Association for Computing Machinery
Welcome to the July 8, 2011 edition of ACM TechNews, providing timely information for IT professionals three times a week.

Also, please download our new ACM TechNews iPhone App from the iTunes Store by clicking here and our new ACM TechNews iPad App by clicking here.


Rutgers-Led Experts Assemble Globe-Spanning Supercomputer Cloud
Rutgers Today (07/06/11) Carl Blesch

Researchers at Rutgers University, the University of Texas at Austin, and IBM have created a massive virtual supercomputer cloud designed to solve complex computing tasks. By combining different computing resources, integrating them into a virtual cloud, and disbanding resources that are no longer needed, the researchers demonstrated how to build a federated cloud, says Rutgers professor Manish Parashar. "Our goal is to make these federated, high-performance computing clouds more useful to industry," Parashar says. IBM's Thomas Watson Research Center provided access to supercomputers while Rutgers and University of Texas researchers contributed expertise in autonomic computing. "Modeling the systems requires simulating complex physics and exploring large numbers of parameters," says IBM's Kirk Jordan. The researchers used IBM's Blue Gene supercomputers based in New York as well as IBM supercomputers at the King Abdullah University of Science and Technology in Saudi Arabia. They ran a demonstration of their system through an Apple iPad tablet computer. "Accessing such a large supercomputer configuration from an iPad to easily run a complex application makes supercomputing much more accessible to scientists and engineers," says Rutgers graduate student Moustafa AbdelBaky.

Israeli Algorithm Sheds Light on the Bible
Associated Press (07/05/11) Matti Friedman

Israeli researchers have developed software that analyzes style and word choices to differentiate parts of a single text written by distinct authors, and it supports earlier research that the Bible may have been crafted by multiple authors. The research team included Tel Aviv University computer scientist Nachum Dershowitz; his son, Idan Dershowitz, a Bible scholar at Hebrew University; and computer science doctoral student Navot Akiva. When the researchers ran the algorithm on the Torah, it uncovered a division between priestly and nonpriestly authors, and it corresponded with the traditional academic division at a rate of 90 percent. The software recognizes repeated word selections, such as uses of the Hebrew counterparts to "if," "and," and "but," and also takes note of synonyms. The program then partitions the text into strands it thinks were authored by different writers. The researchers had to objectively demonstrate that the program could correctly distinguish between different authors before it could be applied to the Bible. To test the program, they jumbled the Hebrew Bible's books of Ezekiel and Jeremiah into a single text and ran the software, which sorted the randomly blended text into its component parts "almost perfectly," the researchers say.

Eric Lander and Larry Summers Talk Innovation, R&D
CCC Blog (07/05/11) Erwin Gianchandani

Government's role in nurturing innovation and the economic impact of science and technology research and development were topics discussed at the Brookings Institution by Larry Summers, former U.S. National Economic Council director, and Eric Lander, co-chair of the President's Council of Advisers on Science and Technology. "The United States has been the master of basic research funding and ... this kind of innovation policy where the public sector does recognize that it makes those investments in the transitional stage," Lander said. He also emphasized the value of science, technology, engineering, and math education in the United States, noting that many U.S.-based jobs are going unfilled because there are not enough people with relevant skills. Summers cited three competing elements contributing to the success of U.S. information technology innovation--the country's lead in relevant types of basic research, the Defense Advanced Research Projects Agency and its need for researchers to communicate, and a culture of venture capital and entrepreneurism. Summers said the first two elements are at least equal in importance in terms of innovation policy. "Just as I think about what the role of people like us in the system is, I think carrying the torch hard for number one and number three--which don't have alternative torch-bearers--needs to be a large part of where we are," he stressed.

Biomolecular Computer Can Autonomously Sense Multiple Signs of Disease (07/06/11) Lisa Zyga

Weizmann Institute of Science researchers have developed a biomolecular computer that can autonomously sense several different types of molecules at once, and could be integrated with a biomedical database of diseases to all systems to determine which drugs to give to a patient. "These devices would have the ability to sense disease indicators, diagnose the disease, and treat it by administering or activating a therapeutic biomolecule," says Weizmann researcher Binyamin Gil. The system is based on the researchers' previous demonstration of a biomolecular computer that consists of a binary system made of biological components. The new system can detect disease indicators from miRNAs, proteins, and small molecules such as ATP. The researchers say that the computer's detection method is simpler than before, requiring fewer components and less interactions with the disease indicators. "The biggest challenge is operating such devices in living surroundings like the blood stream or cell's cytoplasm," Gil says.

