Google and IBM have announced a major research initiative to help universities provide the technical training needed for powerful and highly complex computing, which even the most elite universities currently lack. The two companies will build large data centers that students will be allowed to access on the Internet for remote programming and research, known as "cloud computing." Online services offered by Google, Yahoo, Microsoft, Amazon, and eBay are simple, consumer-based examples of cloud computing, but cloud computing is also being used to handle increasingly large computing challenges, which often involve searching the Internet and data sources for patterns and insights. Corporations have been responsible for most of the innovation in cloud computing, but computer scientists say a lack of skills and talent in graduates could limit further growth. "We in academia and the government labs have not kept up with the times," says Carnegie Mellon University's dean of computer science Randal E. Bryant. "Universities really need to get on board." Six universities--Carnegie Mellon, MIT, Stanford University, UC Berkeley, the University of Maryland, and the University of Washington--will participate in the project. Google and IBM will each set up a data center, which will run an open source version of Google's data center software and will be large enough to run ambitious Internet research. "This is a huge contribution because it allows for a type of education and research that we can't do today," says University of Washington computer science professor Edward Lazowska.
Click Here to View Full Article
to the top

Data Storage Discovery Earns Nobel
Washington Post (10/10/07) P. A3; Vedantam, Shankar

The Royal Swedish Academy of Sciences awarded the 2007 Nobel Prize in physics to Peter Gruenberg of Germany and Albert Fert of France for their discovery of giant magnetoresistance (GMR), the phenomenon that allows ultra-thin slices of metal to have different electrical properties in a magnetic field, which led to the ultra-small hard drives that have made iPods and high-capacity laptops possible. While Gruenberg and Fert worked independently, they will share the award of about $1.5 million. GMR describes a phenomenon at the junction of electricity and magnetism. When two layers of metal, such as iron, are separated by a thin layer of a second metal, such as chromium, the electrical resistance of the structure can be manipulated by a magnetic field. In GMR devices, a mechanical reader "head" moves over the data, altering the resistance within the head and controlling the flow of electricity, which is translated into the ones and zeros used for digital information. Without the discovery of GMR it would still be possible to store massive amounts of information in a tiny space, but it would be impossible to read it. Gruenberg and others credit IBM researcher Stuart Parkin with discovering the practical application of GMR, and some had expected Parkin to share the prize. IBM's vice president of research Mark Dean says Parkin was key to identifying the structures and materials used to create hundreds of devices that allow the computer industry to make dramatic leaps forward every year. "The raw understanding of how nature works is a great thing," Dean says. "The application of that knowing how nature works in the creation of something my mother can use is another great breakthrough--and as significant."
Click Here to View Full Article
to the top

New HPC Ph.D. Fellowship Program Announced
HPC Wire (10/08/07)

Full-time faculty members at Ph.D. granting institutions have until Oct. 30 to nominate Ph.D. students for the new High Performance Computing Ph.D. Fellowship Program. ACM, the IEEE Computer Society, and SC Conference Series announced the fellowship in response to recent reports that have stressed a need for highly trained HPC scientists and engineers. "The ACM/IEEE-CS HPC Ph.D. Fellowship Program is designed to directly address this recommendation by honoring exceptional Ph.D. students throughout the world with the focus areas of high performance computing, networking, storage and analysis," says Bill Kramer, general manager of the National Energy Research Scientific Computing Center (NERSC) at Lawrence Berkeley National Laboratory. "The ACM/IEEE-CS HPC Ph.D. Fellowship Program also supports our long-standing commitment to workforce diversity and encourages nominations of women, members of under-represented groups, and all who contribute to diversity." Kramer is a member of the standing committee that will lead the fellowship program. Selections will be based on Ph.D. students' research potential, how well their technical interests match those of the HPC community, their academic progress, and a demonstration of how they are likely to use HPC resources. Complete information on the fellowship program is available at http://www.sigarch.org/HPC_Fellowships.html.
Click Here to View Full Article
to the top

