Association for Computing Machinery
Welcome to the October 6, 2010 edition of ACM TechNews, providing timely information for IT professionals three times a week.

HEADLINES AT A GLANCE


One Goal: 10 Quadrillion Calculations
Wall Street Journal (10/05/10) Daisuke Wakabayashi

Japan is funding a supercomputer project aimed at producing the world's fastest computer. The K Computer, a joint venture between Fujitsu and Japan's Institute of Physical and Chemical Research, will be able to run 10 quadrillion calculations per second and is expected to be completed in 2012. The system is being designed to solve complex problems related to climate change and weather patterns, and also could be a powerful computational tool in the search for new developments in drugs, materials, and new technologies. "This is a tremendously important project for Japan's international competitiveness and for the advancement of science and technology in the country," says Fujitsu's Hideyuki Saso. The K Computer will combine 80,000 processors, each equipped with eight cores, for a total of 640,000 electronic brains. The United States, Russia, and China also are working to build faster supercomputers. For example, IBM is building a system, which aims to break the 10 petaflop barrier, at the National Center for Supercomputing Applications. "There is a growing recognition of the close link between supercomputing and scientific advancement as well as industrial competitiveness," says IDC's Steve Conway.
View Full Article - May Require Paid Subscription | Return to Headlines


Cell Phones Make a Touch Surface Smarter
Technology Review (10/06/10) Duncan Graham-Rowe

PhoneTouch is a prototype system developed by University of Lancaster researchers that lets users manipulate objects on a tabletop touchscreen by touching any part of their phone to the screen's surface. PhoneTouch makes it possible to transfer files between the phone and the surface and also can personalize interactions, says Lancaster professor Hans Gellersen. "Surfaces in general are good for working together in parallel," Gellersen says. "But when people work together they also want to bring information into the group." PhoneTouch uses a camera beneath the surface to recognize finger contact. The system also can determine the pattern made when the edge of a phone touches the surface. PhoneTouch can identify the accelerometers built into connected phones to see which of them experienced a slight bump at the moment of contact, which enables it to determine which user is in contact with the surface. "PhoneTouch not only establishes a connection but allows the phone to be used as a stylus on the surface, to select specific widgets," Gellersen says.


Aiming to Learn as We Do, a Machine Teaches Itself
New York Times (10/04/10) Steve Lohr

Carnegie Mellon University (CMU) researchers are developing the Never-Ending Language Learning (NELL) system, a computer that can master semantics by learning more like a human. NELL was provided with basic knowledge in various categories and connected to the Web with a mission to teach itself. "For all the advances in computer science, we still don't have a computer that can learn as humans do, cumulatively, over the long term," says CMU professor Tom M. Mitchell. NELL scans hundreds of millions of Web pages for text patterns that it uses to learn facts, with an estimated 87 percent accuracy. The learned facts are added to NELL's database. NELL is one of several projects aimed at enabling computers to better understand the meaning of language. IBM's Watson has attained advanced semantic understanding in history, literature, and sports. Google's Squared research project has demonstrated a grasp of semantic categories as it finds and presents information found on the Web. "What's exciting and significant about it is the continuous learning, as if NELL is exercising curiosity on its own, with little human help," says the University of Washington's Oren Etzioni. NELL's tools include programs that extract and classify text from the Web, programs that identify patterns, and programs that learn rules.


Researchers Argue for Smarter Traffic Lights
IDG News Service (10/01/10) Joab Jackson

Two European researchers have proposed using a combination of sensing technology, analytics, and networking to change the way traffic lights are timed. "Instead of waiting for a certain point in time before switching to green, we now wait for a critical number of vehicles [to be] ready for service at [the] maximum rate, which [is] given by the saturation flow," write Swiss Federal Institute of Technology Zurich's Dirk Helbing and Dresden University of Technology's Stefan Lammer. The researchers say the approach could reduce congestion by as much as 30 percent. They propose giving each traffic light a sensor so it will know when to change instead of being on a fixed timer. "In contrast to a fixed-time controller ... the green times are requested only when there is definite demand for them," the researchers write. "The cycle time is not fixed, and the service is not necessarily periodic." However, the researchers acknowledge that complete traffic light autonomy could lead to chaos, and they have proposed a dynamic control method that enables the various traffic lights to coordinate traffic flow with each other.


