Welcome to the November 18, 2016 edition of ACM TechNews, providing timely information for IT professionals three times a week.
Updated versions of the ACM TechNews mobile apps are available for Android phones and tablets (click here) and for iPhones (click here) and iPads (click here).
HEADLINES AT A GLANCE
K-Computer Leads Graph 500 List of World's Fastest Supercomputers
Inside HPC (11/16/16)
Japan's K-Computer is the world's fastest supercomputer, topping the latest Graph 500 list, which was released this week at SC16 in Salt Lake City, UT. The Graph 500 list, which is updated every six months, measures performance against three kernels: search, optimization, and edge-oriented. The rankings on these kernels are informative for application performance in a range of areas, including cybersecurity, medical informatics, social networks, data enrichment, and symbolic networks such as the human brain. "It is exciting to see the Graph 500 list evolve as we continue to push forward on large-scale data analytics and architectural challenges we face in developing memory and interconnects for these powerful machines," says Micron's Richard Murphy, cofounder of the Graph 500. The K-Computer, which was built by Fujitsu and is operated by Japanese research institute RIKEN, has topped the Graph 500 list since July 2015. Lawrence Livermore National Laboratory's Sequoia supercomputer ranked second on the Graph 500 list, and Argonne National Laboratory's Mira supercomputer ranked third. "In this new age of big data, we need to measure not just how quickly computers can chew on sets of numbers, but rather how quickly computers can build knowledge from massive-scale datasets," says the Georgia Institute of Technology's David Bader, who helped develop the Graph 500.
New AI Algorithm Taught by Humans Learns Beyond Its Training
U of T News (11/16/16) Marit Mitchell
Researchers at the University of Toronto say they have developed an algorithm that can learn directly from human instructions rather than a set of examples. Researchers usually provide neural networks with labeled data and teach the system how to make decisions based on the samples. With the new heuristic training model, humans program the algorithm with instructions that are used to classify training samples. Researchers Parham Aarabi and Wenzhi Guo trained their algorithm to identify people's hair in photographs. "Our algorithm learned to correctly classify difficult, borderline cases--distinguishing the texture of hair versus the texture of the background," says Aarabi. "What we saw was like a teacher instructing a child, and the child learning beyond what the teacher taught her initially." Using the new method, the algorithm outperformed conventional training techniques by 160 percent and outperformed its own training by 9 percent. The researchers say the heuristic method could be used to classify previously unlabeled data, such as cancerous cells for medical diagnostics or to classify objects surrounding a self-driving car.
ACM SIGHPC/Intel Computational & Data Science Fellowship Winners to Be Recognized at SC16
HPC Wire (11/16/16) Tiffany Trader
Intel and ACM SIGHPC, the special interest group for high-performance computing (HPC), announced the winners of their first Computational and Data Science Fellowship program over the summer, and the recipients formally recognized at the SC16 awards ceremony on Thursday. Established to increase the diversity of students pursuing computation and data science graduate programs, the fellowship provides financial assistance to students from underrepresented groups, including women and those with diverse racial and ethnic backgrounds. According to statistics from the U.S. Bureau of Labor, the computing industry has more than 200,000 unfilled jobs, with that figure expected to grow to about 1 million by 2022. A greater emphasis on workforce diversity is needed to ensure future HPC systems, software, and tools leverage the most diverse and creative ideas available, according to ACM vice president Cherri Pancake. "There is a growing realization that we cannot build the kinds of products that will help people address global challenges effectively without a workforce encompassing the broadest possible set of skills and life experiences," Pancake says. Recipients of the fellowship program were chosen from candidates who demonstrated potential as mentors and role models as well as expertise in their area of study. Eighty-five percent of the awardees are female, and 30 percent are from underrepresented minority groups in computing.
EU Project Developing Symbiotic Robot-Plant Biohybrids
IEEE Spectrum (11/15/16) Evan Ackerman
The European Union-funded Flora Robotica project aims to develop and investigate closely linked symbiotic relationships between robots and natural plants, and to explore the potential of a plant-robot society that can produce architectural artifacts and living spaces. The research team, which features scientists from six groups based in Poland, Denmark, Germany, and Austria, says they want to use robots to provide support and guidance to plants, so those plants will be healthier and more robust. In addition, the plant-robots will be somewhat trainable in that they will react to stimuli coming from the robots, enabling users to incorporate them into living, hybrid structures such as benches, walls, and roofs. "The robots are implemented as hardware modules that allow to implement an artificial growth process and to keep pace with the natural growth of the plants," according to the researchers. The researchers note the robot assemblies support the biological plants through appropriate scaffolding to help them maintain homeostasis, and the plants support and control the robots by guiding them through growth and support their weight in later growth phases.
