Welcome to the December 1, 2017 edition of ACM TechNews, providing timely information for IT professionals three times a week.
ACM TechNews mobile apps are available for Android phones and tablets (click here) and for iPhones (click here) and iPads (click here).
To view "Headlines At A Glance," hit the link labeled "Click here to view this online" found at the top of the page in the html version.
The online version now has a button at the top labeled "Show Headlines."
|
|
|
AI Is Dreaming Up New Kinds of Video Games
Technology Review
Simon Parkin
November 29, 2017
Michael Cook at Falmouth University in the U.K. has developed Angelina, an artificial intelligence (AI) that can imagine new video games from scratch, and which has produced hundreds of experimental games since 2011. Angelina can generate games from images it pulls from license-free depositories, as well as building out premises and rules with characters and ideas borrowed from online newspapers or social media. Cook says this information is written to a text file that can be run by a standalone application. He notes Angelina "tries to build games that match its notion of what a good game is." In addition, the more interesting things Angelina uncovers via its experimentation, the more effort it commits to a project. Angelina is one example of researchers tapping increasingly sophisticated machine-learning methods to design games, and some envision a time when game artists could provide only the vaguest sketches of game elements for AIs to flesh out.
*May Require Free Registration
|
Robots Could Soon Replace Nearly a Third of the U.S. Workforce
The Washington Post
Danielle Paquette
November 30, 2017
The growth of automation will force as many as 70 million workers in the U.S. to find new jobs by 2030, according to a new McKinsey Global Institute study, which predicts nearly a third of the American workforce could soon need to develop new skills or change their field of work. The study found the jobs most at risk involve repetitive tasks, and notes about half the duties workers handle globally could be automated, although less than 5 percent of occupations could be entirely taken over by computers. The study also found that the need for people doing "predictable physical work," such as construction equipment installation and repair, car detailing, security guarding, and dishwashing and food preparation will fall by 30 percent. Around the world, other advanced economies such as Germany and Japan will see at least a third of their workforce similarly disrupted, while China's workforce decline will be around 12 percent.
*May Require Free Registration
|
Can a Zebra Crossing Change Its Stripes?
CNN
Andrea Lo
November 30, 2017
The Starling Crossing is an interactive crosswalk that responds dynamically to its environment, and it could be the future of how humans interact with cities. The Starling Crossing keeps the familiar "zebra" pattern, but it is able to change its layout, size, and color on demand because the markings emerge from a 23-meter by 7.5-meter waterproof network of LED lights embedded in the street. The system also includes two cameras positioned at opposite ends of the road, which are programmed to take about 25 images of the street per second. A computer processes these images, distinguishing between pedestrians, cyclists, and cars, and the system uses this information to decide how to behave. "One of the principles of an interactive crossing is that it should be able to learn over time the way that people use it," says Starling Crossing developer Usman Haque.
|
New #AskAbout AI Campaign Encourages Educators to Prepare Students for the Future Use of Artificial Intelligence
Education Dive
Katie Vander Ark; Jessica Slusser
November 30, 2017
The Ask About AI: The Future of Work and Learning campaign (#AskAboutAI), powered by Getting Smart and sponsored by eduInnovation, came out of a multi-year research project investigating the implications and influence artificial intelligence (AI) will have on young people in the future. The #AskAboutAI campaign includes a range of multimedia resources such as blog posts, articles, podcasts, videos, and a white paper. The campaign focuses on three main topic areas: employment, including the impact on future labor markets and required competencies; ethics and identifying the social and civic implications of exponential technology; and education, including providing advice to educators, parents, and policymakers on the knowledge, skills, and dispositions likely to be important in the automation economy. "We spent the year traveling the world talking with professionals about the future of AI and have concluded it will have a significant impact on employment, learning, ethical issues, and more," says Getting Smart CEO Tom Vander Ark.
