介绍:
TI②
Johnny:
Google self-driving car hits a bus, Google bears the responsibility
Google’s self-driving car initiative is one of the most ambitious and controversial projects currently under the scrutiny of the public eye. Although the search giant has had six years of experience under their belt with only 17 minor accidents occurring across over two million miles of test driving, this month marks the first time that a self-driving vehicle bore any culpability in a traffic incident.
On Valentine’s Day, February 14, 2016, a self-driving Lexus RX450h attempted to navigate around some sandbags placed in a wide lane. A bus traveling at 15 mph approached the Google car from behind, occupying same wide lane as the smart car. The Lexus attempted to re-enter the center of the lane while moving at a speed of 2 mph.
Google reported in a statement last week that the smart car anticipated that the bus would slow down to allow the self-driving car to continue, but the bus did not. When the smart car re-entered the center of the lane, it struck the side of the bus causing minor damage to the front left fender, front wheel, and a sensor. There were no injuries, and no police report was filed.
Iris:
In spite of the lack of police involvement, which usually determines culpability, Google says they are at least partially responsible for the incident. Their traffic algorithms take into account the common practice of vehicles slowing down to accommodate the merging of others. It would appear that buses and similar large vehicles are less likely to slow down in these situations, and Google says they are making adjustments based on this and “thousands of variations on it in our simulator,” to make their vehicles “more deeply understand that buses (and other large vehicles) are less likely to yield to us than other types of vehicles.”
Google is taking a stance that this is all a part of the learning process, but even extremely small incidents like this can be damaging to the development of self-driving cars. Although many futurists believe that this revolution is inevitable, public fear and concerns that feed on smart-car culpable accidents may seriously delay this technology becoming mainstream.
Tim：
AI learns to predict human reactions by reading our fiction
It's certainly easier than hand-coding a sense of morality.
A team of Stanford researchers have developed a novel means of teaching artificial intelligence systems how to predict a human's response to their actions. They've given their knowledge base, dubbed Augur, access to online writing community Wattpad and its archive of more than 600,000 stories. This information will enable support vector machines (basically,learning algorithms) to better predict what people do in the face of various stimuli.
"Over many millions of words, these mundane patterns [of people's reactions] are far more common than their dramatic counterparts," the team wrote in their study. "Characters in modern fiction turn on the lights after entering rooms; they react to compliments by blushing; they do not answer their phones when they are in meetings."
Johnny:
In its initial field tests, using an Augur-powered wearable camera, the system correctly identified objects and people 91 percent of the time. It correctly predicted their next move 71 percent of the time.
This isn't the first time developers have turned to books to teach computers, mind you. Facebook just this week handed its AI a 1.96GB stack of children's books in hopes of teaching it a similar lesson.
Mercedes replaces robots with people on its assembly line
Robots can't keep up with all the custom options.
We've all heard stories about how the robots are taking our jobs. While that may be true, Mercedes-Benz is replacing some of its high-tech workers with real live humans. As it turns out, robots can't keep up with the degree of customization that the automaker offers on its S-Class sedans. To be fair, there are four different types of caps for the tires alone, not to mention options for carbon fiber trim and temperature-controlled cup holders.
Iris:
"Robots can't deal with the degree of individualization and the many variants that we have today," said the company's head of production Markus Schaefer. "We're saving money and safeguarding our future by employing more people."
The change comes at a time when a number of companies are replacing people with robotic devices. International Federation of Robotics (IFR)released a survey today stating that 1.3 million of the industrial robots will be in use by 2018. In fact, the number of those machines rose 43 percent in a year between 2013 and 2014. However, as Mercedes continues to expand the options available on its vehicles, the robots aren't able to adapt to new tasks. They're better suited for doing the same jobs repeatedly.
"The variety is too much to take on for the machines," Schaefer explained. "They can't work with all the different options and keep pace with changes."
Tim:
Robots aren't getting the boot from Mercedes-Benz Sindelfingen factory entirely, though. The machines will work alongside humans instead of being confined behind glass. Mercedes, BMW and Audi are all working on sensor-packed robots that can operate safely alongside their living breathing colleagues.
Australia uses drones to spot sharks and rescue their victims
Rescue authorities in Australia are using drones to identify sharks and provide assistance to swimmers threatened by them. The camera-equipped drones were launched this week as part of a trial program, and will patrol the coastline of the Australian state of New South Wales (NSW). There were 14 shark attacks in NSW last year, including one fatality, and state authorities have been supporting the development of new technology to prevent future attacks.
The $180,000 drone trialled this week, known as the "Little Ripper," has a high-definition camera that relays aerial footage to computers on the ground. It also contains a rescue pod that drops down in the event of an emergency. As the Sydney Morning Herald reports, the pod includes medical equipment, shark repellant, and an inflatable raft that can carry up to three people. The battery-powered Vapor 55 drone can stay in the air for up to 150 minutes, according to the Daily Telegraph, and researchers are working on software that would allow the drone to automatically identify shark type.
Johnny:
"The Little Ripper is the new, high-tech eye in the sky," said Brian Hartzer, chief executive of Westpac, which is funding the trial. "There are 17 Westpac helicopters around Australia and we hope this is going to work really well and become another very welcome sight around the coastline." If the trial is successful, about 40 other drones will be deployed across the area.
