Association for Computing Machinery
Welcome to the May 16, 2014 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).


FCC Approves Plan to Consider Paid Priority on Internet
The Washington Post (05/15/14) Cecilia Kang

The U.S. Federal Communications Commission (FCC) on Thursday voted 3-2 in favor of considering a plan to let Internet service providers (ISPs) charge websites for higher-quality delivery of content to U.S. consumers. Although not final, the vote is a significant step forward on a controversial idea that is opposed by many consumer advocates, Silicon Valley firms, and Democratic lawmakers. The plan could potentially create a new revenue channel in which an ISP could charge a website for faster video streaming, although it would bar telecom firms from blocking websites entirely. The plan's critics say the new rules would mean the end of net neutrality, a principle that calls for ISPs to treat all Web content equally. FCC chairman Tom Wheeler says the agency will not allow for unfair or "commercially unreasonable" business practices, but critics say such arrangements would hurt consumers and smaller companies that cannot afford to pay ISPs for faster Internet connections. Wheeler says consumers would be guaranteed a baseline of service and the agency would not allow anti-competitive arrangements. "There is one Internet. It must be fast, it must be robust, and it must be open," Wheeler says. "The prospect of a gatekeeper choosing winners and losers on the Internet is unacceptable."
View Full Article - May Require Free Registration | Return to Headlines | Share Facebook  LinkedIn  Twitter 

Error at IBM Lab Finds New Family of Materials
The New York Times (05/16/14) John Markoff

Accidentally leaving out a chemical in the mixture of three components has led to the discovery of a new family of unusually strong and lightweight materials at an IBM lab. The two new types of synthetic polymers, or thermosets, yielded from the mixture also are self-healing and reformable, which can be critical in the manufacture of recyclable products. The thermosets' bone-like strength and rigidity are derived from their three-dimensional network of chemical bonds, and combining them with materials such as carbon nanotubes could further augment their strength. Computer simulation showed the material could be generated either as an extremely rigid polymer or a soft and gooey polymer. The materials are not yet ready for commercial application, but IBM Almaden Lab scientists say composite uses they are currently working on with several universities could have significant ramifications on manufacturing and fabrication in a wide spectrum of fields, including aerospace, microelectronics, and transportation. "Thermosets are designed to be exceptionally stable in terms of temperature and mechanical properties; they are not designed to be reversible," notes Virginia Institute of Technology professor Timothy E. Long. "To think about materials that have all of these properties, and which are also recyclable, is an advance."
View Full Article - May Require Free Registration | Return to Headlines | Share Facebook  LinkedIn  Twitter 

Where the Internet of Things Could Take Society by 2025
Center for Digital Education (05/14/14) Tanya Roscorla

The Pew Research Center Internet Project and Elon University's Imagining the Internet Center on Wednesday released a report predicting how the Internet of Things (IoT) will shape society through 2025. Among the more than 1,600 experts cited, 83 percent said the IoT would have "widespread and beneficial effects on the everyday lives of the public by 2025." The report says both the IoT and wearable computing will advance significantly over the next 11 years, but warns privacy concerns will rise due to a surge in the volume of data gathered by connected devices and resulting growth of profiling. Also by 2025, most people will not take advantage of progress in information interfaces by connecting their brains to the IoT, the report says. Experts believe the IoT will have complex, unintended consequences, and could increase the digital divide. In addition, the report predicts responses to the IoT will alter human relationships. Although the popularity of Google Glass over the next decade remains to be seen, the idea of a device that overlays information onto the physical world will remain, says the Massachusetts Institute of Technology's David Clark. People will use coordinated appliances in conjunction with the scannable physical world, says ACM president Vint Cerf, who cautions that these appliances will be vulnerable to hostile takeover.

New Implanted Devices May Reshape Medicine
UT Dallas News (05/13/14) LaKisha Ladson

Researchers at the University of Texas at Dallas (UT Dallas) and the University of Tokyo have developed electronic devices that become soft when implanted inside the body and can deploy to grip three-dimensional objects. The research is one of the first demonstrations of transistors that can change shape and maintain their electronic properties after they are implanted in the body, says UT Dallas graduate student Jonathan Reeder. The keys to the new technology are shape memory polymers, which were developed by UT Dallas professor Walter Voit. The polymers respond to the body's environment and become less rigid when they are implanted. Moreover, the devices are built with layers that include thin, flexible electronic foils. "In our device design, we are getting closer to the size and stiffness of precision biologic structures, but have a long way to go to match nature's amazing complexity, function, and organization," Voit says. During testing, the researchers found that after implementation, the device morphed with the living tissue while maintaining excellent electronic properties. "Our research comes from a different angle and demonstrates that we can engineer a device to change shape in a more biologically compatible way," Reeder says.

