Welcome to the October 17, 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).
HEADLINES AT A GLANCE
How Vector Space Mathematics Helps Machines Spot Sarcasm
Technology Review (10/13/16)
Researchers at the Indian Institute of Technology Bombay led by Aditya Joshi report developing a new strategy to help computers better detect the presence of sarcasm in sentences. Their algorithm analyzes the similarity of words by examining how they relate to each other in Word2Vec, a database of Google News stories containing about 3 million words. Joshi says extensive analysis determined how frequently words appear next to each other so they can be represented as vectors in a high dimensional space. Similar words are represented by similar vectors, which enable vector space mathematics to capture simple relationships between the words. The researchers say sentences that contrast similar and dissimilar concepts are more likely sarcastic, and they tested this theory using a database of 3,629 quotes, 759 of them tagged by people as sarcastic. By comparing the word vectors in each quote while seeking similarities and dissimilarities, the researchers found the algorithm improved sarcasm detection. Joshi says the algorithm's errors are likely because many of the words contain multiple definitions not captured in the Word2Vec database, or because the word pairs have high similarity scores. He also notes in some cases the algorithm incorrectly identified sentences as sarcastic.
Software That Automatically Recognizes Surfaces Within Complex Three-Dimensional Images Can Benefit Petroleum Extraction
Researchers from King Abdullah University of Science and Technology (KAUST) in Saudi Arabia have developed SurfCut, an algorithm that smoothly detects faults and other three-dimensional (3D) surfaces with high computational efficiency even in noisy and cluttered datasets. Other techniques use fast marching algorithms that track how interfaces evolve with time from an initial seed point, and this method uses efficient computational routines to expand the seed curve step-by-step until mathematical conditions corresponding to a boundary are met. However, requiring software users to define probable surface boundaries makes it difficult to use fast marching algorithms for complex 3D problems. "It's a challenge to extract a surface from an image volume when the boundary is non-empty and unknown," says KAUST researcher Ganesh Sundaramoorthi. "Until now, no algorithm could handle this task." He says SurfCut solves these issues by using fast marching methods to compute the shortest paths between a seed point and a moving interface. Ridge sets are then computed by retracting the interface until rigid topological features emerge. "Our technique allows them to see structures that are impossible to view using [two-dimensional] slices, and is really robust against data imperfections," Sundaramoorthi says.
Bendable Electronic Paper Shows Full Color Scale
Chalmers University of Technology (10/14/16) Mats Tiborn
Researchers from Chalmers University of Technology in Sweden have developed the basis for a new kind of electronic paper that is less than a micrometer thin, flexible, and produces all of the colors that a regular light-emitting diode (LED) display does, while requiring 10 times less energy than a Kindle tablet. The researchers discovered placing conductive polymers on nanostructures is a perfect combination for creating electronic displays as thin as paper. The polymers that cover the whole surface of the electronic paper lead the electric signals throughout the full display and create images in high resolution. The researchers say although the material is not yet ready for application, the underlying technology is in place. The team has tested and built a few pixels, which use the same red, green, and blue colors that together can create all of the colors in standard LED displays. "We are working at a fundamental level but even so, the step to manufacturing a product out of it shouldn't be too far away," says Chalmers professor Andreas Dahlin. He thinks this breakthrough could reduce energy consumption and replace signs and information screens that are not currently electronic with more flexible displays.
Stepping Up Security for an Internet-of-Things World
The New York Times (10/16/16) Steve Lohr
As the reality of an Internet of things comes closer to realization, initiatives must be followed to ensure the technology is secure from hackers, according to Michael Walker at the U.S. Defense Advanced Research Projects Agency (DARPA). "If we want to put networked technologies into more and more things, we also have to find a way to make them safer," Walker says. He and DARPA have created the Cyber Grand Challenge, a contest in which participants are tasked to create automated digital defenses that can spot and correct software vulnerabilities without human assistance. Out of more than 100 teams initially participating in the competition, only three won the finals in Las Vegas in August. The winning teams' leaders say the contest reveals how machine automation and human expertise might be most efficiently integrated with computer security. The winners combined different software methods into automated "cybersecurity systems." Second-place team captain David Melski says, "this was a demonstration that automated cyberdefense is mature enough, and it's coming." Third-place team leader Yan Shoshitaishvili believes systems that combine human expertise and machine power are the likely way forward, especially since humans are superior to computers at understanding context.
