Welcome to the September 24, 2018 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."
|
|
|
Over 2,000 European AI Experts Join Hands to Challenge U.S., China in Artificial Intelligence
South China Morning Post
Alice Shen
September 21, 2018
About 2,100 artificial intelligence (AI) researchers from 29 European countries have announced a joint research alliance and are soliciting funding from the European Union (EU) to counter progress in such research in the U.S. and China. The Confederation of Laboratories for AI Research in Europe (CLAIRE) is calling on the European Commission to deploy an EU-wide AI strategy similar to U.S. and Chinese initiatives, to cultivate and retain AI talent by establishing research and development centers that support "Google-scale" infrastructure. The German Research Center for Artificial Intelligence Research's Philipp Slusallek says, "Europe has a very strong tradition in AI and was also among the leaders in AI in the past." CLAIRE says European progress in AI technology is highly reliant on government funding, as such research is not driven by the private sector to the same degree it is in China and the U.S.
|
Spack, a Lab-Developed 'App Store for Supercomputers,' Becoming Standard-Bearer
HPCwire
September 20, 2018
Researchers at Lawrence Livermore National Laboratory (LLNL) have developed the Supercomputer PACKage manager (Spack), an open source package manager optimized for high-performance computing (HPC). Spack is one of the most popular pieces of software the lab has ever released to the GitHub open source community, becoming the go-to package manager at LLNL, as well as at Argonne, Oak Ridge, Los Alamos, and Sandia national laboratories. Spack also is being used at Lawrence Berkeley Laboratory's National Energy Research Computing Center, Oak Ridge's Summit, and LLNL's Sierra computing facilities; it is the official deployment tool for the Exascale Computing Project. Through GitHub, Spack attracted hundreds of users who have added about 2,800 software packages. LLNL researcher Todd Gamblin says, "The dream is to take the grunt work out of HPC: users get on a machine, assemble a stack of hundreds of libraries in minutes, then get back to focusing on the science."
|
'Robotic Skins' Turn Everyday Objects into Robots
YaleNews
William Weir
September 19, 2018
Yale University researchers have developed "Robotic Skins," allowing users to turn inanimate objects into robots. The skins are made from elastic sheets embedded with sensors and actuators. When these sheets are placed on a deformable object, the skins animate the object from its surface, allowing the makeshift robot to perform different tasks depending on the properties of the soft object and how the skins are applied. In addition, using more than one skin at a time allows for more complex movements. The robotic skins could be used for everything from search-and-rescue robots to wearable technologies. Yale professor Rebecca Kramer-Bottiglio said she came up with the idea for the devices a few years ago, when the U.S. National Aeronautics and Space Administration put out a call for soft robotic systems. “One of the main things I considered was the importance of multifunctionality, especially for deep space exploration where the environment is unpredictable.”
|
Artificial Neural Network Now Capable of Finding Medication Complaints in Social Networks
Kazan Federal University (Russia)
September 17, 2018
Researchers at Russia's Kazan Federal University (KFU) and the Moscow Institute of Physics and Technology have developed recurrent neural networks using semantic vector word representation to match medical complaints such as "cannot fall asleep" or "a little giddy" in social networks with formal medical terms such as insomnia or vertigo. The researchers uploaded medical texts to the software platform to create a special vocabulary. The software used this data to assign a vector to each word. More robust intelligent text analysis for patient complaints in social networks can impact human understanding of how medications affect patients. Andrey Filchenkov, a research associate at Russia’s ITMO University, said, “Recurrent neural networks work well with serialized data because they can find links between elements while taking consideration of the context.”
|
Meet the Winner of Robotics' World Cup
Technology Review
September 20, 2018
NimbRo, a two-legged robot developed at the University of Bonn in Germany, won the soccer World Cup for adult-size robots this summer in Montreal, Canada. NimbRo is 135 centimeters tall and weighs almost 40 pounds. The robot's exoskeleton is made up of a few simple parts which are three-dimensionally (3-D) printed from nylon. NimbRo is equipped with a vision system that performs object detection, robot localization, task planning, and actuator control. These functions give the robot a number of soccer skills, such as being able to locate and approach the ball, avoid obstacles, and kick and dribble. In five games during the competition, NimbRo scored 46 goals while allowing only one; the robot also accumulated 21 points in the technical challenges, giving it a clear victory.