Are America's Software Skills Getting Soft?
Wisconsin Technology Network (10/06/07) DiRomualdo, Tony

Globalization represents a formidable challenge to the software industry, which leads to questions about whether American leaders are responding effectively to the challenge so that the U.S. economy can stay on top. An ACM study projects the continued growth of software offshoring as IT, work and business processes, national policies, and education undergo a shift, which will necessitate a grounding in computing basics, comprehension of the global community, and up-to-date business process and platform knowledge by the workforce. Villanova University professor Steve Andriole, a member of the ACM task force that collected and analyzed data for the study, says American dominance of IT research and development is "challenged" and the United States is in danger of losing its lead position to other nations if current trends continue. Even more distressing is the low score Andriole gives U.S. policy makers in terms of their emphasis and support for technology-focused education and training, and he notes that the U.S. academic community has made little effort to address a major fall-off in undergraduate computer science and management-information systems majors. In a policy brief, the Brookings Institution's Lael Brainard and Robert E. Litan suggest strategies for addressing U.S. leaders' failure to effectively deal with software globalization, including a careful examination of America's tax policies; an increase in funding for science and engineering education and training; greater enforcement of trade agreements; enforcement of regulations that shield against risks; more accurate government collection of official offshoring data; and above all else, the extension of wage insurance, adjustment aid, and training to cover permanently displaced workers.
Click Here to View Full Article
to the top

New Image Search Rules Out Guesswork
New Scientist (10/05/07) Barras, Colin

Southampton University computer scientist Jonathon Hare has developed an image search engine capable of returning results based on the content of the image rather than the words that are near the image on the Web page, which is how most image search engines currently work. The search engine uses a small set of images that have been manually tagged with keywords. Off-the-shelf software then creates a virtual mathematical space of all the features in the image. The search engine then finds images that contain similar geometrical shapes. "The idea is you construct the space from a training data set and then apply it to new images," says Hare. The search engine is able to find images based solely on their visual features, creating more accurate and extensive search results. For example, a search for "water" using Hare's search engine would return more pictures of oceans, rivers, and lakes than a typical search engine. Despite being able to return unique results without relying on keywords, Hare's search engine does have several downsides, as it would be difficult to update the engine's stock set of images and it may struggle with large amounts of varied images, such as what a Web search engine needs to sort through. Hare is currently working on improving the search engine's ranking system.
Click Here to View Full Article
to the top

100 Gb/s Internet2 Completed
TG Daily (10/09/07) Gruener, Wolfgang

The Internet2 consortium announced that its updated infrastructure is ready to go online and will launch with an initial capacity of 100 Gbps in dedicated chunks of 10 Gbps next January. The group demonstrated its new infrastructure during its Fall 2007 meeting, in which a third of a terabyte was transferred within five minutes over a 10 Gbps connection involving the University of Nebraska-Lincoln and Fermilab in Batavia, Ill. The optical improvements help to provide a "uniquely scalable platform on which to build side-by-side networks that serve different purposes, such as network research and telemedicine," according to the consortium. Internet2 President Doug Van Houweling adds, "More importantly, we believe the Internet2 network and its new capabilities will play an integral part in enabling our members to provide the robust cyber-infrastructure our community requires to support innovative research and education." The network will continue to offer an advanced Internet Protocol network that supports IPv6, multicast, and other high-performance networking technologies. The consortium also has plans to develop new 40 Gbps and 100 Gbps technologies.
Click Here to View Full Article
to the top