|
Two New Simulators Tease Future of Quantum Computing
IEEE Spectrum
Jeremy Hsu
November 29, 2017
Two separate U.S. research groups have demonstrated some of the largest quantum simulators ever constructed, which could pave the way for a universal quantum computer. A team led by Harvard University used lasers to tweeze a 51-quantum bit (qubit) array of Rydberg atoms, while a second team from the University of Maryland and the U.S. National Institute of Standards and Technology built a 53-qubit trapped ion device by having electric fields control a string of charged atoms. The Harvard group first shined 101 lasers on a "vapor of rubidium atoms," which were configured into a perfect atomic array and then converted into qubits with additional lasers. The other team's trapped ion quantum simulator harnessed charged ytterbium atoms in single file. "As we learn to better control our systems--both Rydberg simulators and ion simulators--we will be able to improve our coherence properties and perform more sophisticated simulations," says Harvard's Ahmed Omran.
|
Dogs Get the Hollywood Treatment to Make Animal Animations More Realistic
University of Bath
Vicky Just
November 28, 2017
Researchers at the University of Bath's Center for Analysis of Motion, Entertainment, Research & Applications (CAMERA) in the U.K. are developing a new technique that will use human movements to drive a four-legged animal character. "What we want to do is to look at the movements of the human actor and then use a kind of translator to look at a library of real animal data to make the character on the screen move in a realistic way," says CAMERA's Martin Parsons. The dogs wear coats with reflective markers attached to them, then infrared light hits the markers and special cameras record their positions. The researchers say this data can be used to reconstruct the movement of the dogs on the computer screen. They note the animal movement data also will be used in collaborative research and development projects with industrial partners to drive next-generation tools and processes in the visual effects and games industries.
|
The Race for the Fastest Supercomputer
Government Computer News
Matt Leonard
November 29, 2017
The latest list of the 500 fastest supercomputers awarded the top spot to the Sunway TaihuLight in China's National Supercomputing Center in Wuxi, marking the fourth consecutive time that system has held the number one ranking in the semi-annual report. The Sunway TaihuLight system is still significantly more powerful than the other systems on the list, according to University of Tennessee, Knoxville professor Jack Dongarra. The TaihuLight has more than 10 million cores and over 1 million GB in memory, enabling it to reach 93,014.6 TFlop/s on the Linpack performance test. Meanwhile, the U.S. has the fewest machines on the global list of high-performance computers since the rankings began 25 years ago, with Oak Ridge National Laboratory's (ORNL) Titan system ranking highest at number five. Dongarra says ORNL's Summit and Lawrence Livermore National Laboratory's Sierra supercomputers, both of which are scheduled to go online next year, could place the U.S. back in the top spots.
|
UMass Amherst Computer Scientist and International Team Offer Theoretical Solution to 36-Year-Old Computation Problem
UMass Amherst News
Janet Lathrop
November 28, 2017
Researchers at the University of Massachusetts Amherst, the Massachusetts Institute of Technology (MIT), the Max Planck Institute in Germany, and the Dalle Molle Institute for Artificial Intelligence Research in Switzerland say they have discovered a theoretical solution to a 36-year-old problem in RNA folding predictions, a breakthrough that will be extremely useful in understanding genome sequences. In 1980, microbiologist Ann Jacobson published an algorithm for predicting the secondary structure of single-strand RNA, but since then the algorithm's cubic running time has not been improved. The researchers have theoretically demonstrated a faster subcubic algorithm possible for RNA folding computations. They developed the algorithm using a special kind of matrix multiplication that reduced the running time from three times the length of the base pair string to 2.82 times. MIT's Virginia Vassilevska Williams says the new solution "could potentially lead to further breakthroughs in our understanding for fundamental problems such as shortest paths in networks, pattern recognition, and so on."