A Tracking Device That Fits on the Head of a Pin
American Friends of Tel Aviv University (10/05/10)

Tel Aviv University (TAU) researchers have developed nano-sized optical gyroscopes that can fit on standard-sized computer chips without compromising their sensitivity. The gyroscopes can pick up smaller rotation rates and create higher accuracy while maintaining smaller dimensions, says TAU professor Koby Scheuer. The key to the device are extremely small semiconductor lasers. As the device rotates, the properties of the light produced by the lasers change, which makes determining the rotation rates possible. The lasers are a few tens-of-micrometers in diameter, while conventional gyroscopes measure about six to eight inches, Scheuer says. "Conventional gyroscopes look like a box, and weigh two or three pounds," he says. "This is fine for an airplane, but if you're trying to fit a gyroscope onto a smaller piece of technology, such as a cell phone, the accuracy will be severely limited." Nano-gyroscopes integrated into cell phones could provide a better tracking function than existing global positioning systems, and the technology also could be incorporated into medical equipment, Scheuer says.


An Intelligent System for Maritime Surveillance Has Been Created
Carlos III University of Madrid (Spain) (10/05/10)

Universidad Carlos III de Madrid (UC3M) researchers have designed a system for maritime surveillance that can integrate information from different types of sensors using artificial intelligence and data-fusion techniques. The various sensors provide complementary data, which can be combined to obtain better information about what is happening in maritime areas of interest. The objective is to guarantee security in an area by monitoring the different ships that are in a given maritime route. "For that, it is necessary to have a complete, accurate, and up-to-date picture, similar to that which is provided to air traffic controllers, of all the ships that are in the area of coverage to be able to adequately manage maritime traffic and to detect anomalies as much in advance as possible," says UC3M researcher Jesus Garcia. The system can monitor 2,000 identifiable objectives between large and small vessels, and can process the data of up to 10 sensors. "We are able to make it so these vessels never lose their position, thus avoiding collisions or any type of problem in information management regarding the movement dynamics of these ships," says UC3M researcher Jose Luis Guerrero.


Intel Labs Creating Robots of the Future
eWeek (10/01/10) Jeffrey Burt

Researchers from Carnegie Mellon University and scientists from the Intel Labs site on campus recently held an open house to show off their projects, including a collaborative robot and a personal robot. The collaborate robot, CoBot2, led people around the room to different demos of projects whenever someone touched particular areas on its display. CoBot2 would stop and wait if someone was in its way, and, after three seconds, say, "Please excuse me." The collaborative robot stands less than five feet tall, and has a Segway base, a camera, and navigational device on top for orienting itself in its environment, and a tablet device for its interface. The personal robot, HERB (Home Exploring Robot Butler), could be used to assist people with disabilities in their homes. Offering similar maneuverability around a room, HERB is able to perform tasks, learn through imitation, and understand when to hand off work to people. HERB has two arms with fingered hands for grasping objects, a Segway base, and a spinning laser on top to give it a three-dimensional view of its environment.


Therapeutic Computer Games Target Stroke Sufferers
The Engineer (United Kingdom) (09/30/10) Stephen Harris

A therapeutic computer game developed in Ireland uses three-dimensional (3D) goggles to improve the brain's sense of perspective and balance in stroke sufferers. Stroke victims use a motion-controlled game console and 3D goggles to aim at targets. The activity is designed to help adjust their bodily movements to match their new spatial awareness. The Institute of Technology Carlow helped develop the software. "It is based on an existing therapy but the equipment needed to do it at the moment costs a lot of money," says Carlow's Joseph Kehoe. Researchers at Carlow also helped design the IQ-EQ Brain Trainer program, which helps people overcome phobias, addictions, or anxiety disorders, and also increase their IQ. The software assesses the severity of the problem by displaying disorder-related images and measuring the reaction time, and trains users on how to give less attention to those cues. Carlow researchers also are working on a similar app for the iPhone. The Anxt app measures the disorder, teaches users about their anxiety or phobia, provides tasks for controlling and conquering the problem, and gives scores to help people monitor their progress.


INFER Project Seeks Smart Software Solution
Bournemouth University (United Kingdom) (09/30/10)

Bournemouth University is leading the computational INtelligence platform For Evolving and Robust predictive systems (INFER) project, a European Union-funded initiative involving researchers and organizations in three countries launched to develop automated systems that help companies react and adjust to changes in market, behavior, or operational conditions. INFER will focus on pervasively adaptive software systems in developing an open modular platform for commercial settings and industries. The best predictive model will emerge from a novel type of environment either anonymously or by user high-level goal-related assistance and feedback. The project represents a paradigm shift from human labor and knowledge-intensive processes of building predictive systems to anonymous, evolving complex systems, says project coordinator Bogdan Gabrys. "The INFER project takes a novel approach to applying intelligent methods: A combination of modern software, systems, and knowledge engineering, together with application experience of the process industry," Gabrys says. "This will enable development of an effective platform for predictive systems that is easily applicable not only in process, but in many other industries."


