Welcome to the December 11, 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
A Learning Advance in Artificial Intelligence Rivals Human Abilities
The New York Times (12/10/15) John Markoff
Researchers at the Massachusetts Institute of Technology, New York University, and the University of Toronto have disclosed an artificial intelligence (AI) advance in which a computer-vision program surpassed a group of people in identifying handwritten characters based on a single example. The AI uses Bayesian Program Learning (BPL) to quickly assimilate the characters in a range of languages and generalize from that knowledge. BPL differs from machine learning based on deep neural networks in that it can grasp new handwritten characters after "seeing" a few or only one example. The researchers suggest the software program learns in a manner similar to human learning. Meanwhile, on Thursday the organizers of the Imagenet Large Scale Visual Recognition Challenge reported a successful reduction of the error rate in software for finding and classifying objects in digital images. The winning team from Microsoft Research's Beijing lab halved the number of errors in a task requiring their program to classify objects from a set of 1,000 categories. These and other innovations reflect the increasing concentration on AI in Silicon Valley and elsewhere, but computer scientists warn against concluding that such advances imply a challenge to human intelligence.
In Belgium, an Encryption Powerhouse Rises
The Wall Street Journal (12/10/15) Amir Mizroch
Belgium's University of Leuven has become a stage where privacy and surveillance advocates are fighting over digital encryption, which is researched at the school's prestigious Computer Security and Industrial Cryptography group (COSIC). The center is led by professor and privacy proponent Bart Preneel, who has drawn criticism from those who think his work at Leuven complicates law enforcement and government efforts to foil or trace criminals and terrorists. "Academics have to be aware of the need to act responsibly when their discoveries can lead to harm," argues former GCHQ director David Omand. Preneel counters he is not working against government agencies, but he emphasizes "citizens need to be able to do some things in secret legitimately." COSIC's researchers develop new encryption for corporate clients, or assess their in-house antihacking technology. Preneel reports Intel is one client, partly underwriting COSIC's work in protecting different apps on a smartphone so they cannot steal data from each other, or siphon off data from the device's user. A team from Preneel's lab is testing encryption protecting chip-enabled passports, SIM cards, and other kinds of digital identification. Another group is hacking medical devices to boost their security. The highly popular Advanced Encryption Standard also has roots in Leuven, created by two university graduates in 2001.
Why Google's New Quantum Computer Could Launch an Artificial Intelligence Arms Race
The Washington Post (12/10/15) Dominic Basulto
The new D-Wave 2X quantum computer, recently announced by Google and the U.S. National Aeronautics and Space Administration (NASA), could help set off an artificial intelligence (AI) arms race. Google says the D-Wave 2X is 100 million times faster than any other machine, enabling it to theoretically complete calculations within seconds to a problem that might take a digital computer 10,000 years to calculate. Google, which claims the D-Wave 2X represents a major advancement for computing and AI, refers to its research being carried out at NASA's Ames Research Center as "quantum artificial intelligence" because machine-learning problems that are too hard or complex for today's computers could be solved almost instantaneously in the future. The technology is competing against IBM's Watson supercomputer, which is working to optimize the research and development process for a range of industries, as well as the consumer realm. In addition, University of California, San Diego researchers recently developed the memcomputer, which mimics the way the human brain works, storing and processing information simultaneously. Meanwhile, Yale University recently launched its Quantum Institute, which also is working to develop quantum computers. Going forward, Google hopes to show it has learned how to make practical quantum computers for the commercial market.
New Education Bill to Get More Coding in Classrooms
The Wall Street Journal (12/10/15) Yoree Koh
The Every Student Succeeds Act, signed into law by U.S. President Barack Obama on Thursday, recognizes computer science (CS) as important an academic subject as math and English, potentially introducing it into more classrooms across the country. The new law includes CS in the definition of well-rounded education subjects, putting it on the same level as other subjects when state and local policy makers decide how to distribute federal funds. "This week marks a watershed moment for computer science in U.S. schools," says Code.org founder Hadi Partovi. "In just two years, this field has been adopted by all the largest cities, almost 100 school districts. It's great to see the federal government finally recognize this field as a foundational academic subject." Computer science is specifically mentioned in the bill with language relating to curriculum and support around professional development. For example, the law will provide CS educators with the same level of access to professional development opportunities as teachers in other subject areas such as math and science. Partovi notes computer science's acknowledgement in the bill also could prompt states to provide funds for CS education, as many states tend to copy federal law verbatim.