Soft Robots That Mimic Human Muscles
Swiss Federal Institute of Technology in Lausanne (10/12/16) Laure-Anne Pessina
Researchers at the Swiss Federal Institute of Technology in Lausanne (EPFL) in Switzerland are developing soft, flexible, and reconfigurable robots that act like human muscles by changing the air pressure in "soft balloons" serving as the robot's body. The researchers also developed a predictive model that can be used to control the mechanical behavior of the robots' various modules. The researchers note the robots can be used in patient rehab, in the handling of fragile objects, biometric systems, and homecare. The researchers demonstrated their model could accurately predict how a series of modules, consisting of compartments and sandwiched chambers, move the robot's limbs. The cucumber-shaped actuators can stretch up to around five or six times their normal length and bend in two directions, depending on the model. "We conducted numerous simulations and developed a model for predicting how the actuators deform as a function of their shape, thickness, and the materials they're made of," says EPFL's Gunjan Agarwal. The researchers also developed soft robots for medical purposes, including a belt made of several inflatable components, which holds patients upright during rehab exercises and guides their movements. The team also is using soft actuators to create adaptable robots that can navigate in cramped, hostile environments.
Giving Credit Where Credit is Due
The UC Santa Barbara Current (10/14/16) Julie Cohen
Researchers from the University of California, Santa Barbara (UCSB), the University of Pennsylvania, and the U.K.'s University of Edinburgh offer a solution in which citations in academic papers are generated automatically. If researchers want to cite the sources of data they use, they might not be able to, because no standard tool exists for generating database citations, notes UCSB professor James Frew. The team suggests by using the same computing power that makes databases possible, database citations can be made more specific while also accurately accounting for all data authors. Frew says this responsibility normally falls to database managers, who would need to define the various ways the data can be viewed, create citation templates for the standard set of views, and provide a computational mechanism to enable researchers to generate citations for specific queries. The researchers demonstrate the new solution's versatility by applying it to two different scientific databases, which are "radically different in both their structure and how they should be cited," Frew notes. "My hope is that our suggestions for automating citations will encourage managers to implement similar systems and make it easier for those using the data to cite it appropriately." The team's findings appear in the September issue of Communications of the ACM.
Making Computer Science Accessible to All Students
EdSource (10/16/16) Trish Williams
The first national consensus document on computer science education, the K-12 Computer Science Framework, was released Monday and provides standards on computer science concepts and practices by grade span in the areas of computer science, digital citizenship, and computer literacy, writes Trish Williams, the California State Board of Education's liaison on computer science. Originally initiated by Code.org, the framework was guided by representatives from ACM, the Computer Science Teachers Association, the National Math and Science Initiative, and the Cyber Innovation Center. Williams says this new framework will help inform state education boards' curriculum decisions, including an initiative by California to establish computer science content standards. She notes the California Department of Education recently announced three focus groups to solicit input from educators and the public. In addition, Williams says the State Superintendent of Public Instruction has been instructed to convene an advisory panel to develop recommendations for a computer science strategic implementation plan by July of 2018. The panel will address issues related to finding qualified teachers, defining education principles, ensuring equal access to instruction, and supporting college readiness through a computer science curriculum.
Facebook, Microsoft, and IBM Leaders on Challenges for AI and Their AI Partnership
IEEE Spectrum (10/13/16) Prachi Patel
Amazon, Facebook, Google, IBM, and Microsoft last month announced the Partnership on Artificial Intelligence, a nonprofit that aims to be a forum to discuss issues on deployment, best practices, and ethical questions surrounding artificial intelligence (AI). The new group also wants to explain the state of the art in AI and what it can do in the future. Thought leaders from the member companies spoke last week at the White House Frontiers Conference in Pittsburgh, detailing why AI has finally arrived and what challenges could be faced in the future. The new organization will serve to dispel myths and serve as a reliable source of information on where the technology is going, according to IBM's Guruduth Banavar. Teaching machines common sense is one of the major frontiers in AI, says Facebook AI director Yann LeCun. "Most of what we [humans] learned, we learned in the first few years of life just by observing the world, and we don't know how to do that with machines, how to teach them to learn by observing the world," LeCun says. Banavar notes another major hurdle is how to deploy AI in the real world. "We will have to design them with cybersecurity in mind, and keeping in mind the sub-society of the Internet," says Microsoft's Jeannette Wing.
