Welcome to the February 24, 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).

Rethinking HPC, illustration Rethinking HPC Platforms for 'Second Gen' Applications
HPC Wire
John Russell
February 22, 2017

Researchers at the universities of Edinburgh and St. Andrews in the U.K. suggest a new approach to high-performance computing (HPC) infrastructure involving containerization is necessitated by second-generation applications. "Many of the emerging second-generation HPC applications move beyond tightly-coupled, compute-centric methods and algorithms and embrace more heterogeneous, multi-component workflows, dynamic and ad-hoc computation, and data-centric methodologies," the researchers say. They define a set of "container HPC" (cHPC) services as an array "of operating-system-level services and [application programming interfaces] that can run alongside and integrate with existing job via Linux containers to provide isolated, user-deployed application environment containers, application introspection, and resource throttling via the cgroups kernel extension." The researchers also note the main thrust of their work is assessing a prototype system "to carry out detailed measurements and benchmarks to analyze the overhead and scalability of our approach."

Full Article
Artificial Intelligence in Quantum Systems, Too
University of the Basque Country (Spain)
February 22, 2017

Researchers at the University of the Basque Country in Spain are conducting simulations in quantum biomimetics, which involves reproducing properties of living entities through quantum systems. The researchers are emulating natural selection, memory, and learning processes. In the first quantum simulator, a natural selection environment was created in which there were individuals, replication, mutation, interaction between individuals and the environment, and death. The researchers say this quantum mechanism could support self-replicating systems and automatic processes on the quantum scale. The mechanism to simulate memory, which consists of a system governed by equations, could be used to study quantum systems in different ambient conditions or on different scales. Finally, the learning mechanism was designed to optimize well-defined algorithmic tasks and improve the error margins of the operations. The researchers say this model enables them to teach a machine a function without feeding it the result beforehand.

Full Article

human eye with tech overlay Project Looks at Human Eye to Sharpen Sight of Robots and Drones
The Engineer (United Kingdom)
Helen Knight
February 22, 2017

Researchers at Kings College London, University College London (UCL), and Kingston University in the U.K. are leading the Engineering and Physical Sciences Research Council-funded Internet of Silicon Retinas (IOSIRE) project, which aims to develop advanced machine-to-machine communication systems that capture and transmit images from highly efficient vision sensors mimicking the human retina. Conventional cameras waste a lot of memory, computing power, and time by generating entirely new images for each frame, according to UCL principal investigator Yiannis Andreopoulos. However, recently developed dynamic vision sensors (DVS) emulate the way the retina works by only updating the image at those points where a movement or change in the scene has occurred. The researchers say this significantly increases the speed at which the sensors can produce video frames, resulting in rates up of up to 1,000 frames a second, compared to up to 30 frames a second with conventional cameras.

Full Article
Researchers From Google, CWI Break SHA-1 Hash Encryption Function
Jaikumar Vijayan
February 23, 2017

Researchers at Google and the CWI Institute in the Netherlands have broken the U.S. National Security Agency's Secure Hash Algorithm-1 (SHA-1) cryptographic function with a collision attack. They mathematically produced identical SHA-1 hashes for two dissimilar sets of content. "Collision occurs when two distinct pieces of data--a document, a binary, or a website's certificate--hash to the same [value]," the researchers note. They warn an attacker "could then use this collision to deceive systems that rely on hashes into accepting a malicious file in place of its benign counterpart." Two years of development preceded the demonstration, which employed Google's cloud computing infrastructure and involved 9 quintillion computations. The researchers say the first phase required 6,500 years of central-processing unit computation, while the second phase took another 110 computational years with graphics-processing units. The attack's disclosure will likely spur urgings that SHA-1 be deprecated for critical functions.

