Association for Computing Machinery
Welcome to the May 4, 2011 edition of ACM TechNews, providing timely information for IT professionals three times a week.

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.


Wall Street Traders Mine Tweets to Gain a Trading Edge
USA Today (05/03/11) Adam Shell

Wall Street traders are using social networking sites and computer algorithms to decipher the mood of the general public to predict market fluctuations. The use of linguistic analysis of online posts could prompt a computer program designed to interpret the data to place a trade with no human intervention. Researchers know that emotions play a significant role in markets and analyzing millions of tweets is similar to a "large-scale emotional thermometer for society as a whole," says Indiana University professor Johan Bollen, who recently completed a study that links Twitter mood measurements to stock market performance. Bollen used two mood-tracking tools to analyze the text of 9.6 million Twitter feeds over a nine-month period. One tool measured whether tweets were positive or negative, while the other tool categorized tweets into one of six moods. The measuring tools resulted in 87 percent accuracy in predicting Dow stock prices three to four days later, according to Bollen. Meanwhile, Pace University's Arthur O'Connor recently conducted a study showing a positive correlation between stock price performance and the social media popularity of brands such as Starbucks, Coca-Cola, and Nike. Market traders also are using computer algorithms to speed-read and analyze news to search for investing clues.

Robots Learn to Share, Validating Hamilton's Rule
Ecole Polytechnique Federale de Lausanne (05/04/11) Michael David Mitchell

Researchers at Ecole Polytechnique Federale de Lausanne (EPFL) and the University of Lausanne are developing robots with altruistic traits based on Hamilton's rule, which states that an organism is more likely to express altruism with others who possess many genetic similarities. "We have shown that Hamilton's kin selection theory always accurately predicts the relationship between the evolution of altruism and the relatedness of individuals in a species," says EPFL's Markus Waibel. The robots start with a random series of coded information to complete a task. Only those robots that successfully complete the task are able to pass on their code to the next generation. After hundreds of generations, the robots become more efficient and learn to work in groups. The researchers created groups of relatedness that are the equivalent of clones, siblings, cousins, and non-relatives. The groups that shared traits correlating with Hamilton's rule completed the tasks better and passed their genes on to the next generation. "We are using this altruism algorithm to improve the control system of our flying robots and we see that it allows them to effectively collaborate and fly in swarm formation more successfully," says EPFL professor Dario Floreano.

Europe Leads in Pushing for Privacy of User Data
New York Times (05/03/11) James Kanter

Europe is preparing to the lead the way in establishing new regulations for user data privacy. European Union (EU) justice commissioner Viviane Reding plans to propose extending EU-wide rules about privacy breaches to online banking, video games, shopping, and social media. The rules mandate that Internet service providers and phone companies notify customers of any data breach without undue delay. "Any company operating in the EU market or any online product that is targeted at EU consumers should comply with EU rules," Reding says. Georgetown University professor Abraham L. Newman says that Europe's focus on privacy could give companies an opportunity to restructure the way they manage policies globally, using European standards in their approach. Or, he says, they could devise policies to guarantee that data collected in Europe was suitably quarantined to adhere to the regulations, while restricting changes elsewhere. Reding says that differences between Europe and the United States should not draw attention away from areas in which they agree, such as the Obama administration and Congress' efforts to pass a privacy bill of rights to curb the corporate collection or sharing of personal data without an Internet user's permission.

Simulation Model to Improve Safety and Efficiency of Port Traffic
Delft University of Technology (05/03/11)

Researchers at the Delft University of Technology and Jiaotong University are participating in a joint project to develop a traffic simulation model for shipping in congested port areas to improve capacity and safety. "The issue of safety is growing in importance because of the increasing threat of collisions between ships, caused in part by the increased congestion and its effects on both the environment and the local area," says Delft professor Han Ligteringen. The researchers will base their studies on previous simulations of other forms of traffic, such as pedestrians and automobiles. "Although their method of propulsion is completely different, just like pedestrians ships have a great deal of flexibility in terms of the route they can select and how they interact with other ships," says Delft's Winnie Daamen. The researchers will use game theory to predict the behavior of individual ships and their interactions under varying conditions, such as wind, waves, current, and visibility.

