Welcome to the November 4, 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
National Science Foundation Invests $5M-Plus in Big Data Hubs
Network World (11/02/15) Bob Brown
The U.S. National Science Foundation (NSF) this week announced more than $5 million in grants to advance the National Big Data Research and Development Initiative. Seven universities will use the grants to establish four Big Data Regional Innovation Hubs (BD Hubs), which will help organizations from a wide range of fields make better use of big data. Data scientists from Columbia University, the Georgia Institute of Technology, the University of North Carolina, the University of Illinois at Urbana-Champaign, the University of California, San Diego, the University of California, Berkeley, and the University of Washington will help speed the research discoveries of organizations into practice. NSF anticipates spending $10 million on establishing Big Data Spokes, which will focus on specific research areas and address issues such as improving data access and automating the data lifestyle. The first national stakeholders meeting of the BD Hubs is scheduled for this week in Arlington, VA. Participants will discuss governance and sustainability models, coordinate ideas for BD Spokes, and identify next steps. Two public webinars will be accessible on Thursday.
System Automatically Converts 2D Video to 3D
MIT News (11/04/15) Larry Hardesty
A project between researchers at the Massachusetts Institute of Technology and the Qatar Computing Research Institute has yielded a system that automatically renders two-dimensional (2D) soccer game video footage as three-dimensional (3D) video. The technology leverages the graphics-rendering software that powers sports video games, and the converted video can be played back over any 3D device. The researchers configured a Microsoft soccer game to play in a loop, and used Microsoft's PIX video-game analysis tool to continuously store screen shots of the action. For every shot, they also took out the corresponding 3D map. They then weeded out all screen shots except those that best captured the range of possible viewing angles and player configurations the game presented, with an algorithm for gauging the difference between two images. Each screen shot and the associated 3D map was stored in a database. For every frame of 2D video of an actual game, the system seeks about 10 shots in the database that best match it, and then looks for the best pairings between smaller video feeds and smaller screen shot regions. Once those are established, the system superimposes the depth data from the shots on the corresponding sections of the video feed, and then re-integrates the sections to generate a convincing 3D effect.
Materials That Couple Sensing, Actuation, Computation, and Communication
CCC Blog (11/02/15) Helen Wright
University of Colorado at Boulder professor Nikolaus Correll and doctoral student Andy McEvoy have conceived of composite materials that integrate sensing, actuation, computation, and communication thanks to innovations in polymers and the miniaturization of computing devices. They say such materials can perceive their environment at high bandwidth, implement high-speed feedback control to reconfigure their shape or appearance, and address difficult computing problems via distributed algorithms. Applications of such "robotic materials" include a skin that triangulates and classifies textures rubbed against it, smart glass that can shift its opacity and color, and intelligent particles that enable three-dimensional printing of functional mechatronic objects. Correll and McEvoy say it will be possible to design smart composites that mimic various properties of organic systems, such as the camouflage abilities of the cuttlefish. With the addition of autonomy, such composites can lead to robotic systems with human-like dexterity, agility, and intelligence, self-diagnosing and mutable airplane wings, or interactive edifices that let users tailor their appearance and geometry. Addressing the challenges of developing robotic materials entails mapping out their design and control, the distributed algorithms that power them, their applications, and their fabrication. Among the issues that must be resolved to realize these composites is finding a way to predict collective behavior given knowledge on individual interactions.
From Bionics to 'Magic,' MIT Media Lab Celebrates 30 Years of Innovation
Computerworld (10/30/15) Sharon Gaudin
For three decades, the Massachusetts Institute of Technology (MIT) Media Lab has contributed to innovations ranging from the Google Glass precursor to bionic limbs to sophisticated video games. "Magic occurs when you mix technology with the artistry of different disciplines--and that had never been done before," notes the Media Lab's Andrew Lippman. "The time was right to do it." Lippman describes the lab's goal as fostering collaboration between different academic fields to enable students to share ideas and learn to think differently so they can study problems and see unique opportunities. "You don't want people who all come from the same place and read the same books," Lippman says, noting this philosophy has led to the lab supporting "a bubbling brew of differentiation, which brings many perspectives to bear on a problem." MIT graduate researcher Jin Joo Lee, who is developing robots that can teach children storytelling skills and a second language, says she wanted to study at the Media Lab to follow the lead of human-robot interaction pioneer Cynthia Breazeal. Lab research assistant Palash Nandy says he was drawn to the lab by its history of innovation, which laid the groundwork for his study of human empathy toward robots. Meanwhile, MIT professor Cesar Hidalgo has tapped students with computer science, design, and physics backgrounds to create big data visualization technology.
