Intel recently announced plans to launch a new family of microprocessors the company says will speed up and improve online, high-definition video on the Internet. Intel's Sean Maloney says the new chips' increased computing power would begin to transform the blurry, small clips online into high-resolution, full-screen programs that rival HDTV. The new series of processors, developed under the code name Penryn, will first be used in servers and high-end desktops that compress video. The chips will be the first manufactured using a new process that Intel says increases computing performance while reducing power consumption. The chips use a re-engineered transistor that is about half the size of its predecessor, allowing it to switch more quickly while requiring less power to switch, and preventing power loss. The new chips measure just 45 nanometers, and Intel says it will be able to put up to 820 million transistors on a single silicon die. The first products based on the new chips, marketed as the Intel Core 2 Extreme and Intel Core 2 Duo, will be available in the first quarter of 2008. For better video compression, Intel has added a set of 46 instructions, called SSE4, to the microprocessors. Steve Fischer, the lead designer of the new chips, says the instructions will create a new generation of servers that enhance the compression of digital video. "This is a step in the right direction," says analyst Richard Doherty, "and it's probably the best use for this 45-nanometer technology over the next couple of years."
Click Here to View Full Article
to the top

Newcomers Hit Top 10 in Supercomputer List
Nature (11/13/07) Stafford, Ned

The top spot on the TOP500 supercomputer list once again belongs to a computer at Lawrence Livermore National Laboratory (LLNL), but five of the top 10 spots on list were awarded to new machines. The LLNL supercomputer, which uses an improved version of IBM's Blue Gene/L system, computes at 478.2 teraflops per second, nearly three times faster than the second-place finisher. The supercomputer at the Julich Research Centre in Germany, which computes at 167.3 teraflops, was second on the list. The list highlights the increased use of supercomputing power by industry. About 57 percent of the TOP500 supercomputers are owned by businesses, and that percentage is expected to continue to increase, according to Jack Dongarra of the University of Tennessee, Knoxville, who is one of the four experts who compile the biannual list. "The trend is more and more into business and this trend will certainly continue," Dongarra says. Dongarra expected a supercomputer at the Texas Advanced Computer Center (TACC) at the University of Texas at Austin to take the top spot on the list, but it was not operational. He now expects the TACC supercomputer to be on the next list, due in June. Looking further ahead, Dongarra believes that an IBM supercomputer under development at Los Alamos National Laboratory in New Mexico will break the petaflop barrier, probably in 2009.
Click Here to View Full Article
to the top

Robotic Aids for the Disabled and Elderly
Pittsburgh Post-Gazette (11/14/07) Rotstein, Gary

People with limited mobility brought on by age or infirmity could be helped significantly by robotic systems under development by Pitt and Carnegie Mellon University, which received a five-year, $15 million grant from the National Science Foundation to establish the Quality of Life Technology Center. Technologies being developed include systems that monitor people's ability to carry out basic functions in their homes, identify decline, and then call attention to the problem so that assistance can be arranged; robotic devices that can assist with getting out of bed, preparing meals, and other tasks that become impaired as people's abilities deteriorate; "virtual coach" technology that helps people with waning mental skills perform daily tasks via a mobile device such as a watch or cell phone; and in-car programs that maintain mobility by observing people's driving ability, spotting potentially dangerous habits, and providing guidance on how to correct those habits. Such breakthroughs could extend the independence and productivity of disabled people and prolong their time out of nursing homes, which would amount to tremendous savings. Carnegie Mellon Robotics Institute professor Martial Hebert says creating a computerized system that records behavioral observations alone is a major challenge, and the Quality of Life Technology Center's goal is to employ such systems to anticipate future actions. To get feedback from clients on problems and other issues their devices face, the center has teamed up with long-term care providers and disability service agencies. The center also has business advisers because the cost of the potential products to consumers is substantial, while social scientists are enlisted to evaluate the willingness of people to interact with robotic devices and take instructions from them. The center will detail its progress on its assorted projects to the NSF in March.
Click Here to View Full Article
to the top

Up Next: Cameras That Know Who You Photographed
CNet (11/14/07) Shankland, Stephen

