Welcome to the August 21, 2013 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
New Details Show Broader NSA Surveillance Reach
The Wall Street Journal (08/20/13) Siobhan Gorman; Jennifer Valentino-Devries
The U.S. National Security Agency (NSA) surveillance network can reach about 75 percent of all U.S. Internet traffic, casting a much wider net than previously disclosed, according to current and former officials. Telecom companies filter data for NSA, and many unrelated communications are collected in the search for those that originate or end abroad, or are foreign communications passing through the United States. In addition, NSA sometimes retains email content sent between U.S. citizens and filters domestic phone calls made with Internet technology. NSA says the programs are legal and respect Americans' privacy. If American communications are "incidentally collected during NSA's lawful signals intelligence activities," the agency follows "minimization procedures that are approved by the U.S. Attorney General and designed to protect the privacy of United States persons," says NSA's Vanee Vines. The programs use complex algorithms to filter out valuable information, notes a former top intelligence official. Although President Barack Obama recently proposed changes to NSA's monitoring activities that are designed to improve oversight, they would not affect the systems that NSA uses for its most sensitive U.S. surveillance practices.
Facial Scanning Is Making Gains in Surveillance
The New York Times (08/21/13) Charlie Savage; Kitty Bennett
The U.S. government's development of a surveillance system in which computers linked to video cameras use facial recognition to identify individuals in crowds is progressing, according to newly released documents and interviews with project researchers. The Department of Homeland Security tested the crowd-scanning Biometric Optical Surveillance System (BOSS) in the fall, but technical specialists still cite speed and reliability issues as obstacles. The BOSS project aims to address such problems by generating far more information for computers to analyze. The system is comprised of two towers supporting robotic camera structures outfitted with infrared and distance sensors. They capture images of the same subject from different angles, and then a computer processes the images into a three-dimensional signature compiled from data such as the ratios between various points on a person's face to be compared against data about faces housed in a watch-list database. Tests indicate that the system is still not accurate and fast enough for use by U.S. police departments, its intended users. However, Electronic Warfare Associates' Ed Tivol and University of Louisville researcher Aly Farag believe these barriers will be overcome as computer processing accelerates. Tivol says BOSS' goal is to match faces to people with 80 to 90 percent accuracy from up to 100 meters away, with the system identifying matches in less than 30 seconds.
Facebook, Six Other Tech Companies Launch Partnership to Bridge the Digital Divide
IDG News Service (08/21/13) John Ribeiro
Closing the gap between those who have Internet connectivity and those who do not is the goal of Internet.org, a joint project between Facebook and six other technology companies. The project's founding members will develop joint efforts, exchange knowledge, and collaborate with industry and governments to link more people to the Internet using mobile technology, Facebook says. The company estimates that only about 33 percent of the global population has Internet access, and the annual rate of Internet adoption growth is less than 9 percent. Internet.org also intends to focus on the reapportionment of white space spectrum to reduce the cost of Internet use, as "licensing spectrum from governments worldwide costs tens of billions of dollars and these costs are passed along to consumers through data plans," according to a plan posted by Facebook CEO Mark Zuckerberg. The plan aims to lower access costs through initiatives such as inexpensive smartphones and collaboration with mobile operators to implement online connectivity in underserved communities. Facebook and its partners also plan to apply data-compression tools, networks that manage data more efficiently, systems that better cache data, and frameworks for apps to reduce data usage by apps and other services.
Here's What You Find When You Scan the Entire Internet in an Hour
The Washington Post (08/18/13) Timothy B. Lee
An ordinary server can scan every address on the Internet in less than an hour using the ZMap tool announced by University of Michigan researchers at the Usenix conference. Rather than retaining a list of outstanding requests, ZMap encodes identifying data in outgoing packets so it can recognize responses, enabling the tool to scan the whole Internet in 44 minutes with a gigabit network link. Among the insights gleaned from the scans the researchers conducted is the increasing use of encrypted HTTPS by websites over time, with ZMap determining that HTTPS use by the top 1 million sites rose about 23 percent in the last year, while overall HTTPS use increased by nearly 20 percent. The researchers also used ZMap scans to ascertain the extent of Internet disruption wrought by Hurricane Sandy in October. In addition, the researchers say ZMap offers a fast and efficient way to evaluate how long it takes users to apply patches to security vulnerabilities from software manufacturers. For example, the program could enable security researchers and software vendors to pinpoint weaknesses and alert systems administrators before public disclosure.
