Association for Computing Machinery
Welcome to the April 25, 2016 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).


Computers That Crush Humans at Games Might Have Met Their Match: 'StarCraft'
The Wall Street Journal (04/22/16) Jonathan Cheng; Min Sun Lee

The next frontier for game-playing artificial intelligence (AI) could be the popular real-time strategy game "StarCraft," in which players assume the roles of three different warring races competing for domination using stealth and subterfuge. Blizzard Entertainment president Michael Morhaime is eager for StarCraft to be selected for AI trials, noting "we would love to be a milestone on that advance of artificial intelligence, from chess to Go and then us." In March, DeepMind founder Demis Hassabis named StarCraft a potential next target for his AI research team. The game's appeal to developers lies in its uncertainty, as players cannot see the entire strategy board at once, so computers cannot calculate all the possible maneuvers an opponent may perform; this makes bluffing and deception crucial stratagems that give human players an advantage over AI. "In order for a computer to win, it needs to learn how to lie," says professional StarCraft gamer Eugene Kim. The continued dominion of human players over AI programs in tournaments is evident in the former's consistent victories in the annual StarCraft AI challenge run for the past five years by University of Alberta researcher David Churchill. He predicts humans will eventually be dethroned by AI as StarCraft champions, given AI's evolving progress in complex game strategy.

A New Number Format for Computers Could Nuke Approximation Errors for Good
Motherboard (04/24/16) Michael Byrne

Computers have a persistent problem representing fractional numerical values, as it requires a method to encode where precisely within a string of digits a decimal point should be located. The larger the number, the less room there is to represent its fractional value, but high-performance computing scientist John Gustafson has proposed a universal number (unum) format to eliminate the approximation error problem. An unum permits the various "fields" within a binary floating-point number representation to grow and shrink according to required precision. The exponential values needed to scale most decimal numbers are typically a lot less than what can be represented by five bits, as are allocated in a 32-bit float, but the current standard attempts to prepare for the worst case. Gustafson says an exponent field that can contract based on need leaves more digits that can represent actual numerical content. He notes an unum contains three extra fields that make the number self-descriptive, representing whether a number is exact or in between exact values, the size of the exponent in bits, and the size of the fraction in bits. "So not only does the binary point float, the number of significant digits also floats," Gustafson says.

Why Early STEM Education Will Drive the U.S. Economy
CIO (04/22/16) Kenneth Corbin

The Obama administration continued a campaign to advance math and science education last week, focusing on early learning with the announcement of a slate of projects designed to promote the science, technology, engineering, and math (STEM) fields. The projects include an alliance between the Education and Health and Human Services departments and the Too Small to Fail initiative to establish resources for families to help them "incorporate STEM concepts and vocabularies into everyday routines," according to the White House. The Obama administration and the U.S. Education Department see early STEM learning as essential to elevating the national economy, with a forward-looking view of social and economic prosperity. "We would argue [early learning is]...a long-term investment that realizes savings in better long-term academic outcomes, better long-term health outcomes, better long-term success in the workforce," says Education Secretary John King. He refers to research estimating an eight-fold to nine-fold net yield over the initial investment in early STEM education. Roberto Rodriguez, deputy assistant to the president for education, says the STEM fields' central emphasis on exploration and experimentation should naturally appeal to young children. "Part of a high-quality learning experience, whether in a child-care setting, whether in a preschool classroom, or a Head Start program, means having access to high-quality STEM learning," he says.

Outwitting Poachers With Artificial Intelligence
National Science Foundation (04/22/16) Aaron Dubrow

A joint effort between researchers and conservationists in the U.S., Singapore, the Netherlands, and Malaysia seeks to apply artificial intelligence (AI) and game theory to combat poaching and other activities. With backing from the U.S. National Science Foundation and the Army Research Office, a team at the University of Southern California (USC) is behind the development of AI apps such as Protection Assistant for Wildlife Security (PAWS), which aims to boost the efficiency of anti-poaching patrols using data on previous patrols and proof of poaching. PAWS "learns" and improves its planning strategy as it accumulates more data, and it can incorporate complex terrain information, resulting in practical patrol routes that minimize elevation changes. PAWS also can account for natural transit paths with the most animal traffic and thus where poaching is most frequent to generate a "street map" for patrols. The USC team recently merged PAWS with the Comprehensive Anti-Poaching Tool with Temporal and Observation Uncertainty Reasoning to predict the likelihood of poaching attacks with greater accuracy. The team also is developing, via a multi-university/nonprofit alliance, methods to prevent illegal logging in Madagascar using game theory. One such method is Simultaneous Optimization of Resource Teams, an algorithm that integrates maps of national parks and security resource costs to gauge the best resource combination to offer maximum protection to the area in question.

