Welcome to the April 10, 2017 edition of ACM TechNews, providing timely information for IT professionals three times a week.
ACM TechNews mobile apps are available for Android phones and tablets (click here) and for iPhones (click here) and iPads (click here).
|
AI Player AlphaGo to Play Chinese Go Champion
IDG News Service
John Ribeiro
April 10, 2017
DeepMind's Go-playing artificial intelligence (AI) program AlphaGo in May will play against top Chinese Go player Ke Jie in a three-game match in China about a year after AlphaGo defeated a South Korean champion. In addition, AlphaGo will participate in games involving both humans and AIs. DeepMind CEO Demis Hassabis says in the "Pair Go" match, one Chinese professional player will oppose another but they will both have their own AlphaGo collaborator alternating moves, "to take the concept of 'learning together' quite literally." AlphaGo also will compete against a team of five top Chinese professional players, whom Hassabis says will be "working together to test AlphaGo's creativity and adaptability to their combined style." AlphaGo already has played Ke and others in unofficial online matches, and Hassabis says the official, full-length games in China should test the program's mettle and help train it to improve strategy.
|
New Computer Vision Challenge Wants to Teach Robots to See in 3D
New Scientist
Matt Reynolds
April 7, 2017
The progress made in training algorithms to recognize images via the ImageNet Large Scale Visual Recognition Challenge has prompted its 2018 replacement with the more formidable task of teaching robots to perceive the world in three dimensions (3D). Compiling a database of images that includes 3D information would enable robots to be trained to identify objects around them and plot out the best navigation routes, says Victor Prisacariu at the University of Oxford in the U.K. The University of North Carolina at Chapel Hill's Alex Berg says the database for the new contest would be comprised of digital models of real-world environments or 360-degree photos that include depth information. He expects it will be at least five years before robots consistently understand their surroundings and can explain what they see as well as humans. Imperial College London's Andrew Davison notes the new competition also should further virtual and augmented reality technologies.
|
Protecting Driver Privacy in the Connected Age
Ben-Gurion University of the Negev (Israel)
Andrew Lavin
April 6, 2017
Researchers at Ben-Gurion University of the Negev (BGU) in Israel have demonstrated the ability to steal a driver's cloud-based private data for Usage-Based Insurance (UBI) programs using only a fragment of the data collected. The researchers employed an algorithm to collect additional information outside of what is provided to UBI firms without the use of global-positioning system data. "Based on our research, an attacker only needs one part of the information provided to a UBI company to discover a driver's whereabouts, home, work, or who they met with," says BGU professor Michael Segal. "As connected vehicle networks become more widely used to collect driver data and provide information or entertainment, the opportunity for someone to uncover private information will also increase." The researchers note driver data increasingly is being stored in the cloud as part of the Internet of Things' connected-device functionality in vehicles.
|
Computers of the Future May Be Minuscule Molecular Machines
Live Science
Tia Ghose
April 5, 2017
The U.S. Defense Advanced Research Projects Agency (DARPA) is launching a new "Molecular Informatics" program to develop a computer system that can store millions of pieces of information in the attributes of molecules, such as orientation, size, and color. In order to circumvent the dependence on conventional binary systems, DARPA says researchers need to invent a new information architecture. "Millions of molecules exist, and each molecule has a unique three-dimensional atomic structure as well as variables such as shape, size, or even color," says Anne Fischer, program manager in DARPA's Defense Sciences Office. "This richness provides a vast design space for exploring novel and multi-value ways to encode and process data beyond the 0s and 1s of current logic-based, digital architectures." Researchers have developed other molecular storage systems in the past; for example, scientists encoded the entire works of Shakespeare in DNA.
|
Carnegie Mellon Artificial Intelligence Takes on Chinese Poker Players
Carnegie Mellon News (PA)
Byron Spice
April 6, 2017
Researchers at Carnegie Mellon University (CMU) have developed a new artificial intelligence (AI) to compete against six professional Chinese poker players in a 36,000-hand exhibition match in China. The Lengpudashi AI is a variant of Libratus, another CMU-developed AI that in January defeated top poker players at Heads-Up, No-Limit Texas Hold'em. However, CMU professor Tuomas Sandholm, who received the 2001 ACM/SIGART Autonomous Agents Research Award, says the new event is not a scientific experiment. The exhibition will involve Lengpudashi and its human rivals playing for 10 hours a day, with the human players each playing two hands at a time. Lengpudashi will operate on the Pittsburgh Supercomputing Center's Bridges computer. "I am very excited to take this new kind of AI technology to China," says Sandholm. "I want to explore various commercial opportunities for this in poker and a host of other application areas, ranging from recreational games and business strategy to strategic pricing, cybersecurity, and medicine."
|
Software System Connects Devices for the Internet of Things
Technical University of Kaiserslautern (Germany)
Katrin Muller
April 5, 2017
Researchers at the Technical University of Kaiserslautern in Germany are developing a user-friendly software system to link devices from different manufacturers within the Internet of Things. "Our technology is similar to the concept of an adapter used to connect differently-shaped plugs with electrical sockets abroad," says the University of Kaiserslautern's Johannes Kolsch. The researchers are using a software interface to connect the devices, which will enable users to decide who can use the devices. "Users can grant permission to someone else," Kolsch says. "Data would then be encrypted before being transmitted from sender to recipient according to the latest security standards." The research forms part of the European Union-funded VICINITY project, and its potential applications could include integrating machines in production facilities, and easing control of large buildings' distribution systems. The researchers say the system also could distribute power from renewable energy sources.