10 Hottest IT Jobs: Developers, Developers, Developers
Network World (08/16/13) Ann Bednarz
Software developers are overall the most desirable information technology (IT) hires companies are seeking, with mobile developers especially sought-after, according to Modis' list of the 10 most popular IT jobs. Modis' Dan Pollock reports that developers with expertise in iOS and Android platforms are very desirable, while companies also are after professionals familiar with PHP, HTML5, and Ruby on Rails. Network pros' popularity is growing as well as companies transition to cloud environments, mainly to save money on infrastructure costs, Pollock says. "Understanding what to outsource and what's critical to keep in house--that's very important," he points out. Among the hottest hiring markets are the San Francisco Bay Area, Los Angeles, Phoenix, Minneapolis, and Houston. Modis estimates that the IT sector's unemployment rate has fallen from 11.2 percent at the peak of the recession to 5.8 percent today, and that demand for IT people outstrips supply in many markets. Pollock notes the third quarter is particularly busy in terms of hiring, and he says "the war for talent for highly technical people is raging," especially in technology hubs such as Silicon Valley. Moreover, a boom in investment for new technology projects is stoking demand for help desk pros.
How Footprint Recognition Software May Help Zoology
Technology Review (08/19/13) Conor Myhrvold
WildTrack provides a low-cost and noninvasive way to track animals. The organization has developed image-processing software that detects physical footprint characteristics that are hard for an untrained eye to recognize. The method, called footprint identification technique (FIT), relies on professional trackers to photograph footprints with a ruler for scale, and add global positioning system coordinates. The footprints are loaded into software that enables WildTrack to match them to a large number of known footprints from captive animals of the same species. Algorithms are then used to compare elements of the photographed footprint against those in a database of animals whose age and gender are known. FIT has a 90-percent accuracy rate in correctly determining the sex, age, and species, but it does not work well with all animals and remains in an experimental stage. WildTrack has been developing the software over the past decade, and it is being used to track tigers in Russia, tapirs in South America, and polar bears in Canada.
Master's Degree Is New Frontier of Study Online
The New York Times (08/18/13) Tamar Lewin
The Georgia Institute of Technology's offering next January of a low-cost master's degree in computer science via massive open online courses (MOOCs) could herald a sea change in higher education, which up to now has not awarded credit or degrees for completing such courses. "This is the first deliberate and thoughtful attempt to apply education technology to bringing instruction to scale," notes University of Maryland physicist S. James Gates Jr. "It could be epoch-making." The basis for the new program is a partnership between Udacity founder Sebastian Thrun and Zvi Galil, dean of Georgia Tech's College of Computing. The school will provide MOOC content and instructors while Udacity will supply the computer platform and course assistants, with revenue split 60-40. Degree program participants will take proctored exams and will be able to access tutoring, online office hours, and other support services. However, some Georgia Tech faculty members are concerned that the MOOCs could devalue a degree from the university. Association of Public and Land-Grant Universities president M. Peter McPherson speculates that online learning will become more commonplace, but no single absolute model will emerge.
Researchers Seek Better Ways to Track Malware's Family Tree
Dark Reading (08/15/13) Robert Lemos
At the recent USENIX security conference, a trio of researchers from Carnegie Mellon University presented a paper outlining a new method for tracing the ancestry of malware. Jiyong Jang, Maverick Woo, and David Brumley created a program called iLINE that uses the relationships between program code to construct "family trees" for malicious and legitimate software, as well as a system called iEVAL that evaluates the correctness of the family trees. Woo says the researchers initially expected their efforts to be fruitless, but that they instead found that even evaluations based on very simple factors, such as the size of files or sections of code, yielded an accuracy rate of greater than 90 percent. Jang says the ultimate goal of their technique is to find a way to track a particular piece of malicious code back to its root and the individual or group responsible for it. However, there are limitations to the method as it exists now. The researchers note that failing to correctly identify a root program can reduce the accuracy of a family tree by half, but Woo and Jang say they intend to continue to develop the method and bring its accuracy rate closer to 100 percent.
U.S. Schools Need More STEM Training, Better Broadband
CIO (08/19/13) Kenneth Corbin
U.S. Education Secretary Arne Duncan recently reaffirmed the Obama administration's call for faster broadband connections in schools along with greater concentration on science, technology, engineering, and math (STEM) training, through projects such as the White House's ConnectED initiative. Although the administration estimates that the average school has an Internet connection equal to the typical U.S. household, it provides service to 200 times as many users, and less than one-fifth of teachers rate their school's online link as satisfactory. The five-year ConnectED project aims to connect 99 percent of U.S. schools and libraries to high-speed broadband, and the White House is asking the Federal Communications Commission to lead the effort to increase broadband services at schools and libraries to at least 100 Mbps, with a goal of 1 Gbps. Duncan says ensuring that all students have access to super-fast broadband services will help the country maintain its global lead in innovation. Meanwhile, the White House's Educate to Innovate program has so far raised more than $700 million in funding through private-public partnerships to encourage STEM education, as well as growing the ranks of STEM educators by 100,000 over 10 years.