Intel Completes Photonics Trifecta
Technology Review (10/10/07) Greene, Kate

Intel researchers recently announced that they have developed a silicon-based light detector that can be combined with a silicon-based laser and a silicon modulator to form a completely silicon-based photonic chip. The silicon-based light detector is less expensive than traditional detectors as well as more effective, and is able to detect flashes at a rate of 40 gigabits per second. The silicon photonic chip could also be manufactured using techniques commonly used in the microchip industry, further reducing costs. In most light detectors, light is detected when a thin layer of gallium arsenide or indium gallium arsenide has a hole punched through it by certain forms of energy. Silicon, however, does not react the same way so a layer of germanium was added on top of the silicon to detect light. Germanium is also used in some current silicon devices, so adding germanium to the manufacturing process would not be exceedingly difficult. Intel says the next challenge is developing a process for integrating the detector and the other silicon devices onto a single chip. The integration is not expected to create any significant problems, but problems could arise as the devices have only been tested in labs so far. Director of Intel's silicon-photonics lab Mario Paniccia estimates that silicon photonic devices could be available to consumers in about five years.
Click Here to View Full Article
to the top

Tracing Computer History From "Ancient" Times to the Latest Technology
University of Wisconsin-Milwaukee (10/09/07) Quirk, Kathy

University of Wisconsin, Milwaukee assistant professor of information studies Thomas Haigh has been studying how computers have changed business and society. He has interviewed hardware and software pioneers, examined the development of databases, and written about the history of word processing. Haigh has found that although companies are constantly trying to capitalize on the latest technology, that it is not always the best approach. "There's this feeling that anything more than five years old is irrelevant, but one of the things I've found is that people are facing the same types of problems now as they did in the mid 1950s--projects using new technology are usually late and filled with bugs, the return on investment is hard to measure, and computer specialists are expensive and speak an alien language," Haigh says. Haigh is currently working on a social history of the personal computer. "Despite the shelves of books on the history of the personal computer there has been no serious historical study of how people used their computers or why they bought them," Haigh says. Haigh has also examined how early explanations and promises of computers were exaggerations and did not accurately reflect the reality, which led to disappointment among the public. "It's a platitude, but if we don't understand who we are and where we're coming from, how can we understand where we're going," Haigh says. "That's true of religion, culture, Iraq, and it's equally true of science and technology."
Click Here to View Full Article
to the top

Adobe Shows Off 3D Camera Tech
CNet (10/08/07) Shankland, Stephen

Adobe Systems is demonstrating new technology that would allow users to reach into the scene of a photograph and adjust the focus. Such a three-dimensional brush would depend on knowing the 3D nature of every pixel, according to Adobe's Dave Story. The 3D camera technology makes use of a lens that is able to transmit smaller images to a camera. Adobe's technology could also be used to remove items in the background of a photograph, without the artful selection that Photoshop and other robust software require. The idea is to provide computers with an understanding of the depth of images by giving photographs multiple sub-views, which would enable computers to reconstruct a model of the scene in 3D. Computers would be able to make substantial transformations of an image, including an artificial shift in focus from the original photograph, based on the information taken from slightly different vantage points at the same time. "With the combination of that lens and your digital darkroom, you have what we call computational photography," Story says. "Computational photography is the future of photography."
Click Here to View Full Article
to the top

NSF Announces $26 Million Solicitation for Projects That Advance Innovative Computational Thinking
National Science Foundation (10/01/07)

The National Science Foundation plans to spend at least $26 million for Cyber-Enabled Discovery and Innovation (CDI) research in fiscal year 2008, and up to $50 million on CDI projects in each of the next five years. The NSF is soliciting science and research projects that will lead to even more innovative computational thinking about concepts, methods, models, algorithms, and tools. Areas of focus include transforming data into knowledge to improve human understanding and produce new knowledge from heterogeneous digital data. The NSF is also looking for multidisciplinary research into social systems that have multiple interacting elements, and virtual organizations that cater to people and resources across institutional, geographical, and cultural boundaries. CDI proposals should describe how the impact of computational thinking on their project will lead to a paradigm shift in advances in several fields of science or engineering, and make a compelling case for how innovation in computational thinking will produce the anticipated results. They should also make use of intellectual partnerships that will benefit from synergies of knowledge and expertise in several fields and from different types of organizations, including foreign and not-for profit entities.
Click Here to View Full Article
to the top

