Welcome to the May 7, 2018 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).

To view "Headlines At A Glance," hit the link labeled "Click here to view this online" found at the top of the page in the html version. The online version now has a button at the top labeled "Show Headlines."

Man main gait being measured by wearable sensor with data being recording on two computers Gait Assessed With Body-Worn Sensors May Help Detect Alzheimer's
Newcastle University (UK)
May 4, 2018


Researchers at the University of Newcastle in the U.K. recently conducted a pilot study showing that low-cost wearable devices could improve clinical trial efficiency and encourage dementia research investment. As part of the study, the researchers equipped people with mild Alzheimer's disease with body-worn sensors to be used at home and in a clinic. The sensors assessed the users' walking patterns as a cost-effective way to detect Alzheimer's earlier and track the progression of the disease. The volunteers wore a small sensor on their lower backs, and completed walking tasks in a laboratory. The participants then went home and wore the sensor for seven days, carrying out everyday tasks. The researchers found clinically appropriate diagnostic measures can be obtained for walking behavior and pattern, as well as gait characteristics related to pace, timing, variability, and asymmetry of walking.

Full Article

Plastic tic tac toe game on a wooden table This Rehab Robot Will Challenge You to Tic-Tac-Toe
IEEE Spectrum
Evan Ackerman
May 3, 2018


Researchers at Ben-Gurion University of the Negev in Israel are employing a robot arm that can play tic-tac-toe with people undergoing physical rehabilitation. The game has patients perform repetitive reaching and grasping to make the healing process more entertaining and less strenuous. A Kinova robot arm takes turns with a patient in placing cups on the grid, with cups deliberately chosen because grasping such objects is a functional movement related to daily activities that rehab programs seek to restore. A second system involves playing the game with colored lights instead of a robot arm, and a study demonstrated that younger subjects preferred the light-up board to the arm. The researchers also note the robot arm system's ability to monitor patients' success rates and movement patterns is critical, with future iterations possibly tracking performance in real time to modify game parameters.

Full Article
Computer Security Flaws Reveal Faster Isn't Necessarily Better
CORDIS News
May 4, 2018


Researchers at Graz University of Technology in Austria working on the European Union-funded SOPHIA project made up one of four research groups that discovered the Meltdown and Spectre security flaws. Although these vulnerabilities were only recently discovered, they have existed since the 1990s, highlighting the idea that computer security should be prioritized over speed. Both bugs enable intruders to access unauthorized information, but in different ways. Meltdown overcomes memory isolation, which enables malicious programs to read ordinarily non-accessible parts of a computer's memory. The Spectre attack deceives the processor into speculatively executing instruction sequences that should not be conducted during correct program execution, which results in confidential information being leaked. "In today's environment of increasing attacks on computer systems, we need to accept security as a major design criterion," says SOPHIA's Stephan Mangard. "I hope the discovery of the Meltdown and Spectre flaws will trigger a new way of thinking about computer design."

Full Article
MLPerf--Will New Machine Learning Benchmark Help Propel AI Forward?
HPCwire
John Russell
May 2, 2018


An academic-industrial partnership, including Google, Baidu, AMD, and Harvard and Stanford universities, among others, last week released a benchmarking tool "for measuring the speed of machine-learning (ML) software and hardware." Those developers say the MLPerf tool is aimed at expediting ML progress via fair and useful measurement, as well as facilitating fair comparison of competing systems while encouraging innovation to enhance cutting-edge ML. The effort's objectives also include maintaining the affordability of benchmarking to support universal participation, serving both commercial and research communities, and guaranteeing reliable results via enforced replicability. "This benchmark appears to be written for bounded problems that predominate today in early AI (artificial intelligence)," says Hyperion Research's Steve Conway. "Later on we are going to need additional benchmarks as AI starts getting into unbounded problems that will be the most economically important problems." There currently are reference deployments for each of the seven benchmarks in the MLPerf suite, including image classification, object detection, speech recognition, translation, recommendation, sentiment analysis, and reinforcement.

Full Article
Data Science Team Receives ACM Software Prize
Berkeley News
Robert Sanders
May 2, 2018


ACM is recognizing an international team of researchers with the 2018 Software System Award for Project Jupyter, led by Fernando Perez at the University of California, Berkeley, which yielded a tool for sharing and improving computer code, documents, and data visualizations. Jupyter evolved out of IPython, an interactive add-on to the Python programming language that Perez created in 2001. More than 2 million Jupyter Notebooks are currently hosted on GitHub, covering technical documentation, course materials, books, and academic publications. In addition, Jupyter Notebooks are employed by major technology companies such as Microsoft, Google, and IBM, which have created related hosted services. Jupyter also played a role in the research that discovered gravitational waves, as the Laser Interferometer Gravitational Observatory (LIGO) publishes Jupyter Notebooks so anyone can reproduce their original analyses of the mechanics underlying these ripples in spacetime. Perez notes Project Jupyter continues to develop tools for "human-computer interplay for scientific exploration and data analysis."

Full Article

Microscopy image of an electronic device made with 1D ZrTe3 nanoribbons One-Dimensional Material Packs a Powerful Punch for Next-Generation Electronics
UCR Today
Holly Ober
May 1, 2018


University of California, Riverside (UCR) researchers have developed prototype devices made of zirconium tritelluride (ZrTe3) nanoribbons, which can conduct a current density 50 times greater than conventional copper interconnect technology. To accommodate the demand for ever smaller and faster devices, the electronics industry needs alternatives to silicon and copper that can sustain extremely high current densities at sizes of only a few nanometers. The UCR breakthrough pushes the field from two-dimensional to one-dimensional (1D) materials, an important advance for future generations of electronics. The researchers attribute the exceptionally high current density in ZrTe3 to the single-crystal nature of quasi-1D materials. They say in theory, these types of materials could be grown directly into nanowires with a cross-section that corresponds to an individual atomic thread.

Full Article
Could Robots Be Counselors? Early Research Shows Positive User Experience
University of Plymouth (United Kingdom)
Amy McSweeny
May 3, 2018


Researchers at the University of Plymouth in the U.K. have demonstrated that a humanoid social robot can deliver a positive motivational interview, encouraging participants who want to be more physically active to voice their goals and challenges. The researchers programmed a NAO robot with a script designed to provoke ideas and dialogue on how someone could accomplish their objectives. Plymouth professor Jackie Andrade believes that because robots are viewed as non-judgmental, they offer greater benefit than more humanoid avatars for delivering virtual support for behavioral alterations. "Concern about being judged by a human interviewer came across strongly in praise for the non-judgmental nature of the robot, suggesting that robots may be particularly helpful for eliciting talk about sensitive issues," she notes. "The next stage is to undertake a quantitative study, where we can measure whether participants felt that the intervention actually increased their activity levels."

Full Article
Researchers' Attack on Data Privacy System Shows Noise Leaks the Very Data It Is Trying to Protect
Tech Xplore
Lisa Zyga
May 1, 2018


Imperial College London researchers have demonstrated a hack on a new data privacy system called Diffix, showing they could expose an individual's private attributes with up to 99 percent accuracy by using only five such attributes and asking Diffix 10 carefully chosen questions. Diffix's privacy protection mainly stems from the injection