Speeding Swarms of Sensor Robots
MIT News (05/03/11) Larry Hardesty

Massachusetts Institute of Technology (MIT) researchers have developed an algorithm that enables sensor-equipped robots to focus on parts of the environment that change most frequently while keeping track of those aspects that change more slowly. The algorithm assumes that researchers already have a model for how the conditions change over time, and uses those measurements to adjust the robots' velocities as they trace their routes. The researchers also developed an algorithm for underwater sensors that University of Southern California researchers are using to study algae blooms. Each sensor is shaped like an airplane and uses an inflatable bladder to move vertically in the ocean. The rate at which the bladder fills affects the sensor's trajectory through the water. The MIT team developed an interface that enables ocean researchers to select specific areas of interest to study. The researchers are currently modifying the algorithm so that it can change its own computations in light of new evidence.

Google's Research Chief: The Power of Big Data
New Scientist (05/03/11) Peter Aldhous

Google research director Peter Norvig and his team are following the concept of applying simple statistical algorithms to enough data to address some of the most difficult problems in machine learning. One example is automated language translation. Rather than designing software to mimic a human translator's understanding of linguistic rules, Norvig's team has accumulated already translated texts and applied statistical methods to train Google's system to learn translations of unfamiliar words and their contextual use. Another challenge that Norvig believes will be met by the big data, simple algorithms approach is speech recognition. Gathering sufficient data about the spoken word to make speech recognition a reality is one of the goals underlying the rollout of the Google Voice service. "One of the reasons we had this phone service is that we wanted to capture lots of interactions; hear different accents and different voices saying different things," Norvig says. A third research focus of Norvig's is visual search. By drawing more relationships between image data through statistical application, people will be able to employ more powerful ways of connecting real-world queries to Google's search capabilities using mobile devices.
View Full Article - May Require Free Registration | Return to Headlines

Ranking Research
EurekAlert (05/03/11)

Researchers at Chulalongkorn University have created CiteRank, an algorithm that uses social bookmarking tools to evaluate the relevance of research papers. CiteRank combines similarity ranking, which measures the match between a query and a research paper index, with static or query-independent ranking, which measures the quality of a research paper. CiteRank uses a group of factors, including number of groups citing the posted paper, year of publication, research paper post date, and priority of a research paper to determine a static ranking score, and applies the query-independent measure. When testing the ranking algorithm, the team found that the weighted combination of 80 percent similarity ranking and 20 percent static ranking was most effective. The researchers learned that literature professors preferred to read more recent papers or just-posted papers, but also rated highly classic papers that emerged in the results if they were posted across different user groups or communities. "Improving indexing not only enhances the performance of academic paper searches, but also all document searches in general," says Chulalongkorn University's Siripun Sanguansintukul.

Talking to the Wall
Technology Review (05/03/11) Kate Greene

Microsoft and University of Washington researchers have developed a method to convert the ambient electromagnetic radiation generated by power lines and home electronics into a computer interface that can be applied to any wall, transforming it into a touch-sensitive surface. The researchers found that when a person touches a wall with electrical wiring behind it, the person becomes an antenna that tunes the background radiation, creating a distinct electrical signal. This signal can be collected and interpreted by a device close to the person's body. "Now we can turn any arbitrary wall surface into a touch-input surface," says Washington professor Shwetak Patel. The technology could be used to create new interfaces for games as well as enable users to control electronic devices, security systems, light switches, and thermostats from anywhere in the house. The researchers say the next step is to facilitate real-time data analysis. Massachusetts Institute of Technology professor Pattie Maes says the system will need to be more user friendly. "Many interfaces require some visual, tangible, or auditory feedback so the user knows where to touch," she notes.