Dragonfly or Insect Spy? Scientists at Work on Robobugs
Washington Post (10/09/07) P. A3; Weiss, Rick

Researchers are developing insect-sized robots and insects augmented with robotic systems, or robobugs, that could be used for rescue missions, spying, or guiding missiles in combat. Although no government agency has admitted to successfully creating such a system, several, including the CIA and the Defense Advanced Research Projects Agency, have admitted to working on projects that, if successful, would result in such spy devices. In fact, the nation's use of flying robots has increased more than fourfold since 2003, with over 160,000 hours of robotic flight logged in 2006. However, creating insect-sized robots is a significant challenge. "You can't make a conventional robot of metal and ball bearings and just shrink the design down," says University of California Berkeley roboticist Ronal Fearing. The rules of aerodynamics change at such small scales and any mechanical wings would need to flap in extremely precise ways, a huge engineering challenge. Researchers at the California Institute of Technology have developed a "microbat ornithopter" that is capable of flight but smaller than the palm of a person's hand, and Vanderbilt University has developed a similar device. At the International Symposium on Flying Insects and Robots, Japanese researchers revealed radio-controlled flyers with four-inch wingspans that look like hawk moths. A known DARPA project, called the Hybrid Insect Micro-Electro Mechanical Systems project, is inserting computer chips into moth pupae, the stage between caterpillar and adult moth, with the hopes of creating insects with nerves that have grown around internal silicon chips, creating a "cyborg moth," so they can be controlled and used to take surveillance photographs.
Click Here to View Full Article
to the top

'Google 101' Class at UW Inspires First Internet-Scale Programming Courses
University of Washington News and Information (10/08/07) Hickey, Hannah

A University of Washington pilot course designed to teach students how to program using massive numbers of computers has been turned into a national program by Google and IBM. The new programming methods learned will help students and researchers manage Internet-scale applications. "This is a new style of computing in which the focus is on analyzing massive amounts of data, using massive numbers of computers," says University of Washington computer science and engineering professor Ed Lazowska. "Universities haven't been teaching it in part because the software is really complex, and in part because you need a big rack of computers to support it." The success of the pilot class has led Google and IBM to donate and manage hundreds of processors that students will be able to access for large-scale computing on the Web. The program was conceived when Christophe Bisciglia, a sensor software engineer at Google and a graduate of the University of Washington, noticed while interviewing potential Google employees that applicants were able to solve difficult problems that could be done on a single computer but were unable to solve more complex problems. Bisciglia designed the pilot program and served as course director during the test phase. "One of my big intentions was to close the gap between how industry and academia think about computing," says Bisciglia. The program will now be available to students at Carnegie Mellon, MIT, Stanford, UC Berkeley, and the University of Maryland, and has become a full-time job for Bisciglia.
Click Here to View Full Article
to the top

MIT Research Helps Convert Brain Signals Into Action
MIT News (10/02/07) Thomson, Elizabeth A.

MIT researchers have developed an algorithm to enhance prosthetic devices that convert brain signals into actions. The MIT approach unifies several different approaches used by experimental groups that created prototypes for neural prosthetic devices in animals or humans. "The work represents an important advance in our understanding of how to construct algorithms in neural prosthetic devices for people who cannot move to act or speak," says Lakshminarayan Srinivasan, lead author of a paper on the technique published in the October issue of the Journal of Neurophysiology. Previous efforts to create such devices have focused on boundaries related to brain regions, recording modalities, and applications. The MIT researchers used graphical models composed of circles and arrows that represent how neural activity results from a person's intentions for the prosthetic device they are using. The diagrams represent a mathematical relationship between the person's intentions and the neural manifestation of that intention, and could come from a variety of brain regions. Previously, researchers working on brain prosthetics have used different algorithms depending on what method they were using, but the new MIT model can be used no matter what measurement technique is being used, Srinivasan says. Srinivasan emphasizes that neural prosthetic algorithms still need significant improvement before such devices are available for common use, and that the MIT algorithm is unifying but not universal.
Click Here to View Full Article
to the top