Competition: Computer Science Teams Determine the Perfect Baby Name
University of Kassel computer scientists Juergen Muller and Folke Mitzlaff have applied knowledge engineering methods to the task of helping parents-to-be choose the right baby name. In early 2012, Mitzlaff created the name search engine Nameling, which enables users to type in names that they like to generate a list of suggested names that they also might find appealing. The team has since refined Nameling so that users receive recommendations according to their own or previous user search profiles. A new feature also lets users view suggested names based on a list of several names, enabling them to search specifically for first names that complement parent and sibling names. Mitzlaff and his colleagues won the call for proposals to hold the 15th Discovery Challenge, an annual international competition within the scope of the European Conference on Machine Learning and Principles and Practices of Knowledge Discovery, which will take place in Prague this September. Participating researchers will program competing modules capable of answering queries, and the team whose recommendation system delivers the suggestions that best satisfy Nameling users will win. The researchers also plan to incorporate opinion mining to draw conclusions about how names are perceived based on public data from social media.
Image-Processing 1,000 Times Faster Is Goal of New $5M Contract
University of Michigan News Service (08/14/13) Nicole Casal Moore
The U.S. Defense Advanced Research Projects Agency has awarded University of Michigan professor Wei Lu a contract of up to $5.7 million to design and manufacture a computer chip based on self-organizing, adaptive neural networks. Lu says the technology will be able to process images and video 1,000 times faster using 10,000 times less power than current systems, while maintaining accuracy. "Instead of processing all of [the image or video data] or transmitting it fully and wasting precious bandwidth, adaptive neural networks can extract key features and reconstruct the images with a much smaller amount of data," he says. The networks will be comprised of memristors that execute both logic and memory functions, and researchers say their multitasking capacity could facilitate new computing platforms that can process an immense number of signals in parallel and are capable of advanced machine learning. Systems that employ them could manage big data tasks with much greater efficiency than conventional computers. Lu's goal is to construct a network that uses the memristors as artificial synapses between conventional circuits. A simpler version of the network will employ memristors as memory nodes along traditional wired connections between circuits to enhance the efficiency of machine learning.
Google Unveils Open Source Gumbo HTML Parser Tool
eWeek (08/14/13) Todd R. Weiss
Google has unveiled Gumbo, an open source tool for developers that is a C implementation of the HTML5 parsing algorithm. Google's Jonathan Tang says Gumbo gives developers "a simple library that can serve as a basic building block for linters, refactoring tools, templating languages, page analysis, and other small programs that need to manipulate HTML." Tang notes that Gumbo is written in C to make it easier to interface with other languages. In addition, he says, "Gumbo was built from the start to support source locations and correlating nodes in the parse tree with positions in the original text." Gumbo conforms to the HTML5 specification, and is robust and resilient to bad input, according to the Gumbo project page on GitHub. Gumbo includes support for source locations and references back to the original text and has been tested on more than 2.5 billion pages from Google's index. In June, Google also released its open source Cloud Playground environment, where developers can try out ideas without having to commit to setting up a local development environment.
A Personalized Robot Companion for Older People
CORDIS News (08/14/13)
European Union-funded researchers have created a customizable robot companion to assist elderly people, which could become available to consumers within two to three years. The semi-humanoid robot is equipped with wheels, cameras, sensors, audio, and a touchscreen interface. A consortium of research institutes, universities, and technology companies in seven European countries created the robot over 33 months through the EU-funded MOBISERV project. "This has been a very broad project, we've worked not only on the robot but also integrating it with a smart-home system and with smart clothes," says Herjan van den Heuvel of Smart Homes, the Dutch Expertise Center on Home Automation and Smart Living. The researchers developed smart fabrics, which can be worn or used as bed sheets, that incorporate sensors to monitor vital signs or sleeping patterns. The smart-home environment also includes smart sensors, as well as optical-recognition units and home automation elements to monitor eating and drinking, activity patterns, and potentially dangerous situations. The centerpiece of the project, however, is the robot, which learns to approach users at appropriate times with suggestions and conversation. "Older people were extremely positive about the robot," Van den Heuvel says. "They can see the benefits of the cognitive support it provides and also, if they live alone, they like the idea of having something they can interact with."
Abstract News © Copyright 2013 INFORMATION, INC.
To submit feedback about ACM TechNews, contact: [email protected]
Current ACM Members: Unsubscribe/Change your email subscription by logging in at myACM.