UA Uses Big Data to Solve Bus Woes in Brazil
UA News (AZ) (10/12/16) Alexis Blue
The University of Arizona (UA) is collaborating with city planners in Fortaleza, Brazil, to help them better understand and rectify problems with the local bus system by harnessing the power of big data. Sudha Ram, director of UA's INSITE: Center for Business Intelligence and Analytics, and collaborators analyzed several years of data from Fortaleza's bus system, collected from smart cards scanned by passengers. The researchers also analyzed data from the buses' global-positioning system trackers, the locations of Fortaleza's bus stops, weather conditions, and traffic accidents. "We were able to write our own algorithms to derive exactly how much time it takes for a bus to move from one bus stop to the next, so we can see how fast it's moving and where the delays occur and where there are no delays," Ram says. The researchers used this analysis to build an online dashboard that Fortaleza planners can use to refine the bus system. "They can pick a particular bus route, they can pick the direction, they can pick a particular date and a particular hour, and they can see how many people were on the bus and where there were delays," Ram says. She believes the concept is viable for other cities in developed and developing countries.
NSF-Funded Hubs Power Big Data Project 'Spokes'
Campus Technology (10/11/16) Dian Schaffhauser
The U.S. National Science Foundation has announced $11 million for project areas that will serve as "spokes" in the national plan for big data research. Last year, NSF announced funding for four regional hubs focused on data science innovation. Similar to the Big Data Hubs, the 10 spokes will serve in a "convening and coordinating" role, but they only will have a specific mission. For example, one spoke will develop a new standardized modular data-sharing license and a platform that makes data sharing easier and helps to enforce the license. The spoke, which will receive a grant of about $1 million, will partner with various institutions. Other projects in the northeast will examine the use of data in education and the integration of environmental factors and casual reasoning approaches for large-scale observational health research. "The [big data] spokes advance the goals and regional priorities of each [big data] hub, fusing the strengths of a range of institutions and investigators and applying them to problems that affect the communities and populations within their regions," says NSF's Jim Kurose. "We are pleased to be making this substantial investment today to accelerate the nation's big data [research and development] innovation ecosystem."
Why Do Some STEM Fields Have Fewer Women Than Others? UW Study May Have the Answer
UW Today (10/12/16) Deborah Bach
A new University of Washington (UW) study examines why women are more represented in some science, technology, engineering, and math (STEM) fields than others. The researchers focused on six of the largest science and engineering fields with the most undergraduate degrees--biology, chemistry, and math, which have the highest proportions of female participation, and computer science, engineering, and physics, which have bigger gender gaps. After analyzing more than 1,200 papers about women's underrepresentation in STEM, the researchers identified three main factors driving the disparity--a lack of pre-college experience, gender gaps in belief in one's abilities, and a masculine culture that discourages women from participating. Masculine culture is the most powerful factor and its main aspects are stereotypes of the fields that are incompatible with how many women perceive themselves, negative stereotypes about women's abilities, and the lack of role models. The researchers conclude a more inclusive culture across STEM fields is the most effective way to boost female participation, and this can be achieved by developing "subcultures" that make girls feel they belong. "Cultural change is never easy, but there are lots of examples of it being done successfully, and it translates into changing who's in a particular field," says UW professor Sapna Cheryan.
Computer Scientists Explore Random Numbers, Cybersecurity
The Daily Texan (10/07/2016) Freya Preimesberger
University of Texas (UT) at Austin researchers have developed a two-source random extractor, an algorithm that takes two independent low-quality random number generators and combines them into a high-quality one. Calculations in fields ranging from computer security to simulating climate models to polling for public support of a presidential candidate require a great deal of complexity and unknown variables, which are generally too complicated for a computer. Scientists can use random numbers to model this complexity, but conventional random number generation is not completely random, which can alter scientific results or create weaknesses in computer security. The UT researchers found the two-source extractor exponentially improved random number quality, providing greater security. UT professor David Zuckerman says although the extractor must be made more efficient before it becomes practical, it does have applications in theoretical math and computer science. New York University professor Yevgeniy Dodis says random numbers are often neglected in cybersecurity research. "Random numbers are the foundation of everything," he notes. "Without that, you just can't talk about computer security."
Abstract News © Copyright 2016 INFORMATION, INC.
To submit feedback about ACM TechNews, contact: [email protected]