Robo-Paparazzi Learn How to Take the Perfect Photo
New Scientist (07/06/11) Jacob Aron

NAO, a humanoid robot developed by researchers at the International Institute of Information Technology (IIIT), can take aesthetically pleasing photographs. NAO, which has a head-mounted camera, follows two simple photographic guidelines known as the rule of thirds and the golden ratio, and is capable of assessing the quality of its photos by rating focus, lighting, and color. The researchers, led by IIIT computer scientist Raghudeep Gadde, taught NAO to take high-quality pictures by analyzing the top and bottom 10 percent of 60,000 images from a Web site hosting a photography contest, as rated by humans. When an image scores below a certain quality threshold, NAO automatically takes another picture, improving on the first attempt by working out the shot's deviation from the photographic guidelines and making the appropriate correction to its camera's orientation. "Earlier photographer robot systems are predominantly limited to capturing photographs of humans" because they rely on face or skin-color detection, Gadde says. "Our approach is generic and does not rely on the subject of the image being captured."

Magnetic Memory and Logic Could Achieve Ultimate Energy Efficiency
UC Berkeley News Center (07/01/11) Robert Sanders

University of California, Berkeley researchers are developing magnetic microprocessors that consume the least amount of energy allowed by the laws of physics. Conventional silicon-based microprocessors rely on electric currents, while magnetic microprocessors do not require moving electrons, dissipating just 18 millielectron volts at room temperature, which is the minimum allowed by the Landauer limit. The magnetic microprocessors enable researchers to "store and process information using magnets, and if you make these magnets really small, you can basically pack them very close together so that they interact with one another," says Berkeley's Brian Lambson. Berkeley professor Jeffrey Bokor is working with Lambson and graduate student David Carlton at the Center for Energy Efficient Electronics Science to build computers that operate at the Landauer limit. Lambson showed through calculations and computer simulations that a simple memory operation can be conducted with an energy dissipation very close to the Landauer limit. "We are working now with collaborators to figure out a way to put that energy in without using a magnetic field, which is very hard to do efficiently," Bokor says.

Safer Skies
MIT News (07/05/11) Larry Hardesty

The U.S. Federal Aviation Administration (FAA) recently mandated that all commercial aircraft must be equipped with a new tracking system that broadcasts global positioning system data to ground-based radar by 2020. The FAA also called on Massachusetts Institute of Technology (MIT) researchers to lead an investigation of the system's abilities. The researchers developed an algorithm that uses data from the tracking system to predict and prevent collisions between small aircraft. One of the major challenges in designing the algorithm was limiting false positives, says MIT's Maxime Gariel. The researchers developed a two-tiered system of alerts, one of which warns pilots that their trajectories are converging, while a higher alert indicates a severe risk of collision. Each alert corresponds to a volume of space around each plane, with the higher alert being a smaller and fixed size volume, while the other alert has a larger, fluctuating space. The researchers tested the system at MIT's Lincoln Laboratory, generating random trajectories for hypothetical aircraft that match real-world statistics. "From the limited data I've seen, it seems that the algorithms that they're looking at are performing better than the algorithms that are in existing systems that can be bought today," says the FAA's David Gray.

Engineers Work to Ease Internet Data Flow as Demand for Video Grows
Princeton University (06/30/11) Chris Emery

Princeton University researchers have developed the Edge Laboratory, a miniature version of the global communications network, to develop new ideas and systems that will help ensure that future networking infrastructure will meet consumer demand. The Edge Laboratory focuses on edge networks such as 4G, Wi-Fi, U-Verse, and FiOS, which connect computers and other devices to the core communications networks. The researchers, led by Princeton professor Mung Chiang, are working with Internet service providers and other telecommunications companies to test academic theories and help companies develop new systems for pricing and data delivery. For example, Chiang is experimenting with new ways to store and deliver content that could reduce Internet bottlenecks. The Princeton researchers also developed the Time-dependent Usage-based Broadband price Engineering system, which gives consumers more information and control over when they use the Internet and how much they pay. The system enables customers to track their usage through a mobile application and avoid downloading data-intensive files during peak usage times.

