Association for Computing Machinery
Welcome to the January 21, 2015 edition of ACM TechNews, providing timely information for IT professionals three times a week.

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

HEADLINES AT A GLANCE


Program Teaches Low-Income Kids to Code
USA Today (01/20/15) Jessica Guynn

Civil rights activist Van Jones has created Yes We Code, an initiative that aims to teach 100,000 low-income kids programming skills. Yes We Code is helping dozens of organizations around the U.S. that are trying to address high-tech's racial and gender gap, from Black Girls Code to Hack the Hood. It connects those groups with the resources they need, according to Jones. Last year, leading technology companies released data showing African Americans and Hispanics make up just 5 percent of the companies' workforces, compared with 14 percent nationally. Jones says the lack of minority participation means Silicon Valley may be missing out on the next big idea or company because it employs too few women and people of color. He says Yes We Code wants to get communities to redirect young people's talents, and to help the technology industry access that talent from new places, such as community colleges, coding boot camps, tribal colleges, and historically black colleges and universities. "Yes We Code aspires to become the United Negro College Fund equivalent for coding education," Jones says. "Yes We Code exists to find and fund the next Mark Zuckerberg and Sheryl Sandberg in communities you would never expect to find them."


Artificial Intelligence Experts to Explore Turing Test Triathlon
IEEE Spectrum (01/19/15) Lee Gomes

Headlines about the Turing Test being defeated last year, when a chatbot known as Eugene Goostman was able to convince several judges at an artificial intelligence (AI) contest that it was a person, have galvanized many AI researchers to devise better tests for determining whether a computer might be considered intelligent. The scientists will gather this Sunday for a workshop at the annual convention of the Association for the Advancement of Artificial Intelligence in Austin to develop better tests of machine intelligence. One of the leading candidates for such a test is Winograd schema, which require a computer to use commonsense and real-world knowledge to understand the meaning of an intentionally ambiguous sentence. Gary F. Marcus, a New York University research psychologist who is co-chairing the workshop, says the goal is to create what he calls a "Turing Triathlon," a battery of tests that will analyze a computer or program on several levels, not just its proficiency with language. The tests, which Marcus expects will be held once or twice a year, also will be revisited on a regular basis to make sure they are working as intended and that programmers are not attempting to game them the way the Goostman chatbot did.


The New Digital Arms Race: NSA Preps America for Future Battle
Der Spiegel (Germany) (01/17/15) Jacob Appelbaum; Aaron Gibson; Claudio Guarnieri; et al.

Der Spiegel magazine draws from documents provided by former U.S. National Security Agency (NSA) contractor Edward Snowden to outline NSA's various efforts to arm itself for and carry out covert cyberwarfare. NSA in 2013 estimated it needed about $1 billion to maintain and continue developing its cyberwarfare capabilities. The article describes several programs, many of them tied to NSA's Tailored Access Operations (TAO) department and its Remote Operations Center, codenamed S321. Some of the documents viewed by Der Spiegel outline how NSA's priority with its cyberespionage operations is not just to monitor, but to have the ability to control or destroy other nations' critical systems. NSA reportedly pursues this goal with a vast armory of tailor-made malware designed to enable it to snoop on and seize control of everything from a cell phone to a server farm. The article also details S321's counter-intelligence efforts to not just identify foreign cyberespionage programs, but to exploit them to siphon off information about other targets. Another project involves subverting hostile botnets and servers to carry out attacks and hide stolen data, granting NSA plausible deniability.


Microsoft Researchers Use Light Beams to Charge Smartphones
Computerworld Australia (01/19/15) Rohan Pearce

Microsoft researchers in Beijing have developed AutoCharge, a prototype automatic wireless charging system for smartphones. The charger can be mounted on a ceiling and automatically locate a smartphone lying on a table, and then charge it using a directed beam of light. The light charger has two modes: in the detection mode, it uses a camera and image recognition software to detect objects with the size and shape of a smartphone lying on a table. The charger will rotate until it detects an object that looks like a smartphone. The device then enters charging mode and turns on its light. The system relies on the smartphone exchanging messages with the charger via an on-phone light-emitting diode in order to avoid attempting to charge a phone that already has a full battery. The researchers found using a light beam to charge a smartphone could be as quick as many wired chargers, depending on the size of the photovoltaic panel. The charger also is capable of identifying when an object moves between it and the phone, and automatically deactivating the light beam within 50 milliseconds.