Growth in Computer Science Driven by Student Interest, Societal Need
Princeton Engineering News (12/09/15) John Sullivan
In response to growing enrollment and increasing interest in computer science from other disciplines, Princeton University is expanding its computer science faculty by more than 30 percent. The expansion will add 10 tenure-track positions to the current 28, making the computer science department one of the three largest concentrations at Princeton. The department plans to bring in the new faculty members as soon as possible, and the university will support the expansion with funds in the long term. "Computer science brims with intellectual excitement, offering new insights into age-old questions and novel ways to solve major societal challenges," says Princeton president Christopher L. Eisgruber. Computer scientists at Princeton regularly connect with a range of collaborators across campus, and enrollments in computer science have tripled since the department was launched in 1985. In addition, among students on track to graduate in 2017, 35 percent of Princeton computer science majors are women, nearly twice the national average of 18 percent. The introductory computer science course, called General Computer Science and started 20 years ago by Princeton professor Robert Sedgewick, is the university's most popular class. Sedgewick created the course because he believed there were core principles of computer science that could be taught to a general group of students.
Tiny Chip That Powers Itself From Radio Waves
BBC News (12/08/15)
Eindhoven University of Technology researchers have developed a tiny sensor powered by the radio waves it uses to communicate information, a development the researchers say could help advance the Internet of Things industry. The sensor is designed to measure temperature, but Eindhoven professor Peter Baltus says similar sensors could be developed to measure light, movement, and humidity. The sensor is two square millimeters in size and weighs 1.6 milligrams. The device has an antenna that captures energy from a wireless router, and stores the energy up to a certain threshold, at which point it can measure the temperature and send a signal to the router. Although the current model has a range of just 2.5 centimeters, Baltus thinks it can be extended up to one meter, and possibly even five meters. The sensor can operate beneath a layer of paint, plastic, or concrete, making it ideal for incorporation into buildings. The cost of an individual chip will be about 20 cents, according to Baltus.
Super-Literate Software Reads and Comprehends Better Than Humans
New Scientist (12/09/15) Hal Hodson
New software is learning to read and comprehend the nature of relationships between characters, the passage of time in the text, and whether the characters get what they want. One example of this technology is Columbia University's Declassification Engine software. It is sifting through and analyzing declassified cables from the U.S. State Department ranging across five decades--4.5 million documents--to plot out social connections and look for new narratives concerning the behavior of U.S. diplomats overseas, according to Columbia computer scientist Owen Rambow. "If we could extract social networks from these cables then we can study how the networks of the U.S. changed over time," he notes. University of Maryland, Baltimore County researcher Snigdha Chaturvedi says current natural-language processing systems are proficient at answering fact-based questions, but less capable with opinion. She is developing software designed to understand people's opinions via their writings. Computers equipped with such software will be compelling tools based on their ability to read faster and more widely than humans, says Carnegie Mellon University's Tom Mitchell. "They could acquire experience that could far surpass what we could ever get in our lifetime," he notes.
Wikipedia-Mining Algorithm Reveals World's Most Influential Universities
Technology Review (12/07/15)
Researchers from the University of Franche-Comte in France say they have developed a method of ranking the influence of the world's universities by applying an algorithm used by Google to rank search results in Wikipedia. The Pagerank algorithm works by examining the links between nodes on a network to determine which of the nodes is most important. The researchers applied the algorithm to 24 different language editions of Wikipedia, a database containing 68 percent of all Wikipedia articles. The resulting top 100 list varies in many ways from more traditional rankings of influential universities. It tends to favor older universities that have had a greater cultural impact, granting universities more credit for their importance in the arts and humanities than traditional rankings, which favor achievements in science. It also is more diverse. Although U.S. universities still account for more than a third of the schools on the Wikipedia list, they account for more than 50 percent in traditional rankings. Universities from non-English-speaking countries such as Japan and India feature more prominently on the Wikipedia ranking. However, the top 20 is still dominated by U.S. and U.K. universities, with a top five of Cambridge, Oxford, Harvard, Columbia, and Princeton.
This Is How You Count All the Trees on Earth
The Register (UK) (12/09/15) Drew Cullen
Modern algorithms have estimated there are about 3 trillion trees on Earth, but a more specific analysis requires new methodologies. Researchers at the U.S. National Aeronautics and Space Administration (NASA) Goddard Space Flight Center and the University of Minnesota have proposed a tree-counting project that combines better algorithms with less costly and faster data processing. The latter would utilize a mix of NASA Goddard's private cloud, as well as the public cloud by way of Amazon Web Services (AWS). The researchers tested the algorithms in sub-Sarahan Africa because of the lack of deforestation. The project used images from four commercial satellites, but for the trial the project leaders concentrated on data captured by only one. The data was organized into 225 one-kilometer-square chunks, and NASA scientists first had to create a mosaic of those images to produce a single seam and then make this "stand up" before deploying the counting algorithm, a process known as orthorectified mosaicing. The data presents as a parallel problem, making it suitable for parallel processing on general-purpose server clusters, or processing via cloud services. After mosaicing, the data was pre-staged in AWS and processed in 200 instances using AWS spot pricing.
