Association for Computing Machinery
Welcome to the April 26, 2013 edition of ACM TechNews, providing timely information for IT professionals three times a week.

ACM TechNews mobile apps are available for Android phones and tablets (click here) and for iPhones (click here) and iPads (click here).


A Messenger for the Internet of Things
New York Times (04/25/13) Steve Lohr

A group of technology companies will unveil the Message Queuing Telemetry Transport (MQTT) protocol to ease the challenge of enabling smart devices to communicate with one another in an Internet of things. Enabling all devices to communicate amongst themselves remains a critical challenge. The open standard, released through international standards organization Oasis, serves as a messenger and carrier for data exchange in a manner that proponents compare to the Web's Hypertext Transfer Protocol. MQTT was co-invented by IBM's Andy Stanford-Clark, who experimented with automating numerous devices in his own home. To link the devices, Stanford-Clark wrote his own code for machine-to-machine data communication. Through an IBM industry partnership Stanford-Clark met embedded systems expert Arlen Nipper, and the two wrote the first version of MQTT in 1998. Stanford-Clark and Nipper have refined MQTT over the years, and the standard is now used by many companies, including Facebook, which uses MQTT for the live notifications it sends on devices running Apple’s iOS software. The standardization process is a significant endorsement for MQTT as demand for machine-to-machine communications grows. However, for MQTT to advance on the industrial front, it must move beyond the information technology-centric Oasis consortium and win backing from industrial technology firms.

Google Predicts Stock-Market Crashes, Study Suggests
Live Science (04/25/13) Stephanie Pappas

University of Warwick researcher Tobias Preis and colleagues have found a method of predicting whole stock market shifts based on publicly available data on Google Trends search terms. The team tracked 98 search terms such as "debt" and "derivatives" from 2004 to 2011, and compared the searches to Dow Jones Industrial Average closing prices. To test whether the terms searched in the week prior to any given closing day could predict market direction, the researchers developed an investing game. If searches for financial terms went down, they opted to buy stocks and hold them with the expectation of value rising, and if searches went up, the researchers would "short" the market by selling stocks they did not own, with the intention of buying later at a lower price. The researchers theorize that if investors feel concerned about the stock market, they will search for financial information before selling stock. The researchers were able to increase their mock portfolio by 326 percent by holding onto stock when searches on the word "debt" decreased and shorting the market when searches increased. Beyond predicting the stock market, online activity on Google and other sources, such as Wikipedia, could aid in forecasting disease spread, civil unrest, and election outcomes.

Indiana University Supercomputer Aims for Big Ideas
USA Today (04/23/13) Stephanie Wang

Indiana University recently launched Big Red II, a $7.5-million supercomputer that aims to open the world of big data to every Indiana University researcher and student. Craig Stewart, executive director of Indiana University's Pervasive Technology Institute, says Big Red II is the fastest supercomputer owned and funded by a U.S. university and is likely one of the world's 50 fastest. Stewart says Big Red II can analyze digital texts, simulate the creation of stars, and sequence 5,000 genomes. In addition, Big Red II will support research in the life sciences, the humanities, and more than 150 other disciplines. "We don't really know all that we're going to enable," Stewart says. The machine has its own water-cooling system, as well as virus and hacker protection. To date, 75 researchers have expressed interest in using the machine for various projects. "We want as many different people to do as many different things that they've never done before," says Matt Link, Indiana University's director of research technology systems. Big Red II is 25 times faster, 50 percent smaller, and slightly cheaper than Big Red I, according to the university.

Deep Learning
Technology Review (04/23/13) Robert D. Hof

Artificial intelligence (AI) is advancing as tremendous computational ability enables computers to recognize objects and translate speech in real time. Renowned inventor and machine-intelligence futurist Ray Kurzweil met with Google CEO Larry Page last July to discuss plans to start a company geared toward developing a intelligent computer capable of understanding language and making independent inferences and decisions. Instead, Page convinced Kurzweil to join Google as director of engineering. Particularly attractive to Kurzweil were Google’s computing resources and work on a branch of AI called deep learning, which attempts to imitate activity in layers of neurons in the neocortex region of the brain in which 80 percent of thinking occurs. Deep-learning software learns to recognize patterns in digital representations of data such as sounds and images. Although the idea that software can mimic the neocortex’s large array of neurons in an artificial neural network has been around for decades, improvements in computing power and mathematical formulas now enable computer scientists to significantly increase the layers of virtual neurons in the model, leading to great leaps forward in speech and image recognition. For deep learning to expand into applications beyond speech and image recognition, progress must be made on the conceptual and software fronts, as well as with processing power.