Data Mining Reveals the Extent of China's Ghost Cities
Technology Review (11/02/15)
China has experienced a tremendous amount of urban growth in the last 30 years, but even as people have flooded into its cities, many of the newly built residential areas remain vacant. So-called ghost cities have become infamous, places where blocks of apartments sit vacant because there is no demand for the housing they offer. But it is difficult to accurately determine which areas constitute ghost cities. Many tourist areas, for example, become largely empty during the off season, but this does not make them ghost cities. Researchers at Chinese search engine Baidu have harnessed the data they collect on their more than 700 million users (more than half the Chinese population) to perform a scientific study of ghost cities. The researchers monitored the movements of Baidu users over six months between 2014 and 2015 and used a clustering algorithm to calculate their home location. They then compared the results to a dataset of known residential areas to calculate the urban density of those areas, counting any area with a density of 5,000 people per square kilometer or less as a ghost city. The team was able to identify 50 such areas in China, and also provide data about how they relate to the local geography and larger urban settings.
A Data Genius Has Figured Out the Ultimate Beer-Drinking Road Trip
The Washington Post (10/31/15) Ana Swanson
Data scientist Nathan Yau, who runs the Flowing Data blog, has created a map of the top breweries in the U.S., based on a list of the top 100 breweries in the world from RateBeer, a beer and brewery review site. Although RateBeer says 72 breweries are in the U.S., Yau eliminated one because it does not have an official place to visit and another, Anchorage Brewery, because Alaska is too far to drive. Yau then used the "genetic algorithms" technique to determine the most efficient route between the 70 breweries. The algorithm starts with a handful of solutions, takes the best one, and then compares the result to other solutions until it cannot find a better one. The route covers 12,299 miles, stops in 28 states, and passes through 40. Yau calculates the road trip would take about 20 days, including stops at night, time to sample some beers at each place, and then wait to sober up. The researcher also collected more information about outstanding breweries and brewpubs that are on or close to the route, comprising more than 1,400 within a five-mile radius of the route he planned.
Mastery of AI Has Been 'Harder Than Expected' and 'Future Is Uncertain,' Says Microsoft's AI Chief
TechRepublic (11/02/15) Hope Reese
At this week's EmTech MIT technology conference, Microsoft Research lab director Eric Horvitz stressed the value of considering recent advances in artificial intelligence (AI). He partly credited new leadership at Microsoft for a significant portion of researchers exclusively focused on AI research, and new joint ventures as well. Horvitz cited revolutionary developments in probability and machine learning "fueled by almost limitless storage for capturing data." He also noted a resurgence in machine learning as a central support of AI, with a great deal of it derived from predictive models. Other innovation areas Horvitz cited include automatic machine-reading methods and "personal empowerment tools." He cautioned about uncertainty in the pace of machine-intelligence advancement, with the primary concentration being on general intelligences instead of building "narrow, deep wedges of intelligence." Although Horvitz pointed out significant pattern-recognition technology advances, he emphasized the need to combine these new competencies into a integrative system "where the whole is greater than the parts." Among the priorities Horvitz said researchers should stress amid the uncertainties of AI's future is addressing people's fears about the technology, and considering unexpected applications of AI.
Revealing the Mysteries of the Maya Script
Swiss Federal Institute of Technology in Lausanne (11/02/15) Sarah Bourquenoud
Researchers at Idiap Institute, which is associated with the Swiss Federal Institute of Technology in Lausanne (EPFL), are collaborating with Maya writing specialists at Bonn University in Germany to create high-quality representations of the hieroglyphs found in the three surviving codices from the Mayan civilization. Researchers at Idiap and EPFL's new Digital Humanities Laboratory of the College of Humanities plan to digitally catalog the hieroglyphs. The work is of great interest to researchers who would like to translate the complex and still partially unknown Mayan language. The secrets of the classical Maya were lost with the destruction of most works during the Spanish conquest in the 16th century. The University of Geneva also will assist in the effort to create an online database, which the scientific community will be able to use to research, compare, and annotate the texts. "By combining the work of the Maya experts with [information technology]-based tools, we can make fascinating progress," says Idiap researcher Rui Hu. She says the effort could potentially lead to machine learning for historians.
Why Artificial Intelligence Researchers Love 'Super Mario Bros.'
Motherboard (10/30/15) Jordan Pearson
Artificial intelligence (AI) researchers find the video game "Super Mario Bros." especially amenable to testing their work, which includes AI that can learn the subtleties of game play well enough to beat the game and intelligently construct its own levels. Games are ideal AI testbeds because they share with researchers the goals of developing systems that can acquire the same logic, creativity, situational awareness, and decision-making skills that players need to both play and create new levels. Georgia Institute of Technology researchers Mark Riedl and Matthew Guzdial, who are using software to investigate how machines could help people design games, say the appeal of "Super Mario Bros." resides in the algorithmic challenges inherent in the game's mechanics. "The side-scrolling nature of the game means a lot of the game is unobservable to the AI, whereas many simpler arcade games have all information on screen at once," they note. Moreover, the levels of the game are both patterned and abstract, which Riedl and Guzdial say can "push the envelope" of AI research. The game's nested-pattern structure also is useful for algorithms reliant on lexicographic ordering, according to Carnegie Mellon University researcher Tom Murphy. New York University professor Julian Togelius says the game's popularity and influence is its biggest advantage for AI scientists, because their work garners a lot of attention as a result.