|
Algorithms Can Determine a Neighborhood's Political Leanings by Its Cars
Stanford News
Andrew Myers
November 28, 2017
Researchers at Stanford University have developed algorithms that can determine where a neighborhood's political allegiances lie by analyzing the cars on its streets derived from publicly available Google Street View images. The algorithms trained themselves to identify the make, model, and year of every car produced since 1990 in more than 50 million Google Street View images from 200 U.S. cities. They then compared this information against the American Community Survey demographic database and against presidential election voting data to calculate demographic factors. The algorithm sorted the cars in all of the images into 2,657 categories by make, model, and year in only 14 days, whereas it would take 15 years for a human working at a rate of six images a minute to complete the job. Stanford professor Fei-Fei Li says this research "opens up more possibilities of virtually continuous study of our society using sometimes cheaply available visual data."
|
Old Methods May Not Measure Up to 3D Body Scans
GCU Today
Lana Sweeten-Shults
November 28, 2017
Researchers at Grand Canyon University (GCU) have developed a three-dimensional (3D) scanner that links to an iPad, and are using it to capture body images so they can wirelessly monitor and dashboard a body's measurements. The team says the goal is to detect changes in someone's body as a way to keep track of their health. The researchers took the 3D-scanning iPad and walked around a subject, scanning the middle of the body and then moving the scanner from the top of the head to the bottom of the feet. "We apply our algorithms and magic and whatnot to it to make sense of it, and then the physiology is where it's applied," says GCU's Kevin Hoskins. In addition, the researchers detected the segments of a body using machine learning. "We are using an algorithm that can detect certain patterns, and we are defining certain characteristics of areas in the body automatically," notes GCU's Isac Artzi.
|
Child-Proofing the Internet of Things
Lancaster University
Ian Boydon
November 28, 2017
Researchers at Lancaster University in the U.K. are studying how children can maintain their safety and privacy as they use the Internet of Things (IoT). The researchers previously developed the BBC Micro:Bit, a microcontroller with built-in sensors, Bluetooth connectivity, and pins that enable it to be connected to external devices. The team is now working on the year-long Child Proofing the Internet of Things project to address privacy issues and learn how children might use IoT devices. The project aims to discover the privacy and security challenges arising from children using IoT devices, as well as what design and programming considerations are needed to provide greater protection, and what guidelines and advice are needed for children, their families, and teachers for programming IoT devices. "By working closely with child protection experts, this research will help provide a much richer understanding of the potential implications that may arise with children and IoT," says Lancaster University's Bran Knowles.
|
High-Speed Quantum Encryption May Help Secure the Future Internet
Duke Today
Kara Manke
November 27, 2017
Researchers at Duke University, Ohio State University, and Oak Ridge National Laboratory have developed a system that can create and distribute encryption codes at megabit-per-second rates five to 10 times faster than existing methods and about equal to current Internet speeds when running several systems in parallel. In addition, the researchers demonstrated that the new technique is secure from common attacks, even with equipment flaws that could present leaks. The system also could be used to stop hackers from using quantum technology to crack even the toughest of standard Internet security codes. Like other quantum key distribution systems, the system uses a weakened laser to encode information on individual photons of light, but the new technique enables more information to be packed onto each photon. By adjusting the time at which the photon is released and its phase, the system can encode two bits of information for every photon instead of one.
|
The AI Guru Behind Amazon, Uber, and Unity Explains What AI Really Is
Fast Company
Sean Captain
November 29, 2017
In an interview, Danish computer scientist Danny Lange, who has invented machine-learning platforms for technology giants such as Uber and Amazon, addresses the subtleties between true artificial intelligence (AI) and technology that follows human scripts. Lange says AI can be defined in both an external and an internal aspect, with the former oriented around perception. He contends the internal aspect is being disrupted as AI transitions from being programmed to becoming able to learn. "When that comes together, I feel that we are crossing a line, and we start dealing with something that is truly AI," Lange says. He also cites a feedback loop playing an important role in AI evolution, using as an example separate machine-learning systems that respectively detect fake news and produce fake news. "As one of them gets better at detecting fake news, the opponent gets better at generating fake news, because it learns from the feedback loop," Lange says.
|
|
