Tuesday 17 June 2014

A Smartphone Keyboard App That Anticipates What You Want to Type

Typing on the glass of a smartphone can be a headache. The keyboards are tiny. Errors are easy to make, even on the simplest of words. And auto-correct, the function meant to fix typing mistakes, often introduces errors, switching words that have little to do with what people actually mean.
Consumers’ irritation with typing has enticed the likes of Google, Apple and several start-ups to try to put the smart into smartphone keyboards  knowing that there is a big opportunity in having such a prominent place in users’ hands.
Perhaps no company, though, has found more success with enhancing keyboards than SwiftKey, a London start-up that first released a keyboard app in 2010. SwiftKey says the app, which cost $4 until last Wednesday, when it became available free, was downloaded more than 30 million times. It ranked as one of the most popular paid smartphone applications on Google’s Play Store.
The app predicts what you will type next by using artificial intelligence and reams of online data, including what you type on other apps, to learn what words you typically use. It then applies that information to predict what you will type next.

As Robotics Advances, Worries of Killer Robots Rise

From driver-less cars to delivery drones, a new generation of robots is about to revolutionize the way people work, drive and shop. But there is one area where robots are already entrenched and spreading fast: the industrial sector, especially manufacturing and storage.
Robots have long toiled alongside workers in factories and warehouses, where they load boxes with items ordered online, drill and weld car parts, or move food from one conveyor belt to the next.
Now many experts worry about the dangers that robots pose to the humans who work alongside them.
Robots have caused at least 33 workplace deaths and injuries in the United States in the last 30 years, according to data from the Occupational Safety and Health Administration. That may not sound like many, but the number may well understate the perils ahead.
Unlike today’s robots, which generally work in cages, the next generation will have much more autonomy and freedom to move on their own.
“In order for robots to work more productively, they must escape from their cages and be able to work alongside people,” said Kent Massey, the director of advanced programs at HDT Robotics. “To achieve this goal safely, robots must become more like people. They must have eyes and a sense of touch, as well as the intelligence to use those senses.”
Until now, robots have largely been used in manufacturing, particularly in the auto industry. They have mostly been “dumb robots,” designed for repetitive tasks that are dirty, dangerous or dull.
Regulations have required that the robots operate separately from humans, in cages or surrounded by light curtains that stop the machines when people approach. As a result, most of the injuries and deaths have happened when humans who are maintaining the robots make an error or violate the safety barriers, such as by entering a cage.
But the robots whose generation is being born today collaborate with humans and travel freely in open environments where people live and work. They are products of the declining cost of sensors and improved artificial intelligence algorithms in areas such as machine vision. Google’s newest driverless car, for instance, is completely automated, without a steering wheel or a brake pedal.
Along with the new, free-roaming robots come new safety concerns. People worry about what happens if a robot spins out of control, or the first time a driverless car kills someone.
“It’s the fear of robots,” said Bryant Walker Smith, a fellow at the Center for Internet and Society at Stanford Law School who studies driverless cars. “There’s something scarier about a machine malfunctioning and taking away control from somebody.”
As a result, these robots require extra protective measures. The Google car has a padded front to soften any blow if the robot or a human causes an accident. The windshield is plastic, and the front of the car is rounded so it is less likely to hurt or trap pedestrians or cyclists.
Another robot, Baxter, which does repetitive jobs in workplaces like packaging small items, is designed to sense humans and stop before coming in contact with them. It also has a display screen that cues those who are nearby about what the robot is focusing on and planning to do next.
If robots and humans are going to live and work together, Baxter and its progeny will need more of these advances. To develop them, the robots’ creators will need to draw on one of the most human of emotions: empathy.