The Brain: Key to a Better Computer
Sandia National Laboratories (05/15/14) Sue Holmes

The goal of a long-term neuro-inspired computing research project at Sandia National Laboratories is the creation of algorithms that run on computer systems whose functionality is more akin to the brain than to conventional computers. Sandia researcher Murat Okandan says such systems would be perfect for running unmanned aerial vehicles, robots, and remote sensors, as well as solving big data challenges, through their ability to identify patterns and anomalies. Sandia's John Wagner sees neuro-inspired computer architecture as fundamentally distinct from current computers, combining processing and storage in a network arrangement "so the pieces that are processing the data are the same pieces that are storing the data, and the data will be processed with all nodes functioning concurrently." Okandan observes modern computers have no sense of what data is. "Whereas if you think about neuro-inspired computing systems, the structure itself will have an internal representation of the datastream that it's receiving and previous history that it's seen, so ideally it will be able to make predictions on what the future states of that datastream should be, and have a sense for what the information represents," he notes.

Who Did What?
MIT News (05/14/14) Larry Hardesty

Researchers at the Massachusetts Institute of Technology and the University of California, Irvine say they have developed a new activity-recognition algorithm that has several advantages over existing alternatives. For example, the algorithm's execution time scales linearly with the size of the video it is searching, meaning that if one file is 10 times the size of another, the algorithm will take 10 times as long to search, instead of 1,000 times as long, as is the case with current algorithms. In addition, the researchers say the algorithm can make good guesses about partially completed actions, which enables it to analyze streaming video. The amount of memory the algorithm requires also is fixed, regardless of how many frames of video it has already reviewed, which means it can handle video streams of any length. The researchers note the rules relating to subactions are the key to the algorithm's efficiency. As a video plays, the algorithm creates a set of hypotheses about which subactions are being depicted where, and it ranks them according to probability. Although the algorithm cannot limit itself to a single hypothesis because each new frame could require it to revise its probabilities, it can eliminate hypotheses that do not fit its grammatical rules.

AI System Reads Novels, Writes Music for Them
IDG News Service (05/15/14) Tim Hornyak

Researchers have trained an artificial intelligence system to read works of fiction and create music based on the texts. The TransProse project is a collaboration between Hannah Davis of New York University, who explored the idea for a thesis project, and Saif Mohammad, a researcher at the National Research Council Canada, who created a word-emotion lexicon. First, the system reads the text of a work, then based on word scans, it assigns densities of two different states, positive or negative, and eight different emotions--joy, sadness, anger, disgust, anticipation, surprise, trust, and fear. The system then composes a musical piece that chronologically follows the novel, broken up into beginning, early middle, late middle, and end parts. The tempo, key notes, octaves, and other musical variables are determined by the emotional density data. The system renders Joseph Conrad's "Heart of Darkness" as a brooding progression in C minor with fear and sadness as its themes. "One practical application might be in online book stores where a customer can click on a button to listen to the emotional tone of a book before deciding to buy it," Mohammad says.

Kindergarten Bots Teach Language to Tots
New Scientist (05/15/14) Hal Hodson

Massachusetts Institute of Technology (MIT) graduate student Jacqueline Kory is conducting a series of experiments at MIT's Media Lab on how well children can learn from robots, in collaboration with research team leader Cynthia Breazeal. They are testing children to see how well they retain new words learned from robots over time, using social storytelling robots called DragonBots that offer personalized teaching. The DragonBots feature a cute aesthetic design combined with eye movements and other reactive elements and behaviors that enhance the narratives they communicate to children in conjunction with images displayed on tablet computers. Among the insights gained from the tests was the fact that children seem more willing to trust robots that mirror their body language, and their tendency to learn is better as a consequence. The DragonBots currently only function in a controlled environment, but upgraded next-generation devices could be applied to homes and schools. The Sesame Workshop's Michael Levine says the DragonBots surmount the basic challenge of individualization, which is innate to the development of educational content for TV and tablets. Breazeal sees the research offering "a tremendous opportunity to develop new technology to support children's education, especially early childhood learning."

A Smart Watch Controlled by Twists, Tilts, and Clicks
Technology Review (05/14/14) Rachel Metz

Carnegie Mellon University (CMU) researchers say they have developed a better way to control smart watches by enabling users to physically tilt, click, and twist the watch's bezel. Their method could make it easier to set up appointments, play music, or navigate a map on a tiny display. The smart watch is equipped with a 1.5-inch display that sits in front of an ARM processor, and two Hall-effect displacement sensors that measure movements of the screen along two different axes. The watch is connected to a computer with software that processes the user's interactions and runs an app that is shown on the watch face. Although the prototype watch looks bulky, a commercial smart watch with this kind of functionality could be made with pressure sensors instead of mechanical ones, notes CMU professor Chris Harrison. He says when developing the device the researchers tried to follow certain rules, such as designing the watch so users would not have to lift their fingers from the screen while using it. Their work was detailed at the recent ACM CHI Conference on Human Factors in Computing Systems in Toronto.

