Welcome to the May 1, 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).

exascale system, illustration U.S., China Engaged in High-Stakes Exascale Computing Competition
Jeff Burt
April 30, 2017

China and the U.S. are spending billions of dollars developing exascale computing technologies, which promise to make possible computers that are 50 times faster than the most powerful supercomputers currently operating. The Chinese government has rapidly increased its investment in exascale computing efforts, highlighted by three projects currently underway. One of the projects involves the development of the Tianhe-3 prototype, which is scheduled to be operational next year. Meanwhile, the U.S.'s Exascale Computing Project (ECP) aims to unveil the first of its exascale-capable systems in 2021, followed by two more two years later. In addition, governments and developers in the European Union and Japan also have exascale efforts in the works. These exascale projects will enable more detailed and complex simulations, as well as the ability to better store, process, and analyze the massive amounts of data now being generated.

Full Article
Mapping the Edge of Reality
David Glanz
April 28, 2017

Researchers at RMIT University and other Australian institutions, as well as at the University of Cologne in Germany, have developed a genetic algorithm to confirm the rejection of classical ideas of causality. The researchers used genetic programming to automatically find the closest classical models for the data, enabling them to map out many dimensions of departure from classical mechanics that quantum correlations exhibit. "We've lightheartedly called the region mapped out by the algorithm the 'edge of reality,' referring to the common terminology 'local realism' for a model of physics satisfying Einstein's relativity," says University of Technology Sydney researcher Chris Ferrie. The algorithm works by building causal models through simulated evolution imitating natural selection. The researchers used photons to generate the quantum correlations that cannot be described using classical mechanics.

Full Article

tech eye image Researchers Examining How State-of-the-Art Camera That Mimics Human Eye Could Benefit Robots and Smart Devices
Kingston University London (United Kingdom)
April 27, 2017

Researchers at Kingston University London and University College London in the U.K. are developing a camera-based artificial vision system for robots and smart devices that simulates the human eye. The researchers will examine the use of neuromorphic sensors that dramatically reduce computing power and data storage needs by only updating the part of an image where movement occurs. Kingston professor Maria Martini notes these sensors emulate how mammals' eyes process information. "These sensors...sample different parts of the scene at different rates, acquiring information only when there are changes in the light conditions," Martini says. The cameras require far less energy and processing requirements, and Martini's team will explore how high-quality footage could be sourced efficiently from the dynamic sensors and then shared between machines or uploaded to a cloud-based server. The researchers also will consider the role the sensors could play in the Internet of Things.

Full Article

Welsh flag Welsh Universities Get Supercomputing Investment from Government
Sooraj Shah
April 28, 2017

Cardiff, Aberystwyth, Bangor, and Swansea universities have received funding from the European Regional Development Fund for "Supercomputing Wales," a supercomputing program designed to enable Wales to compete globally in researching and solving complex scientific problems. The five-year program aims to provide access to powerful computing capabilities to use in existing science and innovation projects in Wales. Some of the funding will be put toward upgrading supercomputer hubs at Cardiff and Swansea, which will employ software engineers and research teams with specific domain expertise to develop algorithms and customized software that exploits the supercomputing facilities to simultaneously perform multiple computational tasks. "This program of investment will ensure that Welsh university research teams have access to facilities to undertake world-class research and to develop new collaborative projects with industrial and other partners," says Cardiff professor Roger Whitaker.

Full Article
This Robot Knows When It's Confused and Asks for Help
Technology Review
Signe Brewster
April 25, 2017

Brown University researchers, in an effort to get people and robots to work together more naturally and effectively, have developed a robot that measures its own confusion and then asks for help if it thinks it is required. Previous work by the Brown team enabled a robot to analyze both speech and hand gesture cues to infer what is being asked of it, and showed this is more effective than voice commands alone. Now, thanks to the team's latest research, a robot will be able to ask questions in order to get a more specific answer. For example, if a person asks for a wrench but there are two wrenches near each other, the robot will point to one and ask if that is the right one. Brown professor Stefanie Tellex says the robot's response represents the latest step toward mimicking the way two people hold a conversation.

Full Article

school of sharks New Shark Spotter Algorithm Lets Surfers Take to the Break With Confidence
Digital Trends
Rachel Cavanaugh
April 27, 2017

Researchers at the University of Technology in Sydney (UTS) in Australia have developed Shark Spotter, an algorithm that uses video footage streamed from drones to detect sharks and alert swimmers. The researchers say the algorithm is 90-percent accurate in distinguishing sharks from dolphins, rays, whales, and other marine life. They note the algorithm's accuracy is a significant improvement over human spotters using their naked eye from helicopters in flight, which are 18-percent accurate, or those spotting from fixed-wing aircraft, which are 12-percent accurate. They developed the app by flying drones over coastal waters and capturing about 8,000 images. The team then created the algorithm using computer systems modeled after the human brain and nervous system. "The system efficiently distinguishes and identifies sharks from other targets by processing video feeds that are dynamic as well as images where objects are static," says UTS professor Michael Blumenstein.