CMU's Classroom Salon Uses Social Networking to Tap Collective Intelligence of Online Study Groups
Carnegie Mellon University (05/03/11) Shilo Raube

Carnegie Mellon University (CMU) researchers have developed Classroom Salon (CLS), a social networking application that creates online learning communities for students. CLS enables students to share their ideas about reading materials and critiques of each others' writings. "With Classroom Salon, we've tried to capture the sense of connectedness that makes social media sites so appealing, but within a framework that that allows groups to explore texts deeply," says CMU professor Ananda Gunawardena. "So it's not just social networking for the sake of socializing but enhancing the student experience as readers and writers." CLS' interactive tools enable users to highlight hot spots that lead to discussion, cluster similar content, and identify influential comments. The researchers say that CLS could be especially beneficial for at-risk students. The system will be tested at the University of Baltimore, where it will be combined with materials developed for CMU's Open Learning Initiative and traditional face-to-face instruction to create a sustainable social learning model.

EU Pushes Button for Next Internet Age
Agence France-Presse (05/03/11)

The European Union (EU) recently launched a research and development effort to create a future Internet that can handle an exponential increase in data. The EU provided an initial investment of 90 million euros, which will cover the first phase of a five-year public-private partnership. The project is designed to boost innovation and develop Internet solutions "capable of managing the exponential increase in [mobile] online data," says EU commissioner Neelie Kroes. Initial projects that are currently underway include Internet tools for electricity management at the community level and environmental data for weather forecasting. The EU has pledged a 300 million-euro investment by 2016. "We are heading for a world that is much smarter," Kroes says. "Where information generated by people and our environments can be used productively in real time."

MIT Project Makes Smarter Mobile Wi-Fi
CNet (04/29/11) Rafe Needleman

Massachusetts Institute of Technology (MIT) researchers are developing technology that takes current and predicted future device movement into account when connecting to hotspots to increase overall wireless performance on mobile devices. The MIT researchers, led by professor Hari Balakrishnan, developed software that uses a device's sensors that normally go unused. The device can tell where it is going, and link with a hotspot that will be able to connect with for a longer period of time, instead of just using one that is available at the moment of initial contact. The researchers say the software should reduce the number of handoffs between hotspots, increasing the connection quality for mobile devices. Testing shows an improvement in overall network performance from 40 percent to 60 percent for moving devices, and a reduction in hotspot handoffs of about 40 percent, according to Balakrishnan. He says the technology would be more effective if it was incorporated into network protocols and programmed into Wi-Fi routers.

UA Researchers Track Terrorists on the 'Dark Web' (AZ) (04/29/11)

More needs to be done to cut off or restrict the cyberpipelines that terrorists use to recruit people to their cause, says University of Arizona professor Hsinchun Chen, founder of the Dark Web project. The project is essentially a cyberbattle station based at the university that computer scientists use to gather information on terrorists. Chen previously focused on static Web sites, but terrorists are now using social media to recruit people and plot attacks. The researchers have developed advanced cybertools for sifting through millions of daily posts and chats of suspected jihadists on social media, and researchers are able to parse words, videos, and data to create maps and uncover potential terrorist ties. "They try to entice and educate the audience, so it has ranged from text, video of violent content, IEDs--improvised explosive devices, how they put in the right place," Chen says. "They are more volatile, they are more creative, they are more mature, and they are reaching a bigger audience than before." Chen's researchers share the data with military and intelligence analysts.

Evolutionary Biologists Map Enormous Plant 'Tree of Life' (04/28/11) Alice Vincent

Researchers have mapped the last 125 million years of plant evolution in the largest example of a tree of life, or phylogeny, to date. Computer scientists worked with evolutionary biologists found that plants are likely to take a trial-and-error approach to creating a new species. Researchers previously assumed that a new plant species was created when a new physical trait or mechanism develops and rapidly produces, but the findings suggest there is a time lag or a phase of maximum development before mass reproduction. Computing such vast amounts of information was a challenge, says project leader Alexandros Stamatakis. "It was not previously possible to reconstruct such large trees or phylogenies based on statistical models of sequence evolution or mutation," he says. The team had to create software to both reconstruct the evolutionary trees and to extract useful information from them.

Abstract News © Copyright 2011 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