Cameras outfitted with face recognition technology could streamline the organization and retrieval of photos via autotagging, and Fotonation is attempting to tackle the computational challenges to realize such an advancement. Fotonation executive Eric Zarakov noted in an interview at the 6sight digital imaging conference that the challenge is simplified by the tendency among most people to photograph the same 25 or 30 persons, and that a camera could be taught to only recognize those specific persons. Riya is developing technology to search through online photo albums to recognize individuals, while Polar Rose is generating 3D facial models in an attempt to enhance recognition. Cameras could be configured to recognize locations through geotagging, which can be employed to seek photos whose location is known to the user and to determine what precisely is in a picture the user already has at hand. SiRF Technology founder Kanwar Chadha forecasts that in-camera GPS systems that facilitate automatic geotagging will eventually emerge. Marian Stewart Bartlett of the University of California-San Diego's Machine Perception Lab demonstrated her work on expression detection at the 6sight event. A camera cannot currently support expression detection, which requires a sophisticated computer run by people with advanced degrees. But such a milestone may not be too far off, in view of the fact that Sony has rolled out a camera that can detect smiles.
Click Here to View Full Article
to the top

The Ever-Changing Supercomputing Conference
HPC Wire (11/12/07) Vol. 14, No. 1,

Oak Ridge National Laboratory director of information technology services Becky Verastegui says that being chair of the Supercomputing Conference was one of the most challenging opportunities of her entire career. Verastegui, who has been involved with the conference for the past eight years, says more people are using supercomputers than ever before and things that were not thought possible even five years ago are currently being studied using supercomputer simulations. Among the new features at the SC07 conference is the Cluster Challenge, which was developed around the fact that computational power that is so easily accessible today significantly outperforms what was considered the best systems at national labs only 10 years ago. "In fact, a small cluster today--less than a half rack--would have topped the Top500.org Web site at that time," Verastegui says. "The Cluster Challenge will showcase the significance of this and highlight how accessible clusters are to anyone today." Several other new features at the conference include a Doctoral Research Showcase that will allow Ph.D. students graduating within the next 12 months to present a short summary of their research, and a Masterworks session that will include presentations on novel and innovative ways of applying advanced computing, communications, and storage technologies to achieve the breakthroughs needed to guarantee a competitive advantage.
Click Here to View Full Article
to the top

Study Compares States' Math and Science Scores With Other Countries'
New York Times (11/14/07) P. A17; Dillon, Sam

American students, even in low-performing states such as Mississippi and Alabama, perform better on math and science tests than students in most foreign countries, including Italy and Norway, concludes a new American Institutes of Research study. However, American students in even the best performing states such as Massachusetts are significantly outperformed by students in Asian countries such as Singapore and South Korea. "In this case, the bad news trumps the good because our Asian economic competitors are winning the race to prepare students in math and science," says the study's author and chief scientist Gary W. Phillips. The study compared standardized test scores of eight-grade students in all 50 states with those of their peers in 45 countries. Gage Kingsbury, a director at the Northwest Evaluation Group, which administered testing in 1,500 school districts, praised the study's methodology, but says it is difficult to compare scores internationally because in many countries children do not start school at the same age and often not every child attends the eight grade. Kingsbury says such differences mean it would be a mistake to infer too much about the relative success of the education system across the states and nations examined in the study. Education Sector co-director Thomas Toch says the real value of the study is that it provides a high-level perspective of the nation's education system. "It shows we're not doing as badly as some say," says Toch.
Click Here to View Full Article
to the top

SC07 Day 3: Programming Bits and Atoms
ZDNet (11/13/07) Burnette, Ed

Neil Gershenfeld, director of MIT's Center for Bits and Atoms, encouraged listeners to reconsider true statements such as two states make up binary information, during his keynote address Tuesday morning at SC07. Gershenfeld, an expert in digital fabrication, sees personal fabrication as the killer app of the field. From his perspective, it may no longer be about sending computation or energy around the globe, but rather the means for creating it. People prefer to measure and modify the world, and the shrinking scale of biological systems has the world on the cusp of a fabrication revolution, Gershenfeld says, adding that a computer is a tool and a program is a thing. During the Q&A, an audience member noted that, "If programs are things then bugs are consequential." Still, we should not be afraid of digital competition, Gershenfeld said. "Biological systems have been competing successfully for millions of years," he noted. The forward engineering of biology is already working for us, said Gershenfeld.
Click Here to View Full Article
to the top

