Association for Computing Machinery
Welcome to the May 24, 2010 edition of ACM TechNews, providing timely information for IT professionals three times a week.


Seven Atom Transistor Sets the Pace for Future PCs
BBC News (05/24/10)

A working transistor that contains only seven atoms has been built by a team in Australia. The researchers, led by University of New South Wales professor Michelle Simmons, developed the atomic transistor as part of a project to create a quantum computer. The atoms in a silicon crystal were replaced with phosphorus atoms using a scanning tunneling microscope. "Now we have just demonstrated the world's first electronic device in silicon systematically created on the scale of individual atoms," Simmons says. The working transistor was handmade, so there is a need for a process for developing them in large numbers. The researchers say the device could lead to chips that have components that are up to 100 times smaller than those on current processors. Simmons says the development could result in an "exponential" leap in processing power.

CS Career Projections
Computer Science Teachers Association (05/20/10)

The market for U.S.-based computing careers is expected to thrive, with the U.S. Bureau of Labor Statistics (BLS) projecting that computing will be one of the fastest-growing U.S. job markets in science, technology, engineering, and mathematics (STEM) for the foreseeable future. Almost 75 percent of new U.S. science or engineering jobs will be in computing, while just 16 percent will be in traditional engineering positions. The BLS predicts that 27 percent of the new STEM positions will be in software engineering, while new jobs in computer networking and systems analysis will greatly outnumber those in traditional engineering. Despite the abundance of jobs, there is a major decline of undergraduate computer science (CS) degree enrollments, and this is leading to a shortage of graduates. Fierce competition for CS graduates is also causing salaries for computing-related professionals to rise. Statistics indicate that annual STEM job openings through 2018 will come close to 140,000, while the number of college graduates with a degree in computing will barely top 40,000. In contrast, there is an oversupply of engineering, life sciences, mathematics, and physical sciences graduates.

The Automotive Internet, From Vision to Reality
ICT Results (05/21/10)

The European Union-funded CVIS project aims to deploy an advanced communications infrastructure that can facilitate new ways to drastically upgrade the safety, efficiency, and reliability of driving. Project coordinator ERTICO-ITS Europe has developed inexpensive technologies, and the rapid engagement of third-party application developers and service providers is a key aspect of the CVIS project's plan to cover new ground. Critical to the success of the project is organic growth, in which resources cluster naturally in specific areas where they can enable compelling applications with immediate impact. Other applications can later be deployed at these nodes, supporting a virtuous cycle that can set up a continuously broadening infrastructure installation and application development arc. In one field test, the CVIS solution involved trucks "booking" a loading-bay spot ahead of time and receiving a reservation time and duration. The reservations are handled by software and carried out via wireless nodes, and are automatically updated if the truck is ahead of or behind schedule.

DARPA's Self-Learning Software Knows Who You Are
Wired News (05/21/10) Drummond, Katie

New York University researchers Yann LeCun and Rob Fergus are developing Deep Learning, software that can teach itself to spot objects in a picture, actions in a video, or voices in a crowd. "The question we're asking is whether we can create computers that automatically learn feature sets from data," LeCun says. "The brain can do it, so why not machines?" The researchers say Deep Learning--which is backed by a $2 million, four-year U.S. Defense Advanced Research Projects Agency (DARPA) grant--will be inspired by biology, but not modeled after it. LeCun and Fergus are developing software that can identify an object on a first, unsupervised example, using layer upon layer of code to abstract the essential attributes of an object. DARPA also wants Deep Learning systems to recognize activities such as running, jumping, or getting out of a car. The final version of the system will operate unsupervised, by being able to hold itself accountable for errors and automatically correct those mistakes at each algorithmic layer. "Ideally, what we'll come away with is a 'generic learning box' that can identify every data cue," Fergus says.

