Today's store-bought drones are remarkably easy to fly, thanks to features like self-stabilization technology, obstacle avoidance sensors and so on. You could walk out of a shop, charge up your batteries and be airborne for the first time all within a single afternoon. But as the video compilation above shows, it's probably still a good idea to get some practice in before attempting any particularly tricky stunts. Even if drones have all sorts of high-tech features designed to keep them airborne, they aren't impervious to the constant pull of earth's gravity, the branches of an unseen tree, or even the grasp of a curious animal. Watch the video above to see a selection of drone crashes from the aircraft's perspective.

Researchers have come up with all sorts of ways to propel tiny robots deep into the human body to perform tasks, such as delivering drugs and taking biopsies. Now, there's a nanorobot that can clean up infections in blood. Directed by ultrasound, the tiny robots, made of gold nanowires with a biological coating, dart around blood, attach to bacteria, and neutralize toxins produced by the bacteria. It's like injecting millions of miniature decoys into blood to distract an infection from attacking the real human cells. The invention, developed in the labs of Joseph Wang and Liangfang Zhang at the University of California San Diego (UCSD), was described today in Science Robotics.

Intuitive Surgical (ISRG) started with a very simple plan. That is, to make surgery less invasive with surgical robots. For laypeople in 1995, the concept was science fiction. At the time, researchers at the Stanford Research Institute had been kicking around the idea for years. That's because the U.S. Army had hired them in the late '80s to make remote battlefield surgery feasible.

When dermatologists received more clinical information and images at level II, their diagnostic performance improved. However, the CNN, which was still working solely from the dermoscopic images with no additional clinical information, continued to outperform the physicians' diagnostic abilities,

Back in March, an Uber self-driving car killed 49-year-old Elain Herzberg in Tempe, Arizona, after failing to do an emergency stop. After a US federal investigation, it is thought that the car did not stop because the system put in place to carry out emergency stops in dangerous situations was disabled . National Transportation Safety Board officials inspecting the car that killed Mrs Herzberg. So, how do self-driving cars actually work? Most self-driving cars have a GPS unit, a range of sensors such as radar, video and laser rangefinders as well as a navigation system.

A couple in Portland, Ore., discovered that their Amazon Echo had recorded their conversation and sent it to one of their contacts. A couple in Portland, Ore., discovered that their Amazon Echo had recorded their conversation and sent it to one of their contacts. As secret recordings go, the Portland couple's conversation was pretty mundane: They were talking about hardwood floors. But their Amazon Echo was listening and recording their discussion. The device then sent the recording to someone in their contacts -- without the couple's knowledge.

Robots that train themselves in battle tactics by playing video games could be used to mount cyber-attacks, the UK military fears. The warning is in a Ministry of Defence report on artificial intelligence. Researchers in Silicon Valley are using strategy games, such as Starcraft II, to teach systems how to solve complex problems on their own. But artificial intelligence (AI) programs can then "be readily adapted" to wage cyber-warfare, the MoD says. Officials are particularly concerned about the ability of rogue states and terrorists to mount advanced persistent threat attacks, which can disable critical infrastructure and steal sensitive information.

Ricardo Baeza-Yates ([email protected]) is Chief Technology Officer of NTENT, a search technology company based in Carlsbad, CA, USA, and Director of Computer Science Programs at Northeastern University, Silicon Valley campus, San Jose, CA, USA.

Over the past decade, advances in deep learning have transformed the fortunes of the artificial intelligence (AI) community. The neural network approach that researchers had largely written off by the end of the 1990s now seems likely to become the most widespread technology in machine learning. However, protagonists find it difficult to explain why deep learning often works well, but is prone to seemingly bizarre failures. The success of deep learning came with rapid improvements in computational power that came through the development of highly parallelized microprocessors and the discovery of ways to train networks with enormous numbers of virtual neurons assembled into tens of linked layers. Before these advances, neural networks were limited to simple structures that were easily outclassed in image and audio classification tasks by other machine-learning architectures such as support vector machines.

Sensor technology is designed to allow machines to interact with real-world inputs, whether they are humans interacting with their smartphones, autonomous vehicles navigating on a busy street, or robots using sensors to aid in manufacturing. Not surprisingly, three-dimensional (3D) sensors, which allow a machine to understand the size, shape, and distance of an object or objects within its field of view, have attracted a lot of attention in recent months, thanks to their inclusion on Apple's most-advanced (to date) smartphone, the iPhone X, which uses a single camera to measure distance. Indeed, the TrueDepth system, which replaces the fingerprint-based TouchID system on the Apple handset, shines approximately 30,000 dots outward onto the user's face. Then, an infrared (IR) camera captures the image of the dots, which provides depth information based on the density of the dots (closer objects display a dot pattern that is spread out, whereas objects that are farther away create a denser pattern of dots. Altogether, the placement of these dots creates a depth map with 3D data that is used to supply the system with the information it needs to check for a facial identity match, which then unlocks the device.