Full Article
A Trick That Hides Censored Websites Inside Cat Videos
Kaveh Waddell
April 26, 2017

Researchers at the University of Waterloo in Canada have developed Slitheen, a system that uses decoy routing to conceal banned websites inside of innocuous cat videos. Decoy routing enables users to view censored websites while producing a browsing trail that looks as if they are perusing harmless sites. Slitheen sends requests for "overt" websites with encrypted secret requests for the actual target embedded within them. Relay stations with Slitheen software installed would detect the secret requests and decode them with a secret key, and then assemble a novel package that defines the Slitheen method. The overt site is downloaded as camouflage for the sensitive data while the target covert site is downloaded concurrently, and its contents substituted for the overt site's contents. Once this is done, the relay transmits the package back to the user that requested it. The complex traffic generated by Slitheen is indistinguishable from a normal request.

Full Article
Establishing the Boundaries of Quantum Secure Communications
University of York
Alistair Keely
April 26, 2017

As researchers around the world build secure quantum networks on a large scale, they are finding it is important to understand the limits of quantum key distribution (QKD) in terms of maximum rates, or capacities, at which two parties can distribute secret keys in a point-to-point connection. Quantum security is not computational but instead is based on the uncertainty principle, a fundamental law of nature. Researchers at the University of York's Center for Quantum Technologies in the U.K. say they have established these capacities through the most important communication lines, including optical fibers. "This is a breakthrough result because it establishes the ultimate performance that any point-to-point protocol of QKD cannot surpass," says University of York professor Stefano Pirandola. "Setting these limits is extremely important for both theoreticians and experimentalists, because they provide benchmarking for new theoretical protocols and actual experimental implementations."

Full Article
Texas A&M Research Team Suggests How to Defend Against Collective Attention Threats
Michael Kassner
April 24, 2017

Researchers at Texas A&M University led by professor James Caverlee are developing a collective-attention threat awareness system to prevent cybercriminals from phishing online users whose attention is diverted by viral Internet content. "It is imperative to develop new techniques to detect, analyze, model, and defend against collective attention threats in large-scale social systems," Caverlee says. He notes current media-based safeguards offer inadequate protection because their "one-size-fits-all method ignores individual risk profiles and suffers from either blocking too much content or allowing all content." Caverlee's team is developing a personalized app to alert each user of their exposure to collective attention threats by detecting evidence of such threats early on. "The ultimate goal is to build a scientific foundation for the understanding of new threats, including algorithms; frameworks and systems; tools that combat threats against company systems--giving users more power to make sense of their online experience," Caverlee says.

Full Article
Computer Geeks Develop Algorithm to Detect Defects in Fruits and Grade Them Automatically
Outlook India
Dinesh C. Sharma
April 27, 2017

Researchers at Annamalai University in India have developed a new image-processing technique that automatically checks the quality of fruits based on their color and texture. The researchers took a set of healthy oranges and another of oranges with skin defects and developed a database of images, each of which were tagged with a series of attributes. The team then used an image-processing tool called a "grey level co-occurrence matrix" to develop the new fruit-checking algorithm. The researchers say the image-analysis technique was used to classify oranges into two commercial grading stages, and successfully extracted useful and meaningful features to uniquely represent the external surface for classification purposes. They note the technique also could be useful for detecting skin damage on other fruits.

Full Article
Professor Designs Search and Rescue Devices
Las Cruces Sun-News (NM)
Minerva Baumann
April 22, 2017

New Mexico State University professor Zachary Toups is developing human-computer interfaces designed to aid urban search-and-rescue operations, using a CAREER award and an earlier related grant from the U.S. National Science Foundation (NSF). Toups says he is using the NSF funding to create custom-built wearable computers to support the mission of the Texas Task Force 1 response team. He also says the research will employ mixed-reality training simulations that integrate virtual and physical-world environments. Toups says his initial NSF grant was used to focus efforts on wearables that can support search and rescue as science migrates from multiple humans piloting a single drone to one human directing numerous drones. The CAREER award will concentrate on complementary wearable interfaces for human search-and-rescue teams, and Toups expects his work will lead to a design catalog of interfaces and best practices for designing wearable and mixed-reality interfaces.

Full Article
Controlling Robot Swarms With Augmented Reality
IEEE Spectrum
Alyssa Pagano
April 22, 2017

Researchers at New York University are using augmented reality to develop a robot control interface that runs on a conventional smartphone or tablet. The system uses the device's camera to capture details from a scene and overlay virtual objects. Users tap and swipe on the screen to make the robots move or pick up objects. The app can detect robots and objects in the environment and create a virtual grid, along with a coordinate system, to keep track of the objects on the screen. In addition, the user can manipulate the objects on the device and watch as the robots complete the desired actions in the real world. The robots and other objects that are part of the system are equipped with visual tags captured by the device's camera. The researchers plan to further develop the technology in order to bring robotic applications outside the lab.

Full Article
MIT Advances in Imaging

Association for Computing Machinery

2 Penn Plaza, Suite 701
New York, NY 10121-0701

ACM Media Sales

If you are interested in advertising in ACM TechNews or other ACM publications, please contact ACM Media Sales or (212) 626-0686, or visit ACM Media for more information.

To submit feedback about ACM TechNews, contact: technews@hq.acm.org