First Indian Student Supercomputing Challenge Coming This June
HPC Wire (12/07/15) Tiffany Trader
Next June, the inaugural distributed and embedded-High Performance Computing (de-HPC) symposium will host the first Indian Student Supercomputing Challenge (ISSC) to introduce the next generation of students to the HPC community and its technology. As part of the competition, eight to 10 student teams will build a small cluster computer of their own design and will use it to run a series of HPC benchmarks and applications. The teams also will be required to present their findings to a panel of judges to show how fundamentally they understand the applications and results. "I believe ISSC opens up an interesting and innovative platform for the HPC student fraternity in India to experience computer science and computational science," says Indian Institute of Technology Bombay research fellow Umesh Gupta. The de-HPC symposium will be a full-fledged conference highlighting HPC-based research in various domains, along with an opportunity for HPC vendors from around the world to introduce their latest products. "The idea of the de-HPC initiative is to bring the entire HPC community of the country, be it academics, government organizations, and vendors with top HPC products, under one roof and give them a platform to collaborate and work towards solving bigger inter-disciplinary challenges," says Syncthreads Computing director Rhushabh Goradia.
A Cure for Medical Researchers' Big Data Headache
Oak Ridge National Laboratory (12/07/15) Jonathan Hines
Researchers at the Oak Ridge National Laboratory (ORNL) say they have developed an advanced tool for literature-based discovery that has the potential to accelerate medical research. Oak Ridge Graph Analytics for Medical Innovation (ORiGAMI) is designed to reason with the knowledge of every published medical paper every time clinical researchers use the tool, which will help them find unexplored connections in medical literature. The researchers note ORiGAMI unites big data, graph computing, and the Semantic Web. The ORNL collaboration with the U.S. National Library of Medicine relied on the high-performance computing resources of ORNL's Compute and Data Environment for Science to develop the smart data tool. ORiGAMI is a free online application capable of delivering health insights in less than a second based on the combined knowledge of a worldwide medical community. Georgia Tourassi, director of ORNL's Health Data Sciences Institute, describes the tool as "computer-assisted serendipity" because it enhances instead of replaces the person making the discovery. Tourassi says the tool will guide researchers in certain ways to enable those moments of epiphany. ORiGAMI could potentially boost the efficiency of medical research by directing researchers toward the right questions, an outcome that could reduce costs and speed up delivery of new treatments.
Tech Breakthroughs Are Giving Animals the Power to Speak
Wired (12/07/15) Clive Thompson
Scientists are developing new technologies to enhance human-animal communication and interaction. One example is the work of the Georgia Institute of Technology's Melody Jackson. She is equipping service dogs with computerized vests to help summon emergency assistance via audio messages. The wearable computer can help dogs find people when their owner is incapacitated or in distress, and then pull a lever to trigger emergency messages. Such messages include an SOS alert that also provides global-positioning system coordinates. When Jackson created an accompanying touchpad for service dogs to be used in tandem with the vest, the animals learned it would detect a signal if they hovered their nose over the pad. Meanwhile, Open University's Animal-Computer Interaction Lab has developed a sensor that measures the sniffing pattern of cancer-detecting canines as they evaluate samples. Other animal-computer interaction technologies under investigation include a system for helping humans and dolphins communicate, video games that cats can play with humans, and sensors developed by Jackson to monitor horses and broadcast signals of impending lameness.
New PBS KIDS ScratchJr App Launches
MIT News (12/03/15)
The Public Broadcasting Service (PBS) has released an app designed to help kids between the ages of five and eight learn the basic concepts behind coding. The PBS KIDS ScratchJr app, based on the ScratchJr programming language co-developed by the Massachusetts Institute of Technology's Media Lab and Tufts University, tasks kids with using colorful programming blocks to create their own stories and games using more than 150 PBS KIDS characters. Kits can snap the programming blocks together to make characters jump, move, dance, and sing. The app is now available for free on the iOS App Store and the Android Google Play store. The app was developed as part of the Corporation for Public Broadcasting and PBS Ready to Learn Initiative, which is funded by the U.S. Department of Education and aims to promote computer science education for children. The app will be promoted in underserved communities in the U.S. through programs and partnerships with Title I schools supported by PBS member stations, the Verizon Foundation, and the Ready to Learn Initiative. The promotional efforts will include after-school activities, a summer camp, and teacher-training pilot programs.
Abstract News © Copyright 2015 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
|