Demand for Computer Science Forces Washington Colleges to Ramp Up
Seattle Times (04/23/16) Katherine Long

Washington state's colleges and universities are scrambling to meet students' demand for computer science courses, which mirrors the market's demand for skilled computer science workers. The schools are creating new programs and related majors, increasing the hiring of professors, and constructing new buildings to accommodate the influx of computer science students. For example, at Western Washington University (WWU), the demand for computer science classes is so strong the school recently sent a letter to 150 potential transfer students, warning them they may have to wait a year before they can start taking courses in the subject. This year, more than 500 students are pre-majoring or majoring in computer science at WWU, up from about 100 students five years ago. Meanwhile, Bellevue College this year won approval from the state legislature to offer a bachelor of science degree in computer science starting this fall. If it wins approval from a state board and an accrediting agency, Bellevue would become the first community college in the state to offer a bachelor of science in a subject that also is available at a four-year university. At the University of Washington (UW), demand is so great only about 33 percent of students who apply to major in computer science and computer engineering are accepted into the program.

Feds Seek Public Input on the Future of IoT
Computerworld (04/25/16) Patrick Thibodeau

The U.S. Department of Commerce is seeking public comment on the "benefits, challenges, and potential roles for the government in fostering the advancement of the Internet of Things (IoT)." Commerce began accepting comments on April 22, and the comment period will last until May 23. The agency plans to make the responses public, which it says will result in the U.S.'s single largest knowledge dump about the future of technology and where Americans think it should go. The government wants to focus on whether the challenges and opportunities arising from the IoT are similar or different when compared to those governments and societies have previously addressed with existing technologies. In addition, the government wants to learn about the most significant opportunities created by the IoT, as well as which technological issues may slow the development of the technology. The government's goal is to map out its policy role, including research, economic development, standards, and security and privacy. The government will take the public comment data and issue a "green paper," which is the name for a tentative government report, to be followed by a "white paper," which serves as an official statement of policy.

With Simple Process, Engineers Fabricate Fastest Flexible Silicon Transistor
UW-Madison News (04/20/16) Renee Meiller

University of Wisconsin-Madison (UW-Madison) engineers and colleagues have pioneered a nanoscale fabrication method that could enable manufacturers to inexpensively fabricate high-performance transistors with wireless capabilities on huge rolls of flexible plastic. The team fabricated a transistor that operates at a record 38 GHz, but simulations show it could be capable of operating at 110 GHz. The researchers say the method upends conventional lithographic approaches, overcoming such limitations as light diffraction, imprecision that leads to short circuits of different contacts, and the need to fabricate the circuitry in multiple passes. Using low-temperature processes, the researchers patterned the circuitry on a flexible transistor--single-crystalline silicon ultimately placed on a polyethylene terephthalate substrate--drawing on a low-cost process called nanoimprint lithography. The transistor consumes less energy and operates more efficiently than existing technology. The method could enable semiconductor manufacturers to squeeze an even greater number of transistors onto an electronic device. "Nanoimprint lithography addresses future applications for flexible electronics," says UW-Madison professor Zhenqiang Ma. "We don't want to make them the way the semiconductor industry does now. Our step, which is most critical for roll-to-roll printing, is ready."

Ready, Set, Think! Mind-Controlled Drones Race to the Future
Associated Press (04/22/16) Jason Dearen

The University of Florida (UF) last week hosted what it called the world's first drone race involving brain-computer interfaces (BCIs). The competition involved 16 pilots who used BCIs to drive drones through a 10-yard dash over an indoor basketball court. "BCI was a technology that was geared specifically for medical purposes, and in order to expand this to the general public, we actually have to embrace these consumer brand devices and push them to the limit," says UF researcher Chris Crawford. Each competitor was equipped with an electroencephalogram headset calibrated to identify the electrical activity associated with particular thoughts in each wearer's brain, such as where neurons fire when the wearer imagines pushing a chair across the floor. Programmers write code to translate these "imaginary motion" signals into commands that computers send to the drones. The UF researchers are inviting other universities to assemble brain-drone racing teams for next year, with the goal of pushing interest in the technology. As everyday life becomes more reliant on Internet-enabled devices, the researchers say they want to discover how mind-controlled devices could expand and change the way humans play, work, and live.