European Students Need Better Tech Training, Study Says
InformationWeek (04/23/13) Gary Flood

Just 25 percent of European nine-year-olds attend a school with up-to-date information and communications technology (ICT) equipment, according to a recent European Commission (EC) study. Moreover, the study found that 20 percent of students age 11 and older have never used a computer as part of a school lesson. Although European students and teachers are eager to engage with the digital revolution, the use of both ICT and digital skills are spread too unevenly across the continent, the EC study concludes. "Europe will only resume sustained growth by producing highly skilled ICT graduates and workers who can contribute to innovation and entrepreneurship," says European commissioner Androulla Vassiliou. The study notes that although ICT equipment in schools has doubled since 2006, it is not evenly distributed across all 27 member states. The Scandinavian and Nordic countries have the most access to ICT equipment, while the Eastern and Southern parts of the continent have less access. "We want our young people exposed to ICT in school from the very beginning, and we want teachers who are confident to share their knowledge," says European Union commissioner Neelie Kroes.

The Digital Public Library of America: Adding Gravitas to Your Internet Search
Ars Technica (04/21/13) Megan Geuss

The recently launched Digital Public Library of America (DPLA) enables users to browse more than two million archived books, images, records, and sounds. Institutions such as Harvard University, the Internet Archive, and the Boston, Chicago, and San Francisco public libraries have all made content available through DPLA. In addition, DPLA offers an application programming interface (API) to users who want to add access to a third-party application. The library's largest obstacle now is gaining the participation of all relevant organizations, from local libraries to big publishers. Other challenges include organizing metadata to make it easy to use and getting small-town libraries and archives online and connected to the DPLA API. Forty-two state and regional libraries currently have digitized all or part of their archives, and seven of those are searchable through the DPLA site. The level of state and regional library access to which the DPLA aspires would make history available in a new way, enabling users to search local newspapers for great-grandparents’ names, for example, with open access enabling large-scale downloading. Eventually, DPLA must address the issue of copyrighted works with publishers, because waiting 70 years or more for contemporary works would curtail its usefulness.

DHS Creates Cyber Internships for Community College Students, Veterans (04/22/13) Brittany Ballenstedt

A new U.S. Homeland Security Department honors program will help introduce community college students and veterans to the field of cybersecurity. The Cyber Student Initiative will begin at Immigration and Customs Enforcement computer forensic labs in 36 cities nationwide. Participants will receive training and gain hands-on experience within the department's cybersecurity community. The unpaid volunteer program is part of the Secretary Honors Program announced last fall. "The Secretary's Honors Program Cyber Student Initiative will offer current students the opportunity to learn more about the critical skills and experience needed to succeed and protect our nation from evolving cyberthreats," says DHS Secretary Janet Napolitano. The program is open to community college students and veterans pursuing degrees in cybersecurity. Applications will be accepted until May 3, and the student volunteer positions will begin in June.

Can Super Mario Save Artificial Intelligence?
New Yorker (04/22/13) Gary Marcus

Human cognition is limited in terms of memory, logic, and arithmetic ability, but the brain's flexibility allows people to handle tasks that flummox computers. For example, supercomputers can be designed to master a single game such as chess, but computers cannot easily learn new games--a skill at which even children are adept. However, Thomas Walter Murphy VII earlier this month unveiled a program that can learn multiple games without any specialized prior knowledge. Available online, the program defeats original Nintendo games such as Super Mario Brothers, learning tricks and trying strategies. As with all artificial intelligence (AI), the program has difficulty scaling up, in this case to more sophisticated games. Researchers wonder whether an AI system could approximate human flexibility in learning new tasks, and current efforts focus on using big data to imitate humans who know how to complete the task. However big data breaks down when mimicry is not an option, and real-world situations are often nuanced and open-ended. One approach to making AI more flexible would be to examine the human gift of analogy, say authors Douglas Hoftstadter and Emmanuel Sander. They note that people can apply knowledge of one first-person shooter game to a similar game, for example, whereas a computer will approach each game as entirely new.

