Association for Computing Machinery
Welcome to the April 8, 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


Coding for a More Open Cuba
The Wall Street Journal (04/07/15) Kjel Vyas

Technology experts will gather at Facebook's Menlo Park, CA, headquarters later this month to participate in the Code for Cuba hackathon organized by the Roots of Hope nonprofit. Participants will focus on boosting access to information as Cuba starts considering more open telecommunications. They will brainstorm ways for Cubans to circumvent Internet restrictions and connect into the global stream of information, and then teams will vie for prize money to develop the best solutions during the two-day event. Roots of Hope's Natalia Martinez says the purpose of the hackathon is to develop programs for immediate application in Cuba while remaining within the bounds of Cuban law, which bars satellite linkage and limits private imports of products such as wireless routers. Meanwhile, Cuba has set the goal of having 60 percent of its 11 million citizens using the Internet, with connections reaching more than 50 percent of the country's homes, by 2020. Only about 25 percent of Cubans currently access the censored Internet, and home connections are outlawed except for foreigners. "It's going to be baby steps and it's going to feel incremental and slow," notes IDT's Claude Pupkin. "But in five to 10 years, there's going to be a lot more freedom, more cell-phone penetration, and Cubans will be more connected to the world."
View Full Article - May Require Paid Subscription | Return to Headlines | Share Facebook  LinkedIn  Twitter 


If Algorithms Know All, How Much Should Humans Help?
The New York Times (04/06/15) Steve Lohr

As the use of big data analytics to make major decisions has spread throughout the economy, so has unease and anxiety about the results these systems produce, leading to a desire for human oversight. Big data concerns have led to the birth of a new academic field called algorithmic accountability. Academics involved in this field want to find ways of ensuring algorithms yield the desired results without unjustly disadvantaging the people affected by those decisions. One example is the increasing use of analytics in the consumer lending space. Algorithmic systems harvest a wide range of data sources to build a profile of a potential borrower that is better than a credit score at predicting if they will make their payments. However, such algorithms, depending on how they are tuned, also could exclude people who might make good borrowers but have the wrong data profile. Some experts say having a human being checking an algorithm's outputs could be a solution. However, others believe the real solution is to be more exacting in crafting and maintaining algorithms. "The goal is not necessarily to have a human look at the outcome afterwards, but to improve the quality of the classification for the individual," says Gary King, director of Harvard University's Institute for Quantitative Social Science.
View Full Article - May Require Free Registration | Return to Headlines | Share Facebook  LinkedIn  Twitter 


Soon, Students May Learn to Code Instead of Taking French Class
Bloomberg (04/02/15) John Lauerman

At least four U.S. states have either passed or considered measures that would enable high school students to satisfy their foreign language requirement by taking a computer language instead. In 2013, the Texas legislature unanimously passed a bill letting students substitute computer science credits for foreign language requirements. Georgia's state Board of Education considered a similar measure proposed by Gov. Nathan Deal earlier this month. Many students are more passionate about learning code than conjugating verbs, notes Washington state legislator Chris Reykdal. Proponents say the approach would be a good way to support computer science, and ultimately help students get jobs and help businesses compete globally. However, educators say foreign language learning is critical to being able to engage with the rest of the world, and many technologists also agree swapping computer language study for foreign language study is more of an apples-and-oranges substitution. Reykdal put his bill on hold after meeting with foreign language teachers and now is seeking a way to promote both computer science and foreign language courses.


NASA, IBM Team Up on Global Hackathon to Solve Earth's Problems
The Washington Post (04/07/15) Amrita Jayakumar

The U.S. National Aeronautics and Space Administration (NASA) this weekend will hold its fourth annual space apps hackathon challenging developers, scientists, artists, and engineers to design apps addressing the space agency's unique needs. The three-day event will take place at more than 135 locations around the world, including Washington, DC, and New York City. NASA has issued specific challenges in several fields such as robotics, earth science, space exploration, and human health. The agency is looking for solutions such as apps that can control small drones in space, print space food, and improve the ability to grow crops in space. Winners will be selected in five categories, including "best use of data" and "most inspiring," and they will have a chance of NASA adopting their apps for use in the space agency's projects. This year, NASA is partnering with IBM for the event. IBM will provide participants with access to several resources, including its Bluemix platform and its Watson supercomputer. IBM employees also will be available to provide technical support. In addition to offering participants access to its services, IBM will award $120,000 worth of Bluemix services to the 30 winners who design the most innovative apps.
View Full Article - May Require Free Registration | Return to Headlines | Share Facebook  LinkedIn  Twitter 