Smart Phone Suggests Things to Do
Technology Review (11/13/07) Greene, Kate

Researchers at Palo Alto Research Center (PARC) have developed Magitti, software that allows mobile phones to suggest entertainment and dining options based on the user's previous choices and text messages. Magitti initially suggests some general activities, events, and dining options based on the time of day and the user's location, but as Magitti learns more about the user's behaviors and preferences it alters its recommendations. For example, if a user prefers eating lunch at an inexpensive place, but frequently goes to expensive restaurants for dinner, Magitti picks up on this trend by comparing the GPS location of the restaurant with a database of establishments and will make appropriate recommendations. Magitti also examines the phone owner's text messages to make recommendations, as text messages often include information on future plans. Magitti does raise some questions over privacy and security, which is why the system erases text messages after a certain amount of time, but PARC senior researcher Victoria Bellotti says that people will accept a trade-off between privacy and convenience once they see the benefits of the system, much like how consumers are willing to use credit cards even though it creates an electronic record of their movements and purchases. Magitti also needs to be better at distinguishing certain activities, because eating at a restaurant, grabbing a quick bite at a deli, and eating at home are all very different things but are currently categorized as eating in the system.
Click Here to View Full Article
to the top

More Security Education Needed to Avoid a Cybersecurity Disaster, Experts Warn
SearchSecurity.com (11/07/07) Westervelt, Robert

A panel of prominent security experts at the Information Security Decisions conference recently warned that although the United States is currently more prepared than ever for a major cybersecurity attack, more needs to be done to increase awareness about cybersecurity issues and better educate future IT professionals. "We need to provide resources for future problems," said Eugene Spafford, executive director of Purdue University's Center for Education and Research in Information Assurance and Security. "Patching the latest problem isn't getting us anywhere." The panelists agreed that it would probably take a major cybersecurity event for the public to become truly motivated enough to demand better security. The panelists also agreed that backdoor Trojan horse programs and herds of bots would continue to be a problem, but it is unknown if they will be used for isolated incidents for personal gain or to take down national electronic infrastructure. Businesses continue to focus on data protection and external attacks, a necessity as financial gain has become the primary motivation behind the majority of attacks according to the panel, but more needs to be done to protect against internal threats as they become a bigger problem. Spafford says there is a greater temptation for insiders and enterprises no longer have a typical perimeter, necessitating more defenses closer to valuable data. The panel praised vendors' efforts to better educate developers on safe coding practices and to spread best practices in the security development lifecycle. The panel did not call for federal regulations requiring vendors to develop more secure products, arguing that there is not enough public outcry for the government to enact such legislation, but did say that market forces are pushing vendors to enact more standards and to better educate their workforce on security issues.
Click Here to View Full Article
to the top

Finger-Friendly 'Tactile Interface' Could Aid Blind Computer Users
Johns Hopkins University News Releases (11/12/07) Sneiderman, Phil

A team of researchers from multiple universities is using a $330,000 grant from the National Science Foundation to develop a dynamic electronic surface that would allow blind or visually impaired people to "feel" mathematical graphs, diagrams, and other visuals displayed on computer screens. The prototype will be able to simulate relatively simple graphs, while future versions could eventually allow blind people to feel more complex images such as pictures and maps. The project is led by City College of New York assistant professor of chemical engineering Ilona Kretzschmar, while Johns Hopkins University research professor James E. West was chosen to be on the team for his extensive knowledge on how to move electrical charges through plastic or polymer materials. The interface's design features an electro-active polymer film that can slightly rise and possibly even wiggle when exposed to electrical signals. The device may also use sound feedback to help users move their fingers over the graph or diagram. "Eventually, if we can show this is feasible, I think this device will open up the world for people who are blind or visually impaired," West says. "The interface could help them sense contours and changes in shape and texture and use their fingers to perceive some of the computer images that people with normal vision take for granted."
Click Here to View Full Article
to the top

Human Error Puts Online Banking Security at Risk
Queensland University of Technology (11/07/07)

Improved security for online banking is unlikely to eliminate hacker attacks if customers do not do their part to protect their accounts, according to a new study from researchers at Queensland University of Technology. In its study on SMS systems, Mohammed AlZomai, from QUT's Information Security Institute, says usability and human error were more of a problem than technical security issues. Sending a one-time password via SMS to the mobile phone of a customer for each transaction has become a typical method for authentication, says AlZomai. However, customers often do not notice a discrepancy between the account number in the SMS message and the intended account number. When QUT changed five or more digits in the account number, the attack was successful 21 percent of the time, and when it altered only one digit the attack had a 61 percent success rate. "This is a strong indication that the SMS transaction authorization method is vulnerable," AlZomai said. "According to our study only 79 percent of users would be able to avoid realistic attacks, which represents an inadequate level of security for online banking."
Click Here to View Full Article
to the top

