who led TensorFlow's design and heads Google's core artificial- intelligence research group. Inside Google, TensorFlow powers products such as the Google Translate mobile app, which can translate a foreign menu in front of your eyes when you point your phone at it. He originally built his startup Matroid, which helps companies create image recognition software, around a competing tool called Caffe, but he dumped it after trying TensorFlow. Since Google released TensorFlow, its competitors in cloud computing, Microsoft and Amazon, have released or started supporting their own free software tools to help coders build machine-learning systems.
There's a lot of excitement about robots advancing quickly thanks to machine learning. You said no one has lost a job to a Fetch robot, but over time, aren't products like yours going to displace human workers? I personally believe that we as a country should move toward universal basic income. I think of universal basic income as saying we all have the right to live our lives and experience the basic standards of life--and anything above that is yours to get.
Ya-Qin Zhang, president of China's leading search business, says Chinese companies can become worldwide Internet powerhouses too. One of the biggest fish in China's market of 730 million Internet users, Baidu is trying to open new revenue streams both domestically and abroad by investing heavily in artificial intelligence. The company employs more than 1,700 AI researchers, including some at a Silicon Valley research center opened in 2014, and was chosen by the Chinese government to run a new national lab intended to make the nation more competitive in machine learning. Why would top talent in artificial intelligence want to work for Baidu's U.S. lab over that of an American company?
Optogenetics was developed by Stanford neuroscientist and psychiatrist Karl Deisseroth, and it represented a breakthrough in efforts to determine the role of specific parts of the brain. Tye's research may eventually help open the door to drugs that affect only specific neural circuits, reducing anxiety with fewer side effects. When mice learned to associate a sound with a treat of sugar, she found stronger synaptic input to the neurons in the basolateral amygdala that were sending information to the nucleus accumbens, which is part of the brain's reward circuitry. On the other hand, when mice learned to associate the sound with mild electric shocks to their feet, input signals grew stronger in circuits leading from the basolateral amygdala to the centromedial amygdala, which is involved in pain and fear.
It's revolutionized seemingly non-digital industries--think of how different financial services, for instance, are today from what they were two decades ago--and investors expect it to soon transform others, which is why Tesla Motors is worth more than General Motors despite making a tiny fraction as many cars as GM makes and earning a tiny fraction of the revenue. But it positioned the company to benefit from the network effect: having third-party sellers made Amazon more appealing to customers, which in turn made it more appealing to sellers, creating a virtuous cycle for the company. Beyond the network effects is another, related way that the sheer scale of the Big Five helps them stay on top: through the access they have to enormous amounts of user data. That data, which is far more detailed and granular than anything companies have been able to access in the past, helps these companies improve their products and services, which in turn helps them add more users, which gives them access to more data, and so on.