Danger in the Internet Cafe?
University of Calgary (05/21/10)

University of Calgary (UC) computer science professors John Aycock and Mea Wang have identified a type of computer security threat, called Typhoid adware, that gains access to computers through wireless networks found in Internet cafes or other areas where users share non-encrypted wireless connections. "We're looking at a different variant of adware--Typhoid adware--which we haven't seen out there yet, but we believe could be a threat soon," Aycock says. Typhoid adware comes from another person's computer and convinces other laptops to communicate with it and not the legitimate access point. Then the Typhoid adware automatically inserts advertisements in videos and Web pages on the other computers. Aycock and Wang developed several defenses against Typhoid adware. One solution protects the content of videos to ensure that what users see comes directly from the original source, and another solution offers a way to "tell" laptops they are in an Internet cafe to make them more suspicious of contact from other computers.

Eye Tracking for Mobile Control
Technology Review (05/24/10) Grifantini, Kristina

Dartmouth College researchers have created EyePhone, an eye-tracking system that enables users to operate smartphones using their eyes. Dartmouth professor Andrew Campbell says his research team found that keeping track of a gaze using a mobile phone is much more difficult than on a desktop computer because both the user and the phone are moving, and the surrounding environment is so changeable. "Existing algorithms were highly inaccurate in mobile conditions--even if you are standing and there's a small movement in your arm, you'd get a large amount of blurring and error," Campbell says. The researchers developed an algorithm that learns to identify a user's eye under different conditions. EyePhone goes through a learning phase, in which the system is trained to identify a person's eye at varying distances and under different lighting. A user calibrates the system by taking a picture of the left or right eye both indoors and outdoors. Campbell says the system is 76 percent accurate in daylight while the user is standing still and 60 percent accurate when a person is walking.

Teaching a Computer to Win Human Friends and Influence People
CITRIS Newsletter (05/10) Slack, Gordy

University of California, Santa Cruz professor Marilyn Walker is studying human expression in an effort to devise algorithms that will enable digitally animated personae to reproduce those methods in order to express themselves more richly and compellingly. She and her colleagues are examining whether the combination of utterance, body language, and facial expression to those animations can augment or ameliorate an utterance's impact. Walker is experimenting with animated figures capable of expressing openness, conscientiousness, extroversion, agreeableness, and neuroticism in voice, body language, or facial expression. In an online study, participants witness and grade characters that are expressing different combinations of these traits. Walker plans to compile a core program that will let developers imbue their characters with "dynamic adaptation," or the ability to adapt automatically to different personality types and utilize a broad spectrum of responses while engaged in a conversation. The system will assess whom they are conversing with and assume the most appropriate conversational style to meet their goals or the goals of the humans they are talking to.

Computer Solves 400-Piece Jigsaw to Claim World Record
New Scientist (05/21/10) Marks, Paul

Massachusetts Institute of Technology (MIT) researchers have developed software that enabled a computer to solve a 400-piece jigsaw puzzle in three minutes--a new world record. The new probabilistic approach of Taeg Sang Cho and colleagues beat the time of Klaus Hansen at the University of Copenhagen in Denmark and offered improvements in solving jigsaw puzzles in other ways. The software is able to handle any image, including photographs of outdoor scenes, while the Danish team's computer only could work with simple cartoon pictures featuring clear shapes and a limited number of colors. The software is designed to analyze the predominant colors to determine the potential image of the puzzle, then refers to its database of existing images to try to arrange the pieces in their likely positions. Cho believes the algorithm could enhance image-editing products such as Photoshop. "If you move a person from one side of a picture to another, our algorithm could highlight the fact that the image will jar if the pixel values are too different," says Cho.
View Full Article - May Require Free Registration | Return to Headlines

Ultra-Secure Quantum Communications
University of New South Wales (05/20/10) Trute, Peter

University of New South Wales researcher Robert Malaney has used a new quantum communication process that enhances the unbreakable encryption security of quantum communication. "Unconditional location verification" limits access to a secure message to a recipient who must be at an agreed upon geographic point. "If they are not at that location the process would detect that and you can stop the communication," Malaney says. "This is a new application that you can deploy on current and emerging quantum networks." The process involves the sending of paired qubits over a fiber-optic or wireless network to a recipient, who must send a return message, using information from the decoded qubits, to several reference points to open up a secure channel. The amount of time it takes to return the message can be accurately measured because quantum networks operate at the speed of light and quantum information cannot be copied. This accurate measurement of return time ensures the message has come from the right place.