On Your Mark, Get Set, Develop: Google Offers $10M for Android Apps
TechNewsWorld (11/13/07) Noyes, Katherine

The Open Handset Alliance (OHA) has made the Android software development kit available for the new Android platform, and Google has announced it will sponsor a $10 million developer's challenge. "We believe that the Android platform offers developers a unique opportunity to create truly innovative mobile software," says Google's Andy Rubin. "We're challenging developers to stretch their imaginations and skills to leverage the full capabilities of this new platform and to create something amazing." Cash prizes ranging from $25,000 to $275,000 will be awarded to developers whose applications are picked by a panel of judges. Android is a mobile platform built on the Linux 2.6 kernel that offers an operating system, libraries, a multimedia user interface, and phone applications. Android's application model allows developers to extend, replace, and reuse existing software components to create integrated mobile services. The challenge is divided into two periods, with Challenge I accepting programs through March 3, 2008, and Challenge II launching after the first handsets built on the platform become available in the second half of 2008. Judging will be done by a panel of technology and mobile experts chosen from OHA member organizations and the industry as a whole. Awards will be given to developers whose applications utilize all that the Android platform offers in order to provide customers with compelling experiences, Google says. Developers will retain all intellectual property and all other rights to their applications.
Click Here to View Full Article
to the top

Google Scientist to Demo Quantum Computer
ZDNet UK (11/09/07) Judge, Peter

D-Wave Systems will bring in Google scientist Hartmut Neven to demonstrate what D-Wave says is the first practical quantum computer at this week's SC07 conference. Neven, an image recognition specialist, will show an image-recognition algorithm running on the device, which is supposedly the first commercially viable quantum computer, though many have their doubts as to the validity of that claim. D-Wave's system uses "adaibatic quantum computing," in which a device is designed to solve a problem and find an answer using a process called "annealing." D-Wave chief technology officer Geordie Rose says decoherence time is not a problem because the system can operate with thermal noise. In February, D-Wave demonstrated a 16-qubit computer called Orion, but scientists are skeptical that D-Wave demonstrated true quantum computing, particularly as no results have been published in peer-reviewed journals. "Over the last year, rather than answering scientists' questions about what, if anything, they've actually done that's novel, they seem to have descended ever further into the lowest kind of hucksterism," says MIT assistant professor of electrical engineering and computer science Scott Aaronson. Aaronson says Orion is probably a classical computer with 16 very noisy superconducting quantum bits, which let information into the system and act like classical bits. Rose says D-Wave has not had its system externally validated because the only meaningful measure of validation for technology is how it performs against systems currently in use.
Click Here to View Full Article
to the top

New Technology Can Be Operated by Thought
Science Daily (11/09/07)

Physically disabled people can compose and send emails and operate a television by thought thanks to advances in brain-machine interface (BMI) technology, and further breakthroughs may even make the mental operation of prosthetic limbs a reality in time. "By permitting the subject to adaptively recode the generated neural activity, the overall performance of the [BMI] device is dramatically increased," says the University of Pittsburgh's Andrew Schwartz. "Furthermore, as we have progressed in this work, it has become apparent that the basic idea of 'intention' during learning is very important and can be addressed by the direct observation of the neuronal transformations taking place during this fundamental processing." Thought-controlled operation of PCs by severely handicapped patients has been facilitated by a brain computer interface (BCI) developed by the Wadsworth Center in Albany, N.Y., and Wadsworth Center researcher Eric Sellers says the system can conceivably function with little technical oversight and offer significant improvements to communication and quality of life. Meanwhile, Washington University School of Medicine researchers have developed a BCI that allows individuals to mentally control a cursor on a computer screen, a wheelchair, and a robotic arm, and a current focus is the use of the BCI to improve the rehabilitation of stroke and brain injury patients using data demonstrating that one hemisphere of the brain can compensate for functions impaired by damage to the other hemisphere. The project involves repurposing the interface to respond to signals from only one hemisphere. University of Chicago graduate student Dennis Tkach and fellow researchers are concentrating on the modification of BMI systems for use by paralysis victims through the harnessing of congruent neural activity.
Click Here to View Full Article
to the top