UMass Amherst Researchers Improve Security for Credit Cards and Other Devices
University of Massachusetts Amherst (10/03/07)

University of Massachusetts Amherst researchers Kevin Fu, Wayne Burleson, and Dan Holcomb have created a cheap and efficient methodology for augmenting the security of radio-frequency identification tags. "We believe we're the first to show how a common existing circuit can both identify specific tags and protect their data," says Burleson, who presented the research at the annual Conference on RFID Security. "The key innovation is applying the technology to RFID tags, since they're such tiny devices with very small memories." Within the tags are passive systems that respond automatically to electromagnetic fields generated by radio antennas attempting to read the devices' memories, and this technology can be vulnerable to security breaches. The UMass Amherst researchers' security technique exploits the concept of random numbers, which are used to encrypt data transmitted by the tags, and a string of random numbers can be easily produced by machines with the appropriate hardware and software. However, RFID tags are not designed for random number generation, so the researchers' work takes specific machinery committed to that function out of the equation and instead employs special software that allows the tag readers to siphon out unique data from the tags' existing hardware. Variations in each tag's cells can also be tapped as individual tag identifiers, generating a unique fingerprint, Burleson says. The RFID Consortium for Security and Privacy is a collaborative effort between engineers and cryptographers that forms part of a research initiative underwritten by a $1.1 million National Science Foundation grant to enhance security for wireless "smart tag" gadgetry.
Click Here to View Full Article
to the top

UC San Diego Opens Doors to Internet2 Participants
UCSD News (10/02/07) Froelich, Warren R.; Ramsey, Doug

Some of the nation's leading researchers and educators in advanced networking are expected to gather at the University of California, San Diego for the fall 2007 Internet2 Network Performance Workshop, which will include tours of the San Diego Supercomputer Center and the California Institute for Telecommunications and Information Technology (Catlit2). The workshop will also have some of the latest innovations and applications in the Internet2 community on display. Some of the demonstrations at the conference will include delivering remote sensing data through Internet2, enabling virtual organizations, and video communication using scalable video coding. San Diego Supercomputer Center director Fran Berman will moderate a panel called "Cyberstructure: The Way Forward," and Catlit2 director Larry Smarr will deliver a keynote address focusing on "lambda networking." The conference will run from Oct. 8 to Oct. 11.
Click Here to View Full Article
to the top

FSU Researchers' Material May Lead to Advances in Quantum Computing
Florida State University (10/04/07) Ray, Susan

Florida State University scientists in the National High Magnetic Field Laboratory and the university's Department of Chemistry and Biochemistry have developed a new material that could provide a technological breakthrough and rapidly accelerate the development of quantum computing. The material is a compound made from potassium, niobium, oxygen, and chromium ions and could potentially be as important to computers of the future as silicon is to today's computers. High magnetic fields and radiation was used to manipulate the spins on the material to see how long the spin could be controlled. Based on experiments, the material could allow for 500 operations in 10 microseconds before losing its ability to retain information. "This material is very promising," says Naresh Dalal, a FSU professor of chemistry and biochemistry and an author of the paper describing the material. "But additional synthetic and magnetic characterization work is needed before it could be made suitable for use in a device."
Click Here to View Full Article
to the top

GTISC Releases Emerging Cyber Threats Forecast
Georgia Institute of Technology (10/02/07)