Learning Software Development--By Developing Software
MIT News (04/24/13) Larry Hardesty

The Massachusetts Institute of Technology’s (MIT) Department of Electrical Engineering and Computer Science is offering a new course that involves students in large, ongoing, open source software development projects with assistance from industry professionals. As in real-world software development, students work online with geographically-dispersed partners, who are students from about 15 universities worldwide. The concept originated with Stanford University computer science lecturer Jay Borenstein, who obtained funding from Facebook, and formed a working group of educators to create courses and other educational initiatives around development projects. Borenstein also collaborated with members of the open source software community to determine which problems should be addressed, and recruited industry mentors. Working on actual development projects helps students acquire practical skills and engages students in a way that traditional courses do not. The course offers some lectures tailored to individual programming work, but students primarily work with one another. When a section of code is complete, students log it into the open source project’s online code repository, where it is reviewed. This level of individual attention is possible because the course only has 11 students, but MIT is considering ways to scale the class up.

Computing a Better Way to Advertise
Daily Pennsylvanian (04/21/13) Fiona Glisson

A team of University of Pennsylvania students has created Advise, a video billboard that tailors the advertisements it displays to the people walking by based on their age, gender, glasses, and facial hair. A Kinect device on top of the billboard screen periodically takes photos of passersby. The billboard then crops out everything in the photo except for the people’s faces, and uses the Betaface database to determine their age and gender. Advise can either target one person who is looking directly at the advertisement or find the average characteristics of a crowd and target an ad to the group. "Online, there are systems like Google and Facebook which track based on...your age and gender and where you are from and whatever websites you have visited, but there is no real corollary to that in the real world," says Pennsylvania student Jeff Kiske. He notes privacy was a concern when designing Advise, so the system is programmed to periodically destroy all of the pictures it takes. “Anything you are outwardly showing to the public, we think is OK to use to target an ad,” Kiske says. “Any deeper connections, like your Facebook profile, we’re staying away from."

World's First Smartphone for the Blind, Made in India
Times of India (04/19/13) Chitra Unnithan

National Institute of Design post-graduate student Sumit Dagar has led the development of a smartphone for the visually impaired. The prototype features a touchscreen that converts text and pictures into Braille and raised patterns. "We have created the world's first Braille smartphone," Dagar says. "This product is based on an innovative touchscreen, which is capable of elevating and depressing the contents it receives to transform them into touchable patterns." The device uses Shape Memory Alloy technology, which is based on the concept that metals remember their original shape, or expand and contract to their original shape after use. The smartphone's screen has a grid of pins that move up and down as needed. The grid also has a Braille display, where pins come up to represent a character or letter. The screen itself can elevate and depress the contents to form patterns in Braille. "The response during the test has been immense. It comes out as a companion more than a phone to the user," Dagar says. "We plan to do more advanced versions of the phone in the future."

Robot Hands Gain a Gentler Touch
Harvard University (04/18/13) Caroline Perry

Harvard University researchers have developed TakkTile, an inexpensive tactile sensor for robotic hands that they say is sensitive enough to turn any machine into a dexterous manipulator. “Despite decades of research, tactile sensing hasn’t moved into general use because it’s been expensive and fragile," says TakkTile co-creator Leif Jentoft. "The traditional technology also uses very specialized construction techniques, which can slow down your work. Now, Takktile changes that because it's based on much simpler and cheaper fabrication methods." TakkTile is designed to give commercial inventors, teachers, and robotics enthusiasts access to high-end technology. TakkTile features a tiny barometer which senses air pressure and adds a layer of vacuum-sealed rubber to it, protecting it from as much as 25 pounds of direct pressure. The sensor enables robots to know what they are touching. During testing, a TakkTile-equipped robot was able to pick up a balloon without popping it and pick up a key and use it to unlock a door. "Not everyone has the bandwidth to do the research themselves, but there are plenty of people who could find new applications and ways of using this," says TakkTile co-creator Yaroslav Tenzer.

Providing Robotic Carers and Smart Systems for the Elderly
CORDIS News (04/18/13)

European researchers continue to improve and fine-tune an automated carer system for the elderly to make it useful, acceptable, and fun to use. Launched in December 2009, the Mobiserv project consists of a robot companion, wearable smart clothes, and a smart home environment. The robot is designed to serve as a social carer that can remind older adults to eat, drink, take medicines, exercise, and do certain activities, such as call or visit someone if they have not communicated with anyone else for a while. The smart clothes can monitor vital signs or sleeping patterns, and detect falls. The smart home environment uses smart sensors, optical recognition units, and home automation elements to detect eating and drinking patterns, activity patterns, and dangerous situations, among other things. The researchers have scheduled extensive user-evaluation studies for April through June, and they plan to present and demonstrate the project and its companion robot at several events in Europe from June to August.

Abstract News © Copyright 2013 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