Automated Pro-Trump Bots Overwhelmed Pro-Clinton Messages, Researchers Say
The New York Times (11/17/16) John Markoff
Researchers from Oxford University in the U.K. found that an automated army of pro-Donald J. Trump chatbots overwhelmed similar programs supporting Hillary Clinton five to one in the last few days of the U.S. presidential campaign. The chatbots would send messages on Twitter based on a topic, usually defined on Twitter by a word preceded by a hashtag symbol. The chatbots were used to rant, confuse people on facts, or simply muddy discussions, according to Oxford researcher Philip N. Howard. "The use of automated accounts was deliberate and strategic throughout the election," according to the researchers. Although they were unable to directly link the chatbots to either campaign, there was evidence the activity was part of an organized effort. "By the third debate, Trump bots were launching into their activity early and we noticed that automated accounts were actually colonizing Clinton hashtags," Howard says. The researchers based their study on a collection of about 19.4 million Twitter posts gathered in the first nine days of November. The researchers selected tweets based on hashtags identifying certain subjects and identified automated posting by finding accounts that post at least 50 times a day. This strategy of using deceptive social media campaigns has been dubbed "computational propaganda."
New Software Continuously Scrambles Code to Foil Cyberattacks
Columbia University (11/17/16) Kim Martineau
Columbia University researchers have developed Shuffler, software that tries to preempt hackers that exploit errors in software code by enabling programs to continuously scramble their code as they run, effectively closing the window of opportunity for an attack. "Shuffler makes it nearly impossible to turn a bug into a functioning attack, defending software developers from their mistakes," says Columbia researcher David Williams-King. Even after repeated debugging, software normally contains up to 50 errors per 1,000 lines of code, each of which is a potential avenue for attack. Shuffler maximizes address space layout randomization's (ASLR) code-scrambling approach by randomizing small blocks of code every 20 to 50 milliseconds. "By the time the server returns the information the attacker needs, it is already invalid--Shuffler has already relocated the respective code snippets to different memory locations," says Columbia professor Vasileios Kemerlis. The researchers say Shuffler runs faster and requires fewer system changes than similar continuous-randomization software. On computation-heavy workloads, Shuffler slows programs by 15 percent on average, but at larger scales the drop in performance is negligible. The researchers want to make Shuffler easier to use on software they have not yet tested, and they want to improve Shuffler's ability to defend against exploits that take advantage of server crashes.
Women in Tech Earn $0.94 Cents to Every $1 Men Make
TechRepublic (11/15/16) Alison DeNisco
The U.S. technology sector's 5.9 percent gender pay gap is greater than the national pay gap of 5.4 percent, according to a new Glassdoor study, which means U.S. women in tech earn an average of 94 cents for every $1 men earn. "Even when factors such as age, job title, and location are controlled for there is still an unexplainable pay gap between women and men--and that isn't acceptable," says Glassdoor CEO Robert Hohman. Computer programmers had the largest wage gap with a 28.3-percent disparity between men and women, followed by game artists, information security specialists, software architects, and search engine optimization strategists. Hohman says pay transparency could spur employers and employees to close pay gaps. Stanford University professor Shelley Correll partly blames the national gender pay gap on occupational sorting, with men and women employed in different jobs and industries. Correll notes once they have entered the tech industry, women can find themselves diverged from technical careers into areas such as marketing, human resources, and project management, which have fewer opportunities for promotion or wage increases. She also says unconscious prejudices may make managers doubt women's technical prowess.
Most Updates to Mobile Apps Don't Make a Noticeable Difference
University College London (11/16/16) Chris Lane
New research from University College London (UCL) in the U.K. assesses the success of updates to mobile apps by examining user ratings and the frequency of ratings. The UCL researchers analyzed 26,339 app updates released on Google Play over a period of 12 months, from 14,592 apps that appeared in one of Google Play's "Top 540" lists at least once in the preceding year. The researchers found 40 percent of releases involving paid apps impacted subsequent user ratings, compared to only 31 percent of updates of free apps. However, free apps had more positive effects among the impactful updates. Among paid apps, the most successful releases came from the most expensive apps. The analysis also found the most impactful releases offer new functionality and more descriptive release texts. For popular apps that are not in the top echelon, users responded particularly well to bug fixes. "There's a culture among app developers of getting more releases out than your competitors, but our research suggests they should think more carefully before putting out a release, as it might bear very little benefit," says UCL professor Mark Harman.