Computers See Your Face as a Child. Will They Recognize You as an Adult?
The Atlantic (05/13/14) Alexis C. Madrigal

As concerns grow about the ability of computers to track children and their activities into their adult lives, questions emerge about whether facial recognition technology could recognize an adult based on images of that person as a child. University of Kent forensic researcher Stuart Gibson says facial recognition systems would have difficulty matching an image of a child under the age of seven with that same person as an adult, because of the way in which the face changes beginning at age seven. Gibson developed computer models that attempt to artificially age images of children at different ages. His attempts to quantitatively model changes in bone structure, skin texture, and other variables could improve facial recognition technology. University of California, Los Angeles professor Stefano Soatto believes studies of aging and facial recognition could benefit from his paper on eliminating variability with respect to focal length in images, which heavily distorts faces. Researchers could enter numerous images into the model and "learn away the variability," Soatto says. Variability includes actual changes in a person's face over time, as well as "nuisance variability," or image features that are irrelevant to identity. Experts believe research on facial recognition and aging will increase as social networks acquire thousands of images of an individual over a period of time.

Google Hopes Test Drives Steer Americans to Embrace Its Robot Cars
Reuters (05/14/14) Alexei Oreskovic

Google this week embarked on its most concerted effort to provide an up-close look at its self-driving cars. The company provided 30-minute ride-alongs to more than two dozen reporters, showcasing the car's ability to automatically and safely navigate around Mountain View, CA, streets packed with cyclists, pedestrians, and traffic signs. Google managers also discussed the self-driving car technology during morning press briefings at the Computer History Museum. Company founders Larry Page and Sergey Brin believe the driverless car is revolutionary technology because it has the potential to sharply reduce fatalities on the road. Still, consumer acceptance and regulation may be as important as perfecting the technology. The public needs to understand that a self-driving car is "not something that you need to fear but something you need to embrace," says Ron Medford, director of safety for Google's self-driving car project. "We do find that when people experience it, we get remarkable results and responses," Medford says. Brin says the technology could be available by 2017.

Advance Brings 'Hyperbolic Metamaterials' Closer to Reality
Purdue University News (05/12/14) Emil Venere

Purdue University researchers are developing practical applications for "hyperbolic metamaterials," which are ultra-thin crystalline films that could lead to optical advances in a wide range of technologies. The researchers have shown how to create "superlattice" crystals from layers of the metal titanium nitride and aluminum scandium nitride, a dielectric, or insulator. "This work is a very important step in terms of fundamental contributions in materials science and optics as well as paving the way to some interesting applications," says Purdue professor Alexandra Boltasseva. "We believe this demonstration brings a paradigm shift to the field of metamaterials similar to developments that led to dramatic advances in silicon technology." The researchers created the superlattices using a method called epitaxy, which involves growing the layers inside a vacuum chamber with a technique known as magnetron sputtering. The researchers have demonstrated these superlattices also can be developed where the layers could be as thin as 2 nanometers, a tiny dimension only about eight atoms thick. "The fascinating optical properties we see here are a manifestation of extraordinary structural control that we have achieved," says Purdue researcher Bivas Saha.

Meeting the Power Challenge: Natalie Bates on Creating More Energy-Efficient HPC
Scientific Computing (05/08/14)

In an interview, Energy Efficient High Performance Computing Working Group (EE HPC WG) chair Natalie Bates discusses the challenges of improving the energy efficiency of high-performance computers (HPC). Bates says the largest obstacle to more energy-efficient HPC is scaling silicon-based technologies. Although alternatives to silicon such as quantum computing, optical computing, and biomolecular computing could significantly improve energy efficiency, they remain in the early stages of research, she notes. A more realistic option for the near future is energy proportional computing, which allots resources such that energy consumed is optimized to generate useful work. When resources are not in use, their energy consumption should be minimal. Bates says a lack of standardization and metrics also presents a challenge for energy efficiency. However, she says the EE HPC WG has made some significant contributions to energy-efficiency measures and design, including guidelines for liquid cooling inlet temperatures, an improved power measurement methodology for use while running benchmarks and workloads, and an improved data center infrastructure energy efficiency metric. Bates says HPC needs customizable metrics based on a computing center's objectives. Hardware vendors, application developers, and system administrators should all help shape the next-generation of energy-efficient HPC, according to Bates.

Abstract News © Copyright 2014 INFORMATION, INC.
Powered by Information, Inc.

To submit feedback about ACM TechNews, contact: [email protected]
Current ACM Members: Unsubscribe/Change your email subscription by logging in at myACM.
Non-Members: Unsubscribe