Full Article
Can an Algorithm Detect a Speaker's Mood?
The Wall Street Journal
Daniel Akst
February 23, 2017

Researchers at the Massachusetts Institute of Technology (MIT) have developed an algorithm to determine a speaker's mood in real time by registering not only their speech but also their vital signs. MIT's Mohammad Mahdi Ghassemi and Tuka Alhanai fed the algorithm snippets of dialogue tagged as positive or negative so it could deduce telltale patterns that it could later apply in its own labeling. The algorithm also was trained on word definitions. The researchers tested its abilities by having 10 volunteers tell a tale that was happy or sad, while Ghassemi and Alhanai asked questions to approximate a dialogue. A wristband computer worn by the participants collected physiological and movement data transmitted to the algorithm. The algorithm inferred whether a conversation was happy or sad with 83-percent accuracy, and provided a helpful evaluation every five seconds at a rate 14 percentage points better than chance.

Full Article

blinking LED Malware Lets a Drone Steal Data by Watching a Computer's Blinking LED
Andy Greenberg
February 22, 2017

Researchers at Ben-Gurion University in Israel have a developed a method using aerial drones to steal data off computers by bypassing "air gap" protections. Their method involves infecting the computer with malware that can then siphon information from the machine's blinking hard drive light-emitting diode (LED) indicator, if a surveillance tool has a line of sight to the target. The researchers found the technique could transmit data as fast as 4,000 bits a second. They also found when their malware read less than 4 kilobytes from the computer's storage at a time, they could induce the LED indicator to blink for less than a fifth of a millisecond, and use these blinks to transmit messages to cameras and light sensors from a compromised computer via on-off keying. However, the researchers note their exfiltration scenario can be countered by keeping air-gapped machines away from windows or obscuring the LED.

Full Article
Split Decision in First-Ever Quantum Computer Faceoff
Gabriel Popkin
February 21, 2017

For the first time, two different quantum computers have been compared and tested in an algorithm-crunching exercise. One of the computers is built around five ytterbium ions in an electromagnetic trap and manipulated by lasers. The second computer, built by IBM, contains five loops of superconducting metal that can be manipulated by microwave signals. In the ion computer, each quantum bit (qubit) is an ion in which an electron can be placed at one energy level to represent 0, another to represent 1, or both levels simultaneously. In the IBM computer, electric current can circulate with one of two different strengths or at both levels at once. The ion computer achieved a 77.1-percent success rate in one exercise, while the superconducting computer succeeded only 35.1 percent of the time. However, IBM's device completed a two-qubit operation 1,000 times faster than the ion computer.

Full Article
Here's Why Self-Driving Cars May Never Really Be Self-Driving
Lucas Mearian
February 23, 2017

Researchers at Carnegie Mellon University (CMU) are exploring myriad unpredictable issues with autonomous car technology that might be solved with embedded software to avoid accidents. Their proof-of-concept KeYmarea X software seeks to determine how self-driving cars should reason through all possible accident scenarios, and determine when to switch control over to human drivers. Regulations governing self-driving vehicle liability currently are lacking, so insurance firms will likely litigate against the automaker for accident claims and not the driver. "Whenever a car gets into a challenging intersection, the computer should shut down, admitting it cannot handle it and allowing the human driver to take over for now--that's completely acceptable," says CMU professor Andre Platzer. "But if a computer doesn't recognize the situation is outside the bounds of what it was designed for, then it's a situation where the human will always have to be in charge."

Full Article
Report: Reforming STEM Ed at Research Universities
Campus Technology
Dian Schaffhauser
February 16, 2017

Instructional methods for science, technology, engineering, and math (STEM) undergraduate programs must change with education research, according to a report from the Association of American Universities (AAU) and the Research Corporation for Scientific Advancement. The report includes case studies that profile current and past efforts to transform teaching and learning in STEM fields. The case studies include the "Science Education Initiative" at the University of Colorado Boulder and the University of British Columbia in Canada, the "Freshman Research Initiative" at the University of Texas at Austin, a program to innovate within the physics and biology programs at Cornell University, and the "Instilling Quantitative and Integrative Reasoning" (INQUIRE) program at Michigan State University. The report is based on the AAU Undergraduate STEM Education Initiative, which aims to encourage STEM departments at member universities to use teaching practices proven to be effective in engaging students in STEM education.