The Georgia Tech Information Security Center has published its annual forecasting report, the GTISC Emerging Cyber Threats Report for 2008, which describes the five key areas of security risk for enterprise and consumer Internet users. In 2008, cyber security threats are anticipated to grow and evolve in the areas of Web 2.0 and client-side attacks, such as social networking attacks, and targeted messaging attacks, including malware proliferation through video-sharing online and instant messaging attacks. Botnets, particularly the expansion of botnet attacks into peer-to-peer and wireless networks, are another significant area of concern. Threats aimed at mobile convergence, including vishing, smishing, and voice spam, are anticipated to be substantial, as are threats targeting RFID systems. The primary driver behind all five major threat categories in 2008 continues to be financial gain. GTISC recommends improved synchronization among the security industry, the user community, application developers, Internet service providers, and carriers. GTISC director Mustaque Ahamad anticipates that enterprise and consumer technologies will continue to converge in 2008, making it even more essential to protect new Web 2.0-enabled applications and the IP-based platforms they increasingly depend upon.
Click Here to View Full Article
to the top

Alternative Reality: UW Prof Touts Computer Game Learning
Wisconsin Technology Network (09/29/07) Vanden Plas, Joe

University of Wisconsin-Madison professor David Williamson Shaffer believes K-12 education is still centered around an industrial-era learning paradigm that teaches children basic skills for passing tests while short-changing them in terms of critical thinking and problem solving, and he advocates computer games as a tool for stimulating such thinking. "We've believed for 150 years to learn basic facts their first year, then do something more sophisticated, but computers allow us to do this before we master the basics," Shaffer explains. "Kids should learn these basic things just in time and on-demand. You need these skills, but you also need to learn them in a way that tells you why." UW-Madison is channeling resources into the concept of computer game education via its Academic ADL Co-Lab and its Games, Learning, and Society conference. Shaffer is a founding member of the GAPPS research group for games, learning, and society, and author of "How Computer Games Help Children Learn." He has examined the impact of new technologies on people's thinking and learning processes, with a concentration on epistemic games, which are computer and video games where players become professionals and are given the chance to tackle challenges in virtual work environments. Shaffer says the better kinds of computer games nurture a form of collaboration typical of the premier work places. "Children need to do things that are valued in a high-tech economy, and our schools are not very good at fostering innovative thinking," he says.
Click Here to View Full Article
to the top

Baby's Errors Are Crucial First Step for a Smarter Robot
New Scientist (10/06/07) Vol. 196, No. 2624, P. 30; Reilly, Michael; Robson, David

The comprehension of artificial intelligence could be improved through research in which machines experience the same errors of cognition that human infants do. For instance, University College London researchers recently announced their creation of a computer program that could be fooled by optical illusions, raising the possibility that robots imbued with human-like capabilities may consequently be prey to human weaknesses. Scientists are elated at the idea, as errors committed by humans performing one task often signify strong ability at another task. Therefore, an important step toward building true AI could be producing software and eventually machines that make human-like cognitive mistakes. "Intelligent behavior requires that you have stability--which you get from past experience--and flexibility so you can turn on a dime when you need to," observes Indiana University's Linda Smith. Put another way, human intelligence comes from an equilibrium between memory of past experience and adaptability to changing conditions. This theory lies at the root of the A-not-B error that babies make, and researchers attempted to recreate this error in software. The results of their work, presented at the European Conference on Artificial Life in September, indicate that a robot with the ability to learn from past experience not only makes the same errors as a human infant, but can also learn to adapt. The software programs with this ability were equipped with homeostatic networks, which may be the best tools for balancing stability and flexibility in robots.
Click Here to View Full Article - Web Link May Require Paid Subscription
to the top

To submit feedback about ACM TechNews, contact: [email protected]

To unsubscribe from the ACM TechNews Early Alert Service: Please send a separate email to [email protected] with the line

signoff technews

in the body of your message.

Please note that replying directly to this message does not automatically unsubscribe you from the TechNews list.

ACM may have a different email address on file for you, so if you're unable to "unsubscribe" yourself, please direct your request to: technews-request@ acm.org

We will remove your name from the TechNews list on your behalf.

For help with technical problems, including problems with leaving the list, please write to: [email protected]

to the top

News Abstracts © 2007 Information, Inc.