Bytes: What’s new in tech
Unmasked by artificial intelligence
It won’t be long before facial recognition software can figure out who you are even if your face is covered up, said James Vincent in TheVerge.com. A group of researchers based in the U.K. and India say they’ve trained an algorithm to identify people even when they are wearing disguises. The results are “far less accurate than industry-level standards”; for instance, the system can correctly identify someone wearing a cap, sunglasses, and scarf only 55 percent of the time. But the research shows how quickly facial recognition technology is progressing, meaning “staying anonymous in public will be harder than ever before.” Facebook can already recognize people based on their hair, body shape, and posture.
Zello to the rescue
In the aftermath of Hurricane Harvey, a “digital walkie-talkie” named Zello “became the go-to app for rescuers working to save thousands of people trapped by floodwaters,” said Peter Holley in The Washington Post. The free app “relies on cellphone data plans or WiFi and is designed to operate in places where signals are weak,” helping volunteer groups coordinate search-and-rescue efforts. Volunteers monitoring social media used Zello to feed information to rescue boats patrolling flooded neighborhoods. As Hurricane Irma bore down on the Florida coast last week, Zello surged to the top download spots on Google Play and the Apple App Store. The app has 100,000 users worldwide, and is particularly popular with Hong Kong taxi drivers, as well as people in countries like Egypt and Venezuela “where government services struggle to meet demand.”
Watson has a lot to learn
IBM’s Watson supercomputer was supposed to revolutionize how cancer is treated, but so far, it’s “nowhere close,” said Casey Ross and Ike Swetlitz in StatNews.com. Watson for Oncology uses artificial intelligence to analyze huge amounts of data, including doctors’ notes, medical studies, and clinical guidelines, to recommend the best treatments. But the results have been underwhelming, according to interviews with doctors, IBM executives, and AI experts. Only a few dozen hospitals are using the system, which “is still struggling with the basic step of learning about different forms of cancer.” Watson also hasn’t been able to generate new insights or approaches, because its decisions are based on its training by a handful of U.S. doctors. That has also led to complaints of bias toward American medical practices.