Highlights for the week of 10 April 2017
Are we infecting machine learning algorithms with our human biases? A new study says yes. Apparently, an AI learning from existing English language texts will reflect the biases in those texts back at us. The researchers found that the AI was more likely to associate European-American names rather than African-American names with “pleasant” stimuli. It also tended to associate “woman” and “girl” with the arts rather than with mathematics.
Netflix has been protecting video streams with HTTPS encryption since the summer of 2016. But new research indicates that this strategy is not sufficient to keep third party service providers and motivated attackers from getting a peek at what I’m watching.
Previously, graphene photodetectors could achieve large area and position-sensitive detection if they were built into a dense array of many graphene pixels. But it would come at the cost of fabrication complexity. In a new photodetector design, the light interaction and absorption occur on the device’s substrate. This approach allows even a single graphene device to achieve large response area and position sensitivity.
According to documents filed with the U.S. Federal Communications Commission, General Motors and its autonomous technology subsidiary, Cruise Automation, plan to equip 300 self-driving Chevrolet Bolts with short- and medium-range radars. The Bolts, which would hit the roads in San Francisco, Detroit and Scottsdale, Arizona, would be the largest fleet of self-driving cars in the world.
Over the past week or so, two new methods of soft robot locomotion have shown up in the news: One of them using external magnetic fields, and the other using an electric field to power flapping fins.
Source: Spectrum Tech Talk 5 Things You Missed: AI Algorithms Are Biased, They Know What You’re Watching, and More