|
In the Future, We Will Control Our Mobiles Using Gestures
Linnaeus University
Anders Runesson
April 4, 2017
Researchers at Linnaeus University in Sweden are developing a next-generation interface that will enable people to interact with mobile devices using gestures. Real-Time Three-Dimensional (3D) Gesture Analysis for Natural Interaction with Smart Devices aims to give users the same experience of, for example, grabbing and twisting an object in the digital world as in the real world, according to Linnaeus' Shahrouz Yousefi. However, he notes the ability to analyze in real time the movements of the hands and individual fingers requires a highly intelligent system with lots of processing power, because the interface requires large amounts of data and advanced analyses, especially when it needs to track the movements of multiple hands simultaneously. The researchers say one of the major hurdles will be developing and using new big data analysis techniques to interpret gestures and movements.
|
Algorithm and Rhyme
Mountain View Voice
Karla Kane
April 5, 2017
San Jose State University professor Margareta Ackerman has developed the Automated LYrical SongwrIting Application (ALYSIA), an artificial intelligence-based system designed to assist musicians in songwriting. By feeding ALYSIA lyrics one sentence at a time, the system constructs a predictive model so it can generate multiple melodic suggestions played back on a computerized piano and written out in musical notation. In addition, ALYSIA ranks these melody choices according to what it thinks best, with highly repetitive melodies often penalized. Ackerman says ALYSIA is unique in its natural-language processing ability, which enables it to learn how words and syllables, and melodies and individual notes, relate to each other. She envisions such systems as collaborative tools, for both experienced and novice musicians, especially those who practice in the electronic-music genre.
|
Technique Makes More Efficient, Independent Holograms
Harvard University
Leah Burrows
April 4, 2017
Researchers at Harvard University have encoded multiple holographic images in a metasurface that can be unlocked separately with differently polarized light, a breakthrough they say could improve holograms for anti-fraud protection and entertainment, as well as offering more control over the manipulation and measurement of polarization. The metasurface consists of an array of polarization-sensitive pillars, called nanofins, which redirect the incident light. Unlike previous arrays, which were uniform in size, the researchers say these nanofins vary in orientation, height, and width, depending on the encoded images. "This metasurface can be encoded with any two images, and unlocked by any two polarizations, so long as they are perpendicular to each other," says Harvard's Noah Rubin. In addition, users can embed different functionalities, including serving as a lens for one polarization and a hologram for another.
|
Few Keep Track of Their Personal Data on the Net
Karlstad University (Sweden)
Maria Wahl
April 4, 2017
A study by researchers at Karlstad University in Sweden found that few users understand how their personal data are compiled and stored on the Internet, which will become a key issue when the European Union's new General Data Protection Regulation takes effect next year. The new regulation will enable users to access and download their personal information, but Karlstad's Farzaneh Karegar says just downloading the data is not helpful for most users. To help users manage their data, the researchers developed Data Track, a tool that enables users to download and visualize their personal information from data service providers. Data Track also has been used in a project to analyze users' views of their data and its portability. Karegar says the analysis found users "have difficulties in understanding the difference between data saved locally on their own computers and remotely on different websites when using transparency-enhancing tools like Data Track."
|
Touch-Sensitive, Elastic Fibers Offer New Interface for Electronics
NCSU News
Matt Shipman
April 4, 2017
Researchers at North Carolina State University (NCSU) have created elastic, touch-sensitive fibers that can interface with electronic devices. "We have created soft and stretchable fibers that can detect touch, as well as strain and twisting," which could be useful for integrating electronics in new places, including wearable devices, says NCSU professor Michael Dickey. The new fibers are made of tube-like polymer strands containing a liquid metal alloy, eutectic gallium and indium (EGaln), which are only a few hundred microns in diameter. Each fiber consists of three strands, including one completely filled with EGaln, one two-thirds filled with EGaln, and one one-third filled with EGaln. The touch-responsive fibers work because of capacitance, the phenomenon in which electric charge is stored between two conductors separated by an insulator; this gives users the ability to send different electronic signals depending on which part of the fiber is touched.
|
Finger Animation Research Begins With CAREER Award in Hand
The Newsstand (SC)
Paul Alongi
April 4, 2017
Clemson University professor Sophie Joerg is using a CAREER award from the U.S. National Science Foundation to lead a five-year project to enable more realistic hand movements for virtual reality (VR) simulation, video games, and animated films. Joerg plans to automate and accelerate finger animation with new algorithms. Among the project's potential applications are hand motions for sign-language avatars, with Joerg noting the creation of a reusable motion database is critical. "We want to automatically say, 'Let's take a half-second from this motion, and then combine it with a half-second from that motion,'" she says. The algorithm then would select which movements to extract from the database to produce the desired animation. Among the issues Joerg says the project will address is whether a virtual character's hand and finger movements are at all important to VR users in interpreting their intention.
|
|