How Laptops Can Enhance Learning in College Classrooms
University of Michigan News Service (05/20/10) Moore, Nicole Casal

University of Michigan (UM) professor Perry Samson has developed LectureTools, an interactive student response system and teaching module that uses laptops to increase students' engagement, attentiveness, and participation in learning. In a survey, students who used LectureTools said the program made them feel more attentive, engaged, and able to learn, compared with students in classes that did not use the system. "Our surveys showed that while laptop computers can be a distraction, students of this generation feel that they are capable of productive multitasking," Samson says. The system allows students to take notes directly on lecture slides, anonymously ask questions through a chat window, and provide feedback on the quality of the lecture to the professor. "It is the first successful instance I've seen of dramatic use of information technology to augment the real-time classroom experience," says UM's John King. After using the LectureTools system, 78 percent of students said that laptops made them more engaged, while about half said the laptops made them more attentive, and 70 percent said laptops had a positive effect on learning.

New Version of Linux OS Includes Ceph File System Developed at UCSC
University of California, Santa Cruz (05/20/10) Stephens, Tim

University of California, Santa Cruz (UCSC) researchers have developed Ceph, an open source file system that allows multiple clients to access files stored on remote severs. "Our goal with Ceph was to build a scalable system that would allow storage of petabytes and beyond," says UCSC professor Scott Brandt. Although Ceph is still in development, it is being included in the latest version of the Linux kernel. "Having Ceph in the Linux kernel makes it much easier for people to use, so a lot more people will be testing it now and contributing to the project," says UCSC graduate student Sage Weil. A Ceph file system can be expanded by adding new servers to the system and it can automatically distribute data across the new servers. All data are copied across multiple storage devices, and if any device fails, the data are automatically re-replicated to other devices. "A Ceph storage system can be both big and fast, because all the operations are parallelized," Brandt says.

How Researchers Enhanced Using Semantic Technology
Government Computer News (05/18/10) Gallagher, Sean

A trio of Rensselaer Polytechnic Institute (RPI) researchers has tapped the power of the Semantic Web to augment the use of the Web site by converting some of the high-value data sets in the archive into demonstrations enabled by Resource Description Framework (RDF). RPI research scientist Li Ding says that following the data's conversion to RDF, it can be accessed by applications that use various standard Web technologies, such as JavaScript Object Notation. The researchers' work has made it possible to build apps from data in a matter of days rather than months, according to RPI's Dominic DiFranzo. "Anyone can use this technology--they don't have to be a graduate student to make this technology work," he says. The researchers say their innovation serves as an example of how fast and inexpensively visualization and mash-up applications can be put together from government data when it is rendered within a Web-friendly format.

Latest Computer Threat Could Be Worst Yet
University of Texas at Dallas (TX) (05/20/10)

Researchers at the University of Texas at Dallas (UTD) are studying the potential threat of "reactively adaptive" malware. UTD computer scientist Kevin Hamlen and data-mining expert Latifur Khan are developing ways to thwart such attacks, which are still hypothetical. "Today's malware mutates randomly in order to avoid detection, but reactively adaptive malware is more intelligent, learning and adapting to new computer defenses on the fly," Hamlen says. "What we've realized is that the same technology that goes into antivirus software could be turned on its head to make some viruses nearly unstoppable." Reactively adaptive malware would use the same algorithms that antivirus software uses to detect software, but would use them to stay a step ahead of antivirus defenses. The researchers plan to use data-mining techniques to enable antivirus software to update their databases quicker, which would allow the software to adapt before the malware can mutate.

Abstract News © Copyright 2010 INFORMATION, INC.
Powered by Information, Inc.

To submit feedback about ACM TechNews, contact: [email protected]
Current ACM Members: Unsubscribe/Change your email subscription by logging in at myACM.
Non-Members: Unsubscribe