Full Article

NYU Tandon Professors Build AI to Help Autonomous Vehicles Locate Themselves on Digital Maps NYU Tandon Professors Build AI to Help Autonomous Vehicles Locate Themselves on Digital Maps
New York University
February 15, 2017

Researchers at the New York University (NYU) Tandon School of Engineering are building a deep-learning system that will enable autonomous cars to navigate and respond to changing road conditions by matching data from onboard sensors to data on a cloud-based service for automated driving. High-definition maps made for machine-to-machine communication must be extremely accurate, so the researchers aim to enhance their car-to-map precision to within 10 centimeters. Real-time images of the street and surrounding landscape will be taken from cameras, laser-based range-finding technology (LiDAR), and other sensors, says NYU professor Edward K. Wong. Onboard artificial intelligence will link the data to highly detailed information on the cloud-based HERE HD Live Map. In addition, vehicles connected to the cloud could transmit data on road conditions, traffic, weather, obstacles, speed limits, and other variables, enabling the platform to reflect real-time conditions.

Full Article
At Haverford, 'Data Rescue' Aims to Save Vital Climate Research
Philadelphia Inquirer
Susan Snyder
February 18, 2017

The Data Rescue event at Haverford College last week was part of the national Data Refuge project that aims to preserve data on climate change and environment research, which scientists worry could be in jeopardy under the current administration. U.S. President Donald Trump and several of his cabinet picks have expressed skepticism regarding the effect of human activity on climate change. Following Trump's inauguration, nearly all mentions of climate change were removed from the official White House website. Since the preservation project began last month, 36 coordinated events have been held or scheduled around the U.S., each preserving different parts of government websites. At the Haverford events, students and staff downloaded data from sections of the U.S. Department of Energy and the National Oceanic and Atmospheric Administration websites. Data Refuge plans to broaden its focus to include datasets in areas beyond climate and environment science.

Full Article
Catalan Scientists Develop New High-Precision Method for Analyzing and Comparing Functioning and Structure of Complex Networks
Polytechnic University of Catalonia (Spain)
February 10, 2017

Researchers at the Polytechnic University of Catalonia and the University of Barcelona in Spain have proposed a new method for identifying, comparing, and quantifying differences among large nodes of complex networks. Differentiating the function and structure of networks with hundreds of thousands of interconnected nodes has been a difficult task for researchers. Until now, researchers could detect a difference in the number of links in a network or determine the number of connections that were not working, but were unable to identify the location of damaged connections and their importance to the system. In addition to identifying and naming the nodes in a given network, the new high-precision method enables researchers to reliably calculate the distances between points. The team expects the new method will make it possible to compare the functioning of brain networks and other complex systems, such as power distribution networks and social networks.

Full Article
Monash University Wins Up to A$18m Funding to Advance Intelligence Analysis
Monash University
February 7, 2017

Monash University in Australia has received a U.S.$14-million (A$18-million) research grant from the U.S. Intelligence Advanced Research Projects Activity (IARPA) CREATE program to develop artificial intelligence (AI) platforms that can help the U.S. intelligence community. The University of Melbourne in Australia, Syracuse University, and George Mason University also received IARPA grants to work on the CREATE program. The Monash researchers are known for their Bayesian networks research, which involves AI models for reasoning under uncertainty. The Monash researchers will develop a system to enable intelligence analysts to improve the way they build and test arguments about probable outcomes. "What we're developing is a sophisticated tool that will improve the quality of the analysts' reasoning by enabling them to better assess the value of their evidence," says Monash University's Kevin Korb. The new interface should increase the reliability and acceptance of certain arguments, improving the decision making of government officials.

Full Article
Text Data Management & Analysis
ACM Learning Center

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: [email protected]