|
Combining Multiple CCTV Images Could Help Catch Suspects
University of Lincoln
Sophie Belcher
September 21, 2018
Researchers at the universities of Lincoln and York in the U.K., and the University of New South Wales in Australia, have generated a series of pictures via "face averaging," which digitally blends multiple images into a single image free of variants so only features indicating a subject's identity remain. The team compared the effectiveness of humans and computer facial-recognition systems at recognizing people from high-quality images, pixelated images, and face averages. They determined that people and computers could both identify a face better when viewing an average image that combined multiple pixelated images, versus the original poor-quality images. Computers benefited from collectively averaging multiple images that were already of high quality, and in some cases achieved 100% accurate face recognition. The researchers say these techniques could enhance law enforcement and security operations by delivering a standardized way of using images captured from multiple closed-circuit TV cameras.
|
Machine-Learning System Tackles Speech and Object Recognition, All at Once
MIT News
Rob Matheson
September 18, 2018
Researchers at the Massachusetts Institute of Technology (MIT) have developed a system that can learn to identify objects within an image, based on a spoken description of the image. When provided with an image and an audio caption, the system can highlight in real-time the relevant regions of the image being described. The system learns words directly from recorded speech clips and objects in raw images, and associates them with one another. The researchers trained the model on a total of 400,000 image-caption pairs, and held out 1,000 random pairs for testing. Said researcher David Harwath, “We wanted to do speech recognition in a way that’s more natural, leveraging additional signals and information that humans have the benefit of using, but that machine learning algorithms don’t typically have access to. We got the idea of training a model in a manner similar to walking a child through the world and narrating what you’re seeing.”
|
Women Tech Hiring Sees Minuscule Growth, Short of Hopes
The Wall Street Journal
Sara Castellanos
September 20, 2018
Women were represented in 24% of technical roles this year, with growth of 1.09% over last year, according to an AnitaB.org survey. About 13% of technical roles were filled by Black, Hispanic, Native American, Pacific Islander, and multiracial women, according to the report, which did not measure racial diversity last year. Cloud software firm Blackbaud's Mary Beth Westmoreland says workplace diversity improves products and helps businesses in many ways. Westmoreland says promoting diversity has become a priority for her. “I feel really strongly about elevating, assisting and paving the way.”
*May Require Paid Registration
|
Putting Underused Smart Devices to Work
IMDEA Networks
September 17, 2018
Researchers at IMDEA Networks Institute in Spain have developed DisCoEdge, a system that aims to transform the concept of device ownership to improve current utilities and create new services. DisCoEdge distributes major computational tasks and large storage over many simple devices at the "edge" of the Internet, such as home routers and mobile devices. This concept could lead to a new market that private and business users may join as if it were a social network marketplace, buying or selling the partial use of personal or industry devices to store information, run a program, or mine data, among other uses. IMDEA Networks researcher Antonio Fernández Anta says, "The key idea behind DisCoEdge is not to generate a market of free agents, but rather a platform that acts as a market broker, an intermediary, and provides guarantees."
|
Capitalizing on Sleep-Wake Cycle Can Drastically Increase Digital Ad Profits From Social Media, Study Shows
Notre Dame News
Shannon Roddel
September 17, 2018
University of Notre Dame researchers have shown that digital content platforms can increase traffic to their websites from social media by aligning their posting schedules with target audiences' sleep-wake cycles. The researchers interviewed social media managers from several major content platforms to learn how they make posting decisions, and examined a year’s worth (5,700) of Facebook posts and boosting (paid advertising in support of posts) data from a major newspaper on the West Coast. The researchers found social media managers relied on intuition in making posting decisions, without considering the emotions that posts might elicit from readers. Using this data, the researchers developed an algorithm to help social media managers determine the optimal timing for posting content, which posts to boost (with paid advertising), and how many to boost per day to stay within budget. Notre Dame University's Vamsi Kanuri said the algorithm "can help firms make profit maximizing content scheduling decisions."
|
|