Teaching a Computer Not to Forget
The Atlantic (04/06/15) Adrienne LaFrance

Humans gradually forget skills as they learn something new, but computers forget what they know as soon as they learn a new skill. Computer scientists call this problem catastrophic forgetting, and they say researchers will need to resolve this issue if they hope to build artificially intelligent computers and robots that can solve diverse problems. Catastrophic forgetting happens because computer brains often rewire themselves every time they learn, making it harder for machines to retain old lessons and learn tasks that require a sequence of steps. University of Wyoming professor Jeff Clune and colleagues are trying to make an artificial brain act more like a human one. Biological brains exhibit a high degree of modularity, meaning they contain clusters of neurons with high degrees of connectivity within clusters, and the team's approach was to show a computer that modularity is preferable. The researchers report modularity appears to help computers, like humans, compartmentalize what they know. They believe modularity may be a key to ending catastrophic forgetting. "Biological brains exhibit a high degree of modularity, meaning they contain clusters of neurons with high degrees of connectivity within clusters, but low degrees of connectivity between clusters," the researchers note.


Planes Without Pilots
The New York Times (04/06/15) John Markoff

Modern airplanes are highly automated; a recent survey of commercial airline pilots found most spend significantly less than 10 minutes of any given flight in direct control of their planes. However, some researchers say there is room to automate planes even further, especially in the wake of the Germanwings plane believed to have been deliberately crashed by its co-pilot. The relatively old F-16 fighter plane already has a ground-avoidance feature that could have prevented such a crash, automatically taking control of the plane if it approaches the ground too fast or at the wrong angle. Meanwhile, the U.S. Defense Advanced Research Projects Agency (DARPA) is conducting a competition to design a robot co-pilot that could speak and listen to the pilot and even manipulate the flight controls. DARPA plans for the robot to employ a wide variety of advanced technology, ranging from voice recognition and speech synthesis to computer vision. In addition, the U.S. National Aeronautics and Space Administration (NASA) wants to automate the U.S. air traffic control system. NASA estimates more automation could increase the density of air traffic by 20 percent while requiring fewer human controls than today. However, some researchers are leery of taking humans completely out of the cockpit, especially on commercial flights carrying people.
View Full Article - May Require Free Registration | Return to Headlines | Share Facebook  LinkedIn  Twitter 


Diagnosis by Keyboard
MIT News (04/01/15) Anne Trafton

Massachusetts Institute of Technology (MIT) researchers have developed an algorithm that can analyze computer users' keystrokes in order to identify a range of information about the state of their motor functions. The study is based on the idea there might be hidden information about the way people type, according to MIT's Ian Butterworth. He says the research represents the first step toward using keystroke patterns to diagnose conditions that impair motor function. Preliminary results suggest the algorithm can distinguish people who have Parkinson's disease from those who do not. The algorithm captures timing information from computer keystrokes, enabling the researchers to detect patterns that distinguish typing that occurs when motor skills are impaired. The algorithm focuses on the key hold time, which is a measure of how long a key is pressed before being released. The system collects data by using plug-in software that is incorporated into a Web browser to capture keystrokes. The researchers also are developing a smartphone app that could be used to gather the same kind of data from mobile devices. "We thought this was a unique opportunity to have a window into the brain using your normal interactions with an electronic device," says MIT researcher Alvaro Sanchez-Ferro.


Why Women Won't Code Is Topic of New Documentary
USA Today (03/30/15) Marco della Cava

Film director Robin Hauser Reynolds was inspired to create her new documentary about the gender gap in computer science, "CODE: Debugging the Gender Gap," by the experience of her college-age daughter. Reynolds says her daughter, who had expressed a keen interest in computer science, was dismayed to find she was one of only two women in her computer science classes and nearly dropped computer science because of it. Reynolds says the problem of women not pursuing computer science is not just an issue of equality, but an economic one as well. "By 2020, there will be 1.4 million technology jobs, but only 400,000 people will have the skills necessary to get those jobs," she says. "This isn't just about gender. It's about a potential economic crisis." The documentary features interviews with women in tech who talk about facing explicit and implicit gender bias on a daily basis, outsized expectations that make them feel like they have to constantly prove their worth, and a lack of female role models and mentors. However, it also highlights how rewarding tech careers can be for women, featuring interviews with female workers who speak about the satisfaction they derive from their jobs at companies such as Pixar and Pinterest.


Is Your Robot a Little Cheeky? Google May Build It That Way
Computerworld (04/02/15) Sharon Gaudin

Google is developing robots that have individual personalities. "A robot may access a user device to determine or identify information about a user, and the robot may be configured to tailor a personality for interaction with the user based on the identified information," according to a Google patent application, first filed in April 2012 and accepted last week. These personable robots would be designed to identify different people, using speech and facial recognition, and then configure their personality to suit the user. "In some examples, a robot's personality or personalization can be transferred from one robot to another robot, or information stored on one robot can be shared with another robot over the cloud," the application says. The goal of developing robots with personalities is to have them better interact with people in elder care, child care, or hospital settings. "I believe that when done right, a robot with a personality will make humans more comfortable with them," says Moor Insights & Strategy analyst Patrick Moorhead. ZK Research analyst Zeus Kerravala says Google's research into changeable personalities is interesting.
"You want your robot to be customizable," Kerravala says. "...This is moving robots more toward personal usage."