Graphics Chips Rev Up Research
BBC News (11/09/07) Ward, Mark

Scientific researchers are starting to realize that graphics cards are a cheap source of supercomputer-level processing power. "They give a phenomenal bang for the buck," says University of Oxford professor of scientific computing Mike Giles. Giles says the way graphics cards are built makes them very good at the repetitive computational tasks often used by scientists to test theories, models, and predictions. University of Wisconsin-Madison professor of electrical and computer engineering Susan Hagness is using microwaves to scan tissue for cancers, then processing the information through a mini computational cluster made out of several graphics cards. The system allows the results of a screening mammography to be returned in a few hours rather than days. Giles, who uses graphics processors for financial modeling, says the chips excel at doing the same thing many different times, as opposed to traditional chips that are better at doing numerous different things at the same time. He says graphics cards have more processing cores than traditional chips and each of the cores could run one of the repetitive simple tasks. The financial models Giles runs tests the same algorithm on each core, but each core gets different random numbers as input. Additionally, each individual core on a graphics card is just as good at handling simple tasks as traditional processors. "Each core is logically very simple but its floating point capability is the same as an Intel chip," Giles says. Graphics card manufacturers are also making it easier for scientists to use their products, with graphics card maker Nvidia recently releasing software to help researchers write code for their cards.
Click Here to View Full Article
to the top

Researcher Models Effects of a Suicide Bombing: Results of Crowd Configurations
Florida Institute of Technology (10/30/07)

A crowd formation in which people stand or sit in vertical rows is the safest way to minimize injuries and fatalities if a pedestrian suicide bomber were to strike, according to Zeeshan-ul-hassan Usmani, a computer science doctoral student and Fulbright Scholar at the Florida Institute of Technology. Usmani used virtual simulation to come up with his findings, which are preliminary. People in a crowd in the form of a circle are at the greatest risk, considering Usmani discovered the formation has a 51 percent death rate and a 42 percent injury rate, with an effectiveness of 93 percent. Usmani plans to model physical objects such as landscape and furniture and such social variables as crowd behaviors and add these variables to the simulation. Emergency response and counter-terrorism activities stand to benefit from his research, which he plans to present at the Interservice/Industry Training, Simulation, & Education Conference in Orlando, Fla., on Nov. 27.
Click Here to View Full Article
to the top

The State of the Art in Machine Conversation: HAL's Still Pure Hollywood
Ars Technica (11/08/07) Timmer, John

Yorick Wilks, a professor of computer science who directs the Institute for Language, Speech, and Hearing at the University of Sheffield, has written a perspective in the Nov. 8 issue of Science that addresses the latest developments in human-computer conversation. More experts are taking a cue from PARRY, in making statistical analysis of large samples of spoken and written conversations a priority. Google has taken this approach with its new translation service. Though statistical models enabling computers to perform statistical analysis of language in use, ultimately allowing machines to "teach" themselves language, continues to improve, progress is also being made on speech recognition and the recognition of personal and general history. In humans, the latter capability enables them to pick up a contradiction from an earlier statement, which may be the result of a change in opinion or even a joke. The Defense Advanced Research Projects Agency's CALO project is an early attempt to move in this direction. Nonetheless, the general conversation abilities of HAL will take some time.
Click Here to View Full Article
to the top

China Special: Quantum Revolution
New Scientist (11/07/07)No. 2629, P. 68; Huang, Gregory T.

Building a quantum computer by combining quantum memory with the new cluster states architecture is the goal of Jian-Wei Pan of the University of Science and Technology of China. He believes this technique could be scaled up to execute useful calculations with greater ease and reliability than any other scheme developed thus far. Pan is convinced that the harnessing of photons for quantum communication is the best approach, and cluster states are designed to solve the problem of maintaining the stability of quantum entanglement between quantum bits (qubits). In a cluster states scheme, each step of a calculation has its own set of qubits, eliminating the need to manipulate and thus risk disrupting the entanglements during the calculations. The problem is that many more qubits must be entangled before the calculation begins, and these entanglements must be saved for longer than is necessary with other methodologies. Pan's team is focused on addressing these challenges, and in May they disclosed the development of the first four-qubit cluster state using just two photons. Quantum memory serves to keep track of many photons' states to make the approach scalable, and Pan and colleagues are aiming to use this kind of storage to establish entanglements of 100 or more qubits so that practical quantum computing can be initiated.
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 be removed from future issues of TechNews, please submit your email address where you are receiving Technews alerts, at:
http://optout.acm.org/listserv_index.cfm?ln=technews

To re-subscribe in the future, enter your email address at:
http://signup.acm.org/listserv_index.cfm?ln=technews

As an alternative, log in at myacm.acm.org with your ACM Web Account username and password, and follow the "Listservs" link to unsubscribe or to change the email where we should send future issues.

to the top

News Abstracts © 2007 Information, Inc.