Cellphone Principles Help Microfluidic Chip Digitize Information on Living Cells
Georgia Tech News Center (04/20/16) John Toon

The way cellular networks separate the signal from each user could enable cancer cells, bacteria, or viruses to be detected in a fluid sample. Researchers at the Georgia Institute of Technology (Georgia Tech) have developed a technique, dubbed microfluidic CODES, which uses a simple circuit pattern with only three electrodes to assign a unique seven-bit digital identification number to each cell passing through the channels on a microfluidic chip. The technique adds a grid of micron-scale electrical circuitry beneath the microfluidic chip. "The cells and the metallic layer work together to generate digital signals in the same way that cellular telephone networks keep track of each caller's identity," says Georgia Tech professor Fatih Sarioglu. The technique has enabled her team to track more than 1,000 ovarian cancer cells with an accuracy rate of better than 90 percent. Sarioglu's system would require no more than the processor power of smartphones that already handle decoding of code division multiple access signals. The team wants to create inexpensive chips that could be used for sophisticated diagnostic testing in physician offices or remote locations.

Helping Computers Learn to Tackle Big Data Problems Outside Their Comfort Zones
A*STAR Research (04/20/16)

A*STAR Institute for Infocomm Research scientists have developed a framework they say could help computers learn how to process and identify images faster and more accurately. The framework can be used for image segmentation, motion segmentation, data clustering, hybrid system identification, and image representation. One of the challenges of the big data era is to organize out-of-sample data using a machine-learning model based on "in-sample," or observational data, according to A*STAR researcher Peng Xi, who has proposed three methods as part of a unified framework to tackle this issue. The first method focuses on sparsity, while the other two focus on low rank and grouping effects. "By solving the large-scale data and out-of-sample clustering problems, our method makes big-data clustering and online learning possible," Peng says. The framework splits input data into in-sample data or "out-of-sample" data during an initial sampling step. The in-sample data is then grouped into subspaces during the clustering step, after which the out-of-sample data is assigned to the nearest subspace; these points are then designated as cluster members. During testing, the researchers found the new framework outperformed existing algorithms and successfully reduced the computational complexity of the task while still ensuring cluster quality.

Using Data to Explore Poetic Sound
MSUToday (04/19/16) Kristen Parker; Sean Pue

Michigan State University (MSU) researcher Sean Pue is studying the role of sound in modern South Asian poetry. He is conducting the research with the help of an Andrew W. Mellon Foundation New Directions Fellowship, which promotes cross-disciplinary research by assisting humanities faculty with training outside their discipline. Pue will use the $213,000 award to pursue training in computational mathematics, computer science, and natural language processing. "My project will address an understudied aspect of one of the most enduring questions in South Asian literary studies--namely the relationship between language and community--through a completely new method," Pue says. He will use computational tools to study rhyme, stress, and patters of repetition and compare written text and performance. Pue also will develop best encoding practices for linguists and scholarship librarians by creating digital texts and performances for Hindi and Urdu. "This project is part of an emerging strength in culturally engaged digital humanities at MSU and Mellon's support for Sean's work is a telling sign of the quality of the scholarship our faculty is producing," says Christopher Long, dean of the MSU College of Arts and Letters.

One App for a Smooth Multi-Device User Experience
CORDIS News (04/19/16)

The Dynamic Media Service Creation Adaptation and Publishing on Every Device (MEDIASCAPE) project was launched to help broadcasters provide a socially engaging and comprehensive entertainment experience. MEDIASCAPE enables broadcasters to seamlessly interact with viewers across multiple devices, rather than providing content either through a TV or a streaming service. Many of today's applications are designed for a specific device, which means broadcasters have to design, implement, distribute, and maintain a set of complex apps tailored to each individual device. MEDIASCAPE wants to build consistency into the app itself, enabling it to run independently across multiple devices. The project is providing the standardized technology developers need to create the connected service apps required for advanced multi-user services. MEDIASCAPE provides free access to an array of information on the various application program interfaces and user-oriented media application prototypes it is developing, allowing developers to incorporate them into their own applications. The MEDIASCAPE project already has developed the User Interface Engine, which enables developers to create responsive and adaptive user interface layouts for multi-device media applications.

Abstract News © Copyright 2016 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