Puzzle Lamps Attain New Dimensions With Disney Research Computer Design Tool
Disney Research (10/30/15) Jennifer Liu
Disney researchers have developed an interactive design tool they say makes it easier to create more intricate decorations. The researchers say the software enables users to think in terms of the shapes they want to create with the interlocking elements, instead of by the task of determining how to connect the pieces to achieve those shapes. The researchers used the same basic interlocking quadrilaterals commonly found in puzzle lamps. They used existing puzzle lamp designs to build a library of basic shapes that can be created with the pieces. The user starts the virtual design process by selecting the desired base shapes, positioning and orienting them as desired. The software then merges the individual shapes into a single one by quickly and automatically determining which elements to connect to achieve the final shape. The software also enables the user to pick and drag parts of the form to extend it into the desired shape. The researchers used the tool to design and build several shapes. For example, a turtle required 123 pieces and was assembled in about two and a half hours, while a flower with 288 interlocking elements took four hours to build. The researchers will present their work this week at ACM SIGGRAPH Asia 2015 in Kobe, Japan.
Northeastern Researchers Unlock Details of Uber's Surge Pricing--and Suggest Ways to Avoid It
Northeastern University News (10/29/15) Thea Singer
Northeastern University researchers have uncovered the details behind Uber's surge-pricing algorithm, and found the company uses "an opaque...algorithm" that changes fares every five minutes, according to professor Christo Wilson. In addition, he says Uber's algorithm divides the cities it services into "discrete 'surge areas,'" each with its own independent price based on the intensity of demand at a particular point in time. Uber's method results in "corner cases," in which a user can walk across the street and get a different price, Wilson notes. The researchers came to these conclusions by programming Uber apps masquerading as real users at 43 different global-positioning system locations throughout San Francisco and Manhattan over four weeks. They collected data on the surge prices and estimated wait times for each ride as well as the location of the request. The researchers also tracked supply and demand, how those dynamics changed over time and distance, and the way surge prices varied by location. "We did a lot of correlation analysis looking at how many cars were getting booked over time and how many cars were available, and you do see high correlation between supply and demand and the surge," Wilson says. The researchers found users can avoid surge pricing by waiting a few minutes or walking a few blocks, possibly leaving the surge areas.
3D Mapping Your World With a Backpack
KQED.org (10/29/15) Derek Lartaud
University of California, Berkeley professor Avideh Zakhor led the development of 3D-mapping technology that has helped to fundamentally reshape the way people navigate through the world. Her team developed the technology that was licensed by Google in 2007 to add ground-level, 3D views of cities to its Google Earth application. The technology now enables anyone with an Internet connection to explore 3D maps of cities, as well as remote places such as Cambodian ruins. Zakhor and her team have continued the research, and now are looking to adapt and improve their previous mapping technology to produce a wearable device that could be used to build 3D maps of the interiors of buildings. One major challenge was how to identify the position of the backpack as it gathered its data. The previous system used global-positioning systems (GPS) for this purpose, but GPS signals cannot penetrate most buildings. Instead, the team turned to simultaneous localization and mapping, a technique used to help robots build maps of their surroundings. After much experimentation, Zakhor and her team produced a 30-pound backpack featuring laser cameras, sensors, and software, which they think could have a wide range of applications, ranging from construction and real estate to search and rescue.
If We Want Humane AI, IT has to Understand All Humans
Wired (10/30/15) Davey Alba
Computer-vision researcher Fei-Fei Li, director of the Stanford Artificial Intelligence Lab, hopes one day artificial intelligence (AI) will be able to not only see and identify human beings, but also to understand and empathize with them. It is an important goal, especially as public anxiety about a future dominated by remote and coldly logical AI continues to grow. Some within the field, including Li, think one way to prevent such fears from being realized is to help ensure that those who build the AI systems of the future are a diverse group. There have been several recent examples of how the AI and broader tech industry's lack of diversity has resulted in embarrassing, although not dangerous, lapses. Apple, for example, was criticized when it was pointed out in 2014 that its new health-tracking system, HealthKit, had no way of tracking menstruation, one of the most important aspects of women's health. Google also had egg on its face this year when the AI that automatically tags pictures for its Photo app tagged an African-American couple as "gorillas." Li is doing her part to try and make the field of computer science more diverse, recently helping to launch an AI summer camp for girls at Stanford University.
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.