Welcome to the February 18, 2011 edition of ACM TechNews, providing timely information for IT professionals three times a week.
Please note: In observance of the Presidents Day holiday, TechNews will not be published on Monday, Feb. 21. Publication will resume Wednesday, Feb. 23.
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.
HEADLINES AT A GLANCE
At AAAS 2011: Taking Brain-Computer Interfaces to the Next Phase
Ecole Polytechnique Federale de Lausanne (02/17/11) Michael David Mitchell
Ecole Polytechnique Federale de Lausanne (EPFL) researchers are designing a brain-computer interface (BCI) that can learn from the user and account for rest periods and multitasking. In standard BCIs, users can give the system one of three commands--either left, right, or no-command. However, the no-command is very difficult to maintain and calls for extreme user concentration. The EPFL researchers, led by professor Jose del R. Millan, recently conducted a BCI experiment in which they asked volunteers to read, speak, or read aloud while executing as many left, right, and no-commands as possible. The EPFL team found that the BCI could distinguish between left and right commands versus the no-command, meaning the system learned to understand the user's mental intention, enabling users to mentally relax while using the BCI. Millan's work complements the shared-control approach to human-robot interactions, which uses sensors and image processing to help users avoid obstacles. The EPFL method involves decoding the signals coming from electroencephalography readings on the scalp and analyzing the data using statistical analysis and probability theory.
Computer Wins on Jeopardy!: Trivial, It's Not
New York Times (02/17/11) John Markoff
IBM's Watson supercomputer concluded its third and final televised round of Jeopardy! on Wednesday in triumph, defeating the human players and winning the three-day tournament. Watson finished the three rounds with $77,147, while the two other contestants won $24,000 and $21,600. Watson proved very proficient at buzzing in quickly to answer questions--a reflection of its confidence in its answers--and its victory was a vindication for computer science and the notion of developing a thinking machine. The supercomputer excels at parsing language. For example, it responded to "A recent best seller by Muriel Barbery is called 'This of the Hedgehog,'" with "What is Elegance?" IBM plans to announce a collaborative venture with Columbia University and the University of Maryland to develop a doctor's assistant service based on the Watson technology, which will permit physicians to ask questions of a cybernetic assistant. Another collaboration with Nuance Communications will strive to add voice recognition to the assistant, possibly making the service available in as soon as 18 months. IBM executives also are discussing the development of a version of Watson that can engage with consumers on various topics such as buying decisions and technical support.
Robots Learn Human Perception
Max Planck Gessellschaft (02/17/11) Tim Schroder
Instilling computers with the ability to generalize and abstract their surroundings is the goal of Michael J. Black, founding director of the Max Planck Institute for Intelligent Systems. "I want to teach a robot to be as familiar with the world as a two-year-old child," he says. Key to fulfilling this ambition is determining which environmental stimuli are important to generalization and abstraction, and Black is primarily concentrating on visual perception--specifically movements. His colleagues in the United States planted electrodes in the brains of paraplegic patients to measure nerve cell stimulation. Black identified and interpreted clear activation samples from the nerve cells' electrical impulses and enabled a computer to translate those signals into the movement of an onscreen cursor. Black plans to apply insights from such analysis to computer programming. "I would like to understand the neuronal control of movement, understand the models after which the brain is patterned, and then transfer these basic principles to the artificial world," he says. Black has developed statistical computing processes that reduce the complexity of environmental stimuli to a required range, to gradually approximate the environment.
Printed Photos the Blind Can 'See'
Discovery News (02/16/11) Alyssa Danigelis
The International Conference on Intelligent User Interfaces featured a demonstration of software that created recognizable images for the blind. Arizona State University professor Baoxin Li invited attendees to have their pictures taken, then used his team's computer application to automatically generate printable tactile versions. "We convert the photo in such a way so the major facial landmarks are nicely kept--that's very important because we can't render all the features into tactile form," Li says. The software is designed to work with a special tactile printer to produce images with raised lines along the facial features. "It's a tour de force that he can analyze a face and make it feel like a face," says John Gardner of ViewPlus Technologies, which developed the tactile printer used by Li. It currently takes about one minute to generate a tactile photo of a face, but the software could be optimized to work faster. Li says the software will work with paperless tactile displays that are in development. The researchers also are considering creating software that can generate tactile images from online mapping sites.