Team Aims Intelligence at Voice-Recognition Systems
The Engineer (United Kingdom) (07/04/11) Andrew Czyzewski

Researchers from the universities of Cambridge, Sheffield, and Edinburgh are participating in the Natural Speech Technology project, which seeks to make a flexible and adaptable voice-recognition system that is capable of handling normal conversation. Existing systems need a significant amount of data, are designed for a single purpose, and perform poorly with natural speech. The researchers are developing base models and algorithms for synthesis and recognition that will be able to almost instantaneously learn from and adapt to new scenarios and contexts. The researchers also are working on conversion-recognition software for detecting who spoke what, when, and how in any acoustic environment, and are honing speech synthesizers capable of generating the expressive range of natural speech. The team also will consider suitable applications for the technology, such as voice-controlled devices for the home, expressive-assisted speech for people with conditions such as motor neuron disease, and fully automated meeting transcription. "This research could open the door to computer-speech technology becoming commonplace throughout our lives--at home, at work, and in our leisure time," says Edinburgh professor and project lead Steve Renals.

U. Va. Engineers Help Manage Crisis With Crowdsourcing
UVA Today (06/30/11) Andrew Clark

A Web-based crowdsourcing application developed by a team at the University of Virginia could be used to provide the public with critical information in the event of a catastrophe. The tool enables people to quickly disseminate detailed and chronological information about an event, and allows other users to view it. The researchers were inspired by School of Medicine faculty, who were interested in a tool that would provide assistance in managing a potential disaster within the community, such as a bioterrorism attack. Users would be able to submit reports that contain headlines, the nature of the incident, location and time, and photos, among other details, through a Web site or by email. The reports would be embedded on a Google map to identify the affected area, highlighted by a marker and a short description. A key feature is the ability to send mass alerts via email and text messages to the tool's users. "When disaster strikes, the media can announce the Web address of the crowdsourcing software so citizens can participate in the crisis recovery by submitting reports of what they observe," says professor Alfred Weaver. The team is still working on ways to help ensure the reliability of submitted information.

The Next Generation of Computing
University of Cambridge (07/04/11)

Scientists at the University of Cambridge are working to generate spin current more efficiently. Working with a group from Germany's University of Muenster, the team has used the collective motion of electrons' spins, called spin waves, to enhance spin currents. A volume of spin current large enough is needed to support current and future electronic devices. The finding partly addresses this major obstacle to developing viable spin current technology. "You can find lots of different waves in nature, and one of the fascinating things is that waves often interact with each other," says Hidekazu Kurebayashi with the Microelectronics Group at the Cambridge's Cavendish Laboratory. He notes that there are a number of different interactions in spin waves. "Our idea was to use such spin wave interactions for generating efficient spin currents," Kurebayashi says. The team reports that one of the spin wave interactions, called three-magnon splitting, generates spin current 10 times more efficiently than using pre-interaction spin waves.

Wi-Fi 'Napping' Doubles Phone Battery Life
Duke Today (06/30/11)

Duke University researchers have developed SleepWell, software that can double the battery life of mobile devices by making changes to Wi-Fi technology. SleepWell enables mobile devices to go into sleep mode to save power while neighboring devices download information, which saves energy for the sleeping device as well as competing devices. "The SleepWell-enabled Wi-Fi access points can stagger their activity cycles to minimally overlap with others, ultimately resulting in promising energy gains with negligible loss of performance," says Duke graduate student Justin Manweiler, who developed the software with Duke professor Romit Roy Choudhury. "The SleepWell system can certainly be an important upgrade to Wi-Fi technology, especially in the light of increasing Wi-Fi density," Choudhury says. Manweiler says "the testing we conducted across a number of device types and situations gives us confidence that SleepWell is a viable approach for the near future."

Making Reading Better: LiquidText Software Supports Active Reading Through Fingertip Manipulation of Text in Documents
Georgia Tech Research News (06/28/11) Abby Robinson

Georgia Tech researchers have developed LiquidText, software designed to facilitate active reading. "LiquidText offers readers a fluid-like representation of text so that users can restructure, revisualize, and rearrange content to suit their needs," says Georgia Tech graduate student Craig Tashman, who developed LiquidText with Georgia Tech professor Keith Edwards. LiquidText uses common fingertip gestures on a touch screen to provide flexible control of the visual arrangement of a sentence, including annotations. LiquidText also combines a traditional document reading space with a dedicated workspace area where the user can organize comments about the text. Comments can be pulled off, rearranged, grouped with other items, and copied and extracted. "LiquidText's more flexible notion of comments and large workspace area provide space for organizing and manipulating any comments or document excerpts the user may have created," Tashman says. LiquidText also enables readers to collapse, dog-ear, and create magnified views of text. "LiquidText's functionality lets a user view two or more document areas with just one action, parallelizing an otherwise serial task," Edwards says.

Abstract News © Copyright 2011 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