This App Lets You Control Your Phone Using Sonar
Technology Review (11/15/16) Jamie Condliffe
Researchers Wei Wang and Alex X. Liu from China's Nanjing University have built software that enables users to directly control the functions of smartphones via sonar. The software triggers sound emissions from the phone's onboard speakers to measure the proximity of an object, such as the user's hand, to within 4 millimeters by analyzing reflected signals picked up by microphone. Wang says the phone can detect and process movement in 15 milliseconds. The researchers plan to transform the app into an application-programming interface that other developers can use to embed the echolocation system into apps on other handhelds. Both researchers believe this functionality can be added to any smartphone without additional hardware. Wang notes hardware makers could optimize the placement of microphones and speakers on existing devices, as well as boost the upper frequency at which devices transmit and receive sound to realize submillimeter resolutions. Wang and Liu also aim to devise versions for smart watches and virtual reality headsets. They already say they can track the motion of a hand accurately enough to identify characters being typed with more than 90-percent precision.
Cow Goes Moo: Artificial Intelligence-Based System Associates Images With Sounds
EurekAlert (11/15/16) Jennifer Liu
Using artificial-intelligence techniques, a new system from Disney Research and ETH Zurich is capable of learning the association between images and the sounds they make. According to the researchers, a system that can successfully recognize and return the sound of a slamming door or car could be used to add sound effects to film or give audio feedback to people with visual impairments. To train the system, data was collected from videos with audio tracks. A key challenge was the presence of extraneous sounds that were not associated with the visual content, such as background music, narration, and off-screen noises. The team was able to filter out the extraneous sounds by looking for redundancies between videos; for example, a video collection of cars will contain recurring car engine sounds. The researchers say uncorrelated sounds generally are not repeated within other videos and can be filtered out. After the video frames containing uncorrelated sounds are removed, the algorithm learns which sounds match with an image. The researchers found the system learning from filtered videos returned better results than one trained with the original video collection.
Paralyzed ALS Patient Operates Speech Computer With Her Mind
UMC Utrecht (11/13/16)
Researchers from UMC Utrecht in the Netherlands have used a brain implant to help a patient with amyotrophic lateral sclerosis (ALS) operate a speech computer with her mind. The researchers placed electrodes in the patient's brain, enabling her to wirelessly control the computer. "This is a major breakthrough in achieving autonomous communication among severely paralyzed patients whose paralysis is caused by either ALS, a cerebral hemorrhage, or trauma," says UMC Utrecht professor Nick Ramsey. The patient operates the speech computer by thinking about moving her fingers; this changes the brain signal under the electrodes, which is then converted into a mouse click. The patient is presented with a screen showing the alphabet and a few additional functions, each of which light up one by one. The patient selects a letter by using her brain to influence a mouse click at the right moment; the system enables the patient to compose words, letter by letter, which are then spoken by the speech computer. "We hope that these results will stimulate research into more advanced implants, so that some day not only people with communication problems, but also people with paraplegia, for example, can be helped," Ramsey says.
Better Together: Improving the Management of 'Systems of Systems'
CORDIS News (11/11/16)
The European Union-funded Dynamic Management of Physically Coupled Systems of Systems (DYMASOS) project focuses on how independent technologies can be harnessed to work together to optimize outcomes. "The project has made an important contribution in taking the first concrete steps into realizing and concretizing...the Internet of Things," says DYMASOS researcher Iiro Harjunkoski. The project will enable everyday objects to be networked via the Internet, allowing them to send and receive data and giving any system the capacity to be "smart" and coordinate with other systems. The researchers note DYMASOS was based on real industrial case studies underpinned by a thorough analysis of markets, industrial needs, and challenges of the industrial project partners. The focus of the case studies was in the field of chemical production from among the largest chemicals producers in the world, and in the operation and engineering of electric power distribution and electric vehicle charging infrastructures. DYMASOS researchers developed four approaches to modeling systems of systems. "The project has developed a number of practical demonstrations, which will interest other complexes within and across companies and organizations to start to take further interest," says DYMASOS researcher Mark Lewis.
Breakthrough in the Quantum Transfer of Information Between Matter and Light
Polytechnique Montreal (11/10/2016) Florence Scanvic
Researchers from Canada's Polytechnique Montreal and France's National Center for Scientific Research created a quantum bit (qubit) in zinc selenide (ZnSe), making it possible to produce an interface between quantum physics and the transfer of information at the speed of light. The researchers say this breakthrough could pave the way to producing quantum communications networks. ZnSe is a crystal in which atoms are precisely organized, as well as a semiconductor into which it is possible to intentionally introduce tellurium impurities, on which holes are trapped similar to air bubbles in glass. This environment protects the hole's spin and coherence, and helps maintain its quantum information accurately for longer periods. The researchers use photons generated by a laser to initialize the hole and record quantum information on it. To read the information, the researchers excite the hole with a laser and then collect the emitted photons. The process results in a quantum transfer of information between the stationary qubit, encoded in the spin of the hole and held captive in the crystal, and the flying qubit, which travels at the speed of light. The researchers say their technique demonstrates the possibility of generating a qubit faster than any of the methods that have been used to date.
Abstract News © Copyright 2016 INFORMATION, INC.
To submit feedback about ACM TechNews, contact: [email protected]
|