IBM's Watson's Ability to Converse Is a Huge Advance for AI Research
Computerworld (02/15/11) Sharon Gaudin
IBM's Watson supercomputer represents a major advance for artificial intelligence (AI) research through its ability to answer verbal questions posed by people. IBM researchers and industry analysts say this skill makes the system better equipped than any previous machine to organize its responses and engage in verbal conversation with humans. "To reach [a computer] conversationally and have it respond with knowledgeable answers is a sea change in computing," says analyst Richard Doherty. The goal of AI researchers to create a computer capable of mimicking human intelligence took a step forward this week with Watson competing against human players on the game show Jeopardy! "Watson is a significant step, allowing people to interact with a computer as they would a human being," says IBM researcher Jennifer Chu-Carroll. "Watson doesn't give you a list of documents to go through but gives the user an answer." The Jeopardy! match especially showcased the confidence the IBM researchers instilled within the supercomputer, as demonstrated by its frequent buzzing in. Chu-Carroll anticipates computers learning to use actual common sense within the next decade, and Doherty believes such advances will revitalize competition between AI researchers.
Cryptographers Debate Role in 'Post-Security' Era
EE Times (02/16/11) Dylan McGrath
Network and computer security has undergone a paradigm shift from a concentration on keeping intruders out to an assumption that adversaries can and do compromise networks, said U.S. National Security Agency Information Assurance Directorate technical director Dickie George at the RSA Conference. He emphasized that contemporary network monitoring involves maintaining vigilance for uncharacteristic or inappropriate behavior. George and others on the conference's Cryptographers' Panel said that cryptography, despite its limitations, remains the best available tool for guaranteeing network security. Stanford University professor and panelist Martin Hellman cited Cryptography Research's work in identifying the threat of differential power analysis attacks as an example of clear, tangible benefits of ongoing cryptographic research. Massachusetts Institute of Technology professor Ronald Rivest said "the problem [cryptographers] are facing is the rash of technology development. We keep building fences, but the universe keeps growing." Weizmann Institute of Science professor Adi Shamir noted, for instance, that cryptography could not have prevented the Stuxnet computer worm attack or the WikiLeaks disclosure.
Algorithms Can't Solve CS Gender Gap
Brown Daily Herald (02/17/11) Sahil Luthra
The percentage of women in computer science (CS) is one of the lowest of all science, technology, engineering, and math disciplines across the United States, and Brown University and other schools have recently taken steps to understand and reverse this trend. Brown's CS department strives to ensure introductory courses have female teaching assistants and examines which types of assignments are most intimidating to women. Brown's Tom Doeppner notes that a key problem is a lack of interest in CS among incoming female students rather than disenchantment with the field after they arrive. Brown professor Shriram Krishnamurthi believes a misunderstanding about the nature of CS is a bigger problem, and he largely blames high school programs. One school that has significantly increased its cohort of female CS students is Harvey Mudd College, and college president Maria Klawe credits, among other things, the CS department for offering many summer research opportunities to female rising sophomores. In addition, Harvey Mudd has reorganized its introductory CS class, changing the concentration from Java programming to computer-based problem solving. Klawe says the college also began inviting female freshmen to a CS conference in which an overwhelming majority of attendees are women.
A World Wide Web That Talks
Technology Review (02/16/11) Tom Simonite
IBM researchers have enhanced the Spoken Web, which is operated by voice over the telephone and attempts to recreate the text-based Internet for people in developing countries with low literacy levels and minimal technical skills, by adding a search feature to help users find information. "As the number of voice sites grows, and they get more content, people need a way to find what they want quickly," says IBM Research India's Nitendra Rajput. Although voice-recognition technology can take a user's search term and compare it to a database of voice sites, presenting the search results back to the user is more difficult. Instead of the system reading out every search result to the user, the researchers designed a system that tells the user how many search results there are and then asks how to narrow down the search further. When the number of sites is narrowed down to five or fewer, then all are read to the user, who must choose which one to select. The researchers tested the new system on 40 farmers in the Indian state of Gujarat, and had promising results, encouraging IBM to roll out it across the entire Spoken Web. "If there is relevant information on the real Web, we can pull it in to the spoken Web using API calls and text-to-speech technology," Rajput says.
Running on a Faster Track
American Friends of Tel Aviv University (02/16/11)
Tel Aviv University researchers have developed a scheduling tool that will shorten the travel time for passengers who use commuter trains. The Service Oriented Timetable is designed to use computers and algorithms to schedule trains in a way that is fair to all commuters. Professor Tal Raviv, assisted by graduate student Mor Kaspi, studied train timetables to determine how to optimize the scheduling system so passengers make it to their destinations faster. Train planners traditionally focus on the frequency of trains passing through certain stops, but Raviv and Kaspi take into consideration the total travel time of passengers, including their waiting time at transfer stations. "We can find a way to synchronize trains to minimize the average travel time of passengers," Raviv says. Raviv and Kaspi simulated their system on Israel Railway, and it reduced the average travel time per commuter from 60 minutes to 48 minutes. "Our solution is useful for any metropolitan region where passengers are transferring from one train to another, and where train service providers need to ensure that the highest number of travelers can make it from point A to point B as quickly as possible," Raviv says.