Facebook Offers Artificial Intelligence Tech to Open Source Group
The New York Times (01/16/15) Quentin Hardy

Facebook recently announced plans to donate several powerful computing tools to Torch, an open source software project focused on deep-learning techniques. Among the tools Facebook is donating are neural networking tools the company says are capable of speeding pattern recognition by up to 23.5 times. Facebook also is donating new means for training numerous computer processors at the same time to better catalog words when analyzing language and using speech recognition software. Torch was co-founded by Soumith Chintala, a research engineer studying artificial intelligence at Facebook. Chintala says Torch's assorted tools are useful for a variety of projects ranging from neural networks and artificial intelligence to computer vision and text-recognition systems. Chintala says his group has used Torch tools to "read" the "Lord of the Rings" novels, and afterwards they were able to successfully answer basic questions about the books. "It's very early, but it shows incredible promise," says Chintala, who adds some potential applications include being able to read and automatically apply hashtags to Facebook posts.
View Full Article - May Require Free Registration | Return to Headlines | Share Facebook  LinkedIn  Twitter 


A Team at the UPC Designs Technology That Automates the Production of Football Match Highlights
Polytechnic University of Catalonia (Spain) (01/16/15)

A system developed by a team from the Polytechnic University of Catalonia in Spain may help journalists and program editors with highlights of soccer matches. The technology is designed to automatically analyze video of a match and select key moments for highlights. The system "divides the video content into different frames, which are rated according to the relevance of the low-level information that is extracted from the images and audio using mathematical algorithms," notes project principal investigator Francesc Tarres. The technology assigns a numerical mark to aspects of information such as ambient noise levels, including the roar from the crowd and referee blowing the whistle, camera movements, predominant colors in the images, and people or faces that accumulate in a particular segment of the image. The numerical mark goes toward a final mark for each of the video fragments detected, which is implemented as a sign that something interesting is taking place on the field. The system selects all of these moments to create a summary of the match's highlights. The technology selected 70 percent of the goals in a test involving five matches.


Software That Knows the Risks
MIT News (01/16/15) Larry Hardesty

Massachusetts Institute of Technology (MIT) researchers have developed artificial intelligence software that enables a planner to specify constraints and reliability thresholds. Then, on the basis of probabilistic models, the system determines whether a solution exists. However, if a solution does not exist, the software suggests ways in which the planner might relax the problem constraints. The researchers note one unique aspect of the software is that it assesses risk. "We say, 'What’s your budget of risk for this entire mission? Let’s divide that up and use it as a resource,'" says MIT researcher Cheng Fang. First, the algorithm represents a problem as a graph, then it begins adding edges that represent the constraints imposed by the planner. If the problem is soluble, the weights of the edges representing constraints will be greater than the weights representing the costs of transitions between events. Existing algorithms can quickly home in on loops in the graph where the weights are imbalanced, and the MIT system calculates the lowest-cost way of rebalancing the loop, which it presents to the planner as a modification of the problem's initial constraints. "If you expose what you need to consider in the planning stage, then you have a much higher success rate in carrying out the plan," says Nuance Communications research scientist Jiaying Shen.


NASA Langley Research Center Envisions Virtual Helpers
Newport News Daily Press (VA) (01/17/15) Tara Bozick

The U.S. National Aeronautics and Space Administration (NASA) Langley Research Center's Big Data Analytics team wants to develop computer assistants or "virtual helpers" that can do the more mundane work for scientists, enabling human researchers to focus on bigger projects or ideas. NASA Langley has been testing about a dozen virtual helper applications in pilot projects over the last seven months, including robots, high-tech watches, and smartphone apps. Machines capable of deep analytics would not replace scientists but would empower them to focus on higher-level research, says NASA Langley's Manjula Ambur. After NASA Langley learns how to apply big data analytics in various case studies, the Big Data Analytics team will develop a suite of internal tools that can be used or customized for various projects. Ambur says NASA's big data analytics research could be used in many fields, including social media, finance, and retail. She says the creation of such intelligent machines or software also could empower nonscientists or nonexperts to develop their own innovations.


Keeping Citizens Safe While Respecting Their Right to Privacy
Homeland Security News Wire (01/16/15)

European Union-funded researchers working as part of the MOSAIC project have developed a technological surveillance solution designed to keep law-abiding citizens safe while at the same time respecting their privacy. The project comes at a time when there is a growing public debate over the extent to which surveillance is justified, and how the right balance between civil protection and personal privacy can be achieved. However, the rapid increase in information gathered by surveillance cameras has resulted in spiraling costs for storage filtering and data checking, and also has led to concerns that innocent citizens are routinely being tracked. The MOSAIC project aimed to streamline surveillance through the development of several decision-support technologies based on video analytics and using data methods that search tags and fuse data from both multimedia sources and information databases. The distributed intelligence within the platform enables decision support for automated detection, recognition, geo-location, and mapping, including intelligent decision support at various levels to enhance situation awareness, surveillance targeting, and camera handover, according to the researchers. The final integration of all the MOSAIC components achieved an overall satisfaction level of 79.88 percent in final trials with end users.