Physicists Break Color Barrier for Sending, Receiving Photons
University of Oregon (09/30/10) Jim Barlow

University of Oregon researchers have developed a method to change the color of single photons in a fiber-optic cable, a development that could lead to fast, secure transfers of large datasets. The single-photon project involves a dual-color burst of laser light changing the color of a separate single photon of light, which could be important to the future of Internet communications technology, says Oregon researcher Michael G. Raymer. "At the level of single photons, we would like to send data in different channels--colors or wavelengths--at the same time," Raymer says. The researchers used two lasers to create a burst of dual-laser light, which can cause a single photon to change to a new color. The process, called quantum frequency translation, enables devices that talk to one another using a given color of light to communicate with devices that use a different color. "Another big advantage to our technique is that it allows us to change the frequency of a single photon by any chosen amount," Raymer says.


MIT Researchers Tout Network Intrusion Recovery System
Network World (09/29/10) Michael Cooney

Massachusetts Institute of Technology researchers are developing RETRO, a system designed to make it easier for organizations to recover from security breaches. RETRO lets administrators specify offending actions that they want to undo and makes repairs by selectively undoing the offending actions. "Since many adversaries go to great lengths to prevent the compromise from being discovered, it can take days or weeks for a user to discover that their machine has been broken into, resulting in a loss of all user work from that period of time," the researchers say. RETRO uses the action history graph to undo an unwanted action and its indirect effects by rolling back the direct effects. "An important assumption of RETRO is that the attacker does not compromise the kernel," the researchers note. However, security vulnerabilities are occasionally found in the kernel. To get around that problem, the researchers say one solution could be to use virtual machine-based techniques, which they plan to explore in the future.


Truthy.indiana.edu to Search, Identify Smear Tactics, Twitter-Bombs Through November Election Runup
Indiana University (09/28/10) Steve Chaplin

Indiana University researchers have developed Truthy.indiana.edu, a Twitter-based research tool that combines data mining, social networking analysis, and crowdsourcing to uncover political misinformation. The site sifts through thousands of tweets per hour in search of political keywords and then isolates patterns of interest, known as memes. "When we drill down we'll be able to see statistics and visualizations relating to tweets that mention the meme and basically reconstruct its history," says Indiana professor Filippo Menczer. The data is sent to Truthy.indiana.edu, where visitors can see the output of a sentiment analysis algorithm that examines mood-identifying words and assess them on a known psychometric scale. The algorithm identifies a meme on scales ranging from anxious to calm, hostile to kind, unsure to sure, and confused to aware. The site relies on input from users to denote a meme as "truthy," or misinformation represented as fact. "One of the concerns about social media is that people are being manipulated without realizing it because a meme can be given instant global popularity by a high search engine ranking, in turn perpetuating the falsehood," Menczer says.


Service Using Public Data to Relate Spanish Geographical Information to Statistics
Universidad Politecnica de Madrid (Spain) (09/27/10) Eduardo Martinez

Universidad Politecnica de Madrid researchers have developed a system that uses Spanish public data to relate geographical information to statistics. The researchers say the system, part of the GeoLinked Data Initiative, enables users to locate and relate spatial data with an unprecedented level of detail. For example, users can analyze the relationships between Spain's coastal regions and statistics about unemployment, population, housing, and industry. The data are displayed as time series by years, and depict the trend of a selected parameter over a 10-year period. All the information from the service is integrated into the Linked Open Data initiative, a worldwide data network that contains more than 4.5 million references organized by subject and connected to each other via the Internet.


Researchers Honored for Influential Contributions to Software Engineering Field
Kansas State University News (09/29/10) Jennifer Torline

"Bandera: Extracting Finite-State Models From Java Source Code," published in 2000, received the Most Influential Paper Award at the recent International Conference on Software Engineering. The award is given to the paper that has the biggest influence on the theory and practice of software engineering in the 10 years since it was published. Kansas State University professors John Hatcliff and Robby were part of the seven-member team that in 1998 began to research how different technologies could test for problems that arise when computer programs multitask. "Our main contribution was to illustrate how model checking, a technique based on discrete mathematics and formal logic, could be applied to software systems to reason about systems that were multitasking in nature," Hatcliff says. "What was interesting about this project was that it brought together a number of different technologies, and that was really one of the things that made this paper stand out," Robby says.


Abstract News © Copyright 2010 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