Carbon Nanotube Transistors Could Lead to Inexpensive, Flexible Electronics
PhysOrg.com (02/16/11) Lisa Zyga
A new way of fabricating carbon nanotube networks could make high-performance, flexible, transparent devices such as e-paper and radio frequency identification tags a reality. Researchers at Nagoya University and Aalto University have optimized the characteristics of metallic and semiconducting nanotubes by fabricating a nanotube network with unique properties. The network has a morphology that consists of straight, relatively long nanotubes, and uses more Y-junctions than X-junctions between nanotubes. The team used the nanotube network to fabricate thin-film transistors (TFTs) that simultaneously demonstrate a high charge-carrier mobility and on/off ratio, offering much better performance than previous carbon nanotube-based transistors. The researchers also fabricated an integrated circuit (IC) capable of sequential logic, which is the first based on carbon nanotube transistors. The researchers say that scaling up the fabrication process and using better printing techniques could lead to the development of large-scale, inexpensive, and flexible electronics. "Our near-future plan is to demonstrate roll-to-roll fabrication of CNT-based TFT arrays and ICs," says Nagoya's Yutaka Ohno.
Obama Sets $126M for Next-Gen Supercomputing
Computerworld (02/17/11) Patrick Thibodeau
President Obama's 2012 budget proposal calls for $126 million for the development of next-generation exascale supercomputers, with about $91 million going to the U.S. Department of Energy's (DOE's) Office of Science and $36 million going to the National Nuclear Security Administration. The funding is part of a general DOE advanced computing request of $465 million for 2012, which marks a 21 percent increase over the 2010 budget. Exascale systems will beat the power of the fastest current supercomputer by 1,000-fold, and the White House's funding for such systems reflects its plan for a predictable future pathway for high-performance computing. The creation of an exascale system is expected by 2020, but that depends on the development of software systems that can use what may amount to 100 million cores. Meanwhile, DOE is constructing 10 petaflop systems. Modeling and simulation are the chief supercomputing applications, and with an increase in system size comes a gain in resolution. Faster networking and other technological milestones that must be achieved to build exascale systems may eventually migrate to business-class servers.
Build Your Online Networks Using Social Annotations
EurekAlert (02/16/11) Shigeaki Sakurai
Toshiba researchers are studying ways to unite groups of online users using social annotations, including the tags, keywords, comments, and feedback that content creators and consumers submit to online networks. The researchers say the three-step approach could be used to get more reliable search engine results and to develop more effective targeted marketing strategies. Photo-sharing Web sites and social-bookmarking sites use social annotations to differentiate between digital objects. Toshiba researchers Shigeaki Sakurai and Hideki Tsutsui say the social annotations also can be used to develop networks. The researchers divide online users into groups according to their interests, with each group receiving a different annotation. First, the system identifies the subject of interest, followed by calculating the similarities between target objects discussed in blog posts based on social annotations. Finally, the system calculates how the target objects are related based on impression words in the blog posts.
W3C: HTML5 Spec Due in 2014
eWeek (02/16/11) Darryl K. Taft
The World Wide Web Consortium (W3C) recently released a stricter timeline for the HTML5 specification, the main component of its Open Web Platform for application development. W3C made May 2011 the last call date and 2014 the deadline for a final full HTML5 specification. The last call phase in 2014 will enable the HTML Working Group to test the new specification's technical reliability. After the last call phase, the Working Group will focus on obtaining implementation experience by developing a comprehensive test suite. The 2014 deadline is part of a larger timeframe that releases the final specification in 2022. "The decision to schedule the HTML5 Last Call for May 2011 was an important step in setting industry expectations," says W3C's Jeff Jaffe. Although the W3C has cautioned developers not to utilize HTML5 capabilities too soon, it does want them to use them when appropriate, such as with Apple's iPad and Microsoft's Internet Explorer 9 browser technology. The hard deadline for HTML5 standardization will lead to broad interoperability of new Web technologies over different platforms, according to W3C officials.
Abstract News © Copyright 2011 INFORMATION, INC.
To submit feedback about ACM TechNews, contact: [email protected]
Current ACM Members: Unsubscribe/Change your email subscription by logging in at myACM.