Making the Internet Accessible to Everyone
CORDIS News (01/15/15)

The third pilot of the Internet accessibility CLOUD4ALL project will kick off in Athens, Berlin, and Madrid in the coming months. The European Union-funded project seeks to use cloud technologies to achieve simple and instant Web accessibility for everyone, using any device anywhere. The focus is on developing the technical components needed to enable the auto-configuration accessibility features of different devices, platforms, and applications. Once a user has selected the settings needed to improve accessibility, such as display in high contrast or use of a screen reader, those settings would apply to any device used, which would be Cloud4All compatible. The pilot will test the final auto-configuration from preferences process, in which users switch between devices that will adapt their accessibility features to their needs. In 2013, for the first pilot, participating researchers tested whether the concepts were viable. Last year, the second pilot tested prototypes, including the preference management tools, on users.


How Does a Machine Smell? Better Than It Did
University of Manchester (01/16/15) Sam Wood

Researchers from the University of Manchester and the University of Bari in Italy have developed a chemical sensor platform that can smell more accurately than humans. The team incorporated odorant binding proteins, found in the mucus of the nose, into field-effect transistors. The proteins work with olfactory receptors and help create a person's perception of smell. The researchers report they were able to measure the unique changes in current as the proteins reacted to odor and record them, which they say is equivalent to the machine smelling the odor. The project marks the first time machines were able to differentiate smells that are very similar, such as chiral molecules. Moreover, the scientists have found a way to manufacture the proteins in quantities that would enable them to be used in biosensors. "Now we have done this it will allow much better sensors to be developed and these could have many uses in industry," says Manchester professor Krishna Persaud. "We shall be able to create biosensors which are accurate enough to be able to tell when food has gone off, or even smell how much pollution is in the atmosphere."


Rice-Sized Laser, Powered One Electron at a Time, Bodes Well for Quantum Computing
Princeton University (01/15/15) Catherine Zandonella

Princeton University researchers have developed a rice grain-sized laser powered by single electrons tunneling through artificial atoms known as quantum dots, a breakthrough they say could lead to the creation of quantum computers built out of semiconductor materials. Quantum dots can communicate through the entanglement of photons, so the researchers designed dots that emit photons when single electrons leap from a higher energy level to a lower energy level to cross the double dot. The researchers fabricated the double quantum dots from extremely thin nanowires made of a semiconductor material called indium arsenide. To construct the new device, they placed the two double dots about six millimeters apart in a cavity made of niobium, which requires a temperature near absolute zero to superconduct. "This is the first time that the team at Princeton has demonstrated that there is a connection between two double quantum dots separated by nearly a centimeter, a substantial distance," says University of Maryland-National Institute of Standards and Technology professor Jacob Taylor, who collaborated on the research. The device enables energy levels within the dots to be fine-tuned to generate light at other frequencies. "The double quantum dot allows them full control over the motion of even a single electron, and in return they show how the coherent microwave field is created and amplified," says Princeton professor Claire Gmachl.


Improved Interface for a Quantum Internet
University of Innsbruck (01/15/15)

University of Innsbruck researchers have improved an interface between quantum processors and optical fiber-based communication channels, making use of so-called superradiant states. "In our interface, we position two ions between two highly reflective mirrors, which form an optical resonator," says University of Innsbruck researcher Tracy Northup. "We entangle the ions with one another and couple both of them to the resonator." She says this process enables the collective interaction between the particles and the resonator to be tuned in order to enhance the creation of single photons. The researchers demonstrated the interface is suitable for quantum information processing by encoding a quantum state in the entangled particles, and then transferring this state onto a single photon. "Thanks to superradiance, the process of information transfer from the particle to the photon essentially becomes more robust," says University of Innsbruck researcher Bernardo Casabone. The researchers also were able to create subradiant states, in which the emission of a photon is suppressed rather than enhanced. "These states are also interesting because the stored information becomes invisible to the resonator, and in that sense, it's protected," Northup says.


Abstract News © Copyright 2015 INFORMATION, INC.
Powered by Information, Inc.


To submit feedback about ACM TechNews, contact: technews@hq.acm.org
Current ACM Members: Unsubscribe/Change your email subscription by logging in at myACM.
Non-Members: Unsubscribe