Connecting Vehicles
Oak Ridge National Laboratory (04/02/15) Morgan McCorkle

Oak Ridge National Laboratory researchers are developing a computational framework for connected-vehicle technologies that facilitates vehicle-to-vehicle communication, as well as communication between vehicles and traffic control systems. The system will help vehicles exchange information, such as location, speed, and destination, to create individualized instructions for drivers. "By telling drivers the optimal speed, the best lane to drive in, or the best route to take, we can eliminate stop-and-go driving and improve safety," says Oak Ridge Urban Dynamics Institute deputy director Andreas Malikopoulos. The system relies on decentralized control algorithms that govern how vehicles communicate locally in order to optimize traffic flow across a city. The decentralized control algorithms are effective because all of the vehicles in a city cannot communicate information to a central control center due to the massive amount of data that would be involved. The project's first phase will validate the framework through simulation, and the second phase will connect the team's communication framework with a transportation analysis simulation system that uses data analytics to simulate traffic conditions and predict congestion. In addition, as part of the second phase, researchers will begin exploring questions related to cybersecurity and possible incentives for drivers to follow connected vehicle instructions.


NYU and DARPA Dive Deeper Into the Web
Baseline (04/02/15) Samuel Greengard

New York University Polytechnic School of Engineering (NYU-Poly) has received a $3.6-million grant from the U.S. Defense Advanced Research Projects Agency to develop more sophisticated methods of locating and exploring hard-to-find information on both the public Internet and the dark Web. Called Memex, the program has three primary focuses: domain-specific indexing of public Web content, domain-specific search capabilities, and applications specifically tailored for the U.S. Department of Defense that could aid in fighting crime or strategic information-sharing within the military. Memex will not exist as a standard search engine, but instead as a component of Java-based open source software that will run on all major operating systems. NYU-Poly professor Juliana Freire says Memex will use deep-learning techniques to classify and categorize topics. "As we crawl the Web, we discover new beta sources for information, and we update the crawler," Freire says. One initial area of focus is human trafficking, with an effort to glean data from escort service websites and services that offer to deliver brides. In the future, Memex could be used by law enforcement to gather information in areas such as antiterrorism, child pornography, and drug enforcement.


'Free' Apps May Not Be So Free After All: They Take a Big Toll on Your Phone
USC News (04/01/15) Robert Perkins

Applications with advertisements use an average of 16 percent more energy than apps without ads, lowering the battery life of a smartphone from 2.5 hours to 2.1 hours on average, according to researchers from the University of Southern California (USC), Rochester Institute of Technology, and Queens University in Canada. They also report apps with ads use 22 percent more memory and take up an average of 48 percent more central processing unit time. Moreover, they use 79 percent more network data, costing an estimated 1.7 cents every time they are used, based on the average cost per megabyte charged by AT&T. The researchers compared 21 top apps from the past year, culled from a list of 10,750 that had been in the top 400 of each of Google Play's 30 categories from January to August, and then measured their effect on phones using analysis tools loaded onto a Samsung Galaxy SII. The researchers hope developers will take note of the findings because "apps are the future of software," says USC professor William Halfond.


Five Years Ago the iPad Changed Clicks to Touches--but Another Tablet Revolution Is Coming
The Conversation (04/03/15) Jason Alexander

Next-generation displays are expected to evolve from the flat, two-dimensional form factor popularized by the iPad into three-dimensional (3D) tablets equipped with self-actuated, physically reconfigurable surfaces, writes Lancaster University lecturer Jason Alexander. He says these deformable flat screens will be capable of changing the shape of their display surface to better represent on-screen content and offer more ways to convey information tactilely instead of only visually. Computer interaction will undergo a fundamental transformation via the advent of dynamic physical geometry, as physically protruding pixels will enable developers to augment applications such as architecture, design, terrain modeling, and photography by rendering computer-generated 3D scenes in three dimensions. Alexander says this will create opportunities for unique applications in team collaboration, tangible entertainment, and ways of improving computer access for the disabled. He notes Lancaster University researchers have developed an interactive bar chart to represent common data visualization tasks such as displaying and filtering data, organizing it into rows and columns, navigating between large datasets, and producing annotations. They determined the physical nature of dynamic bars encouraged users to directly manipulate data points for annotation and comparison-style tasks, and that traditional touch-based controls are conducive to navigation and organization. Non-technical researchers can experiment with reconfigurable displays via ShapeClip, which adds a z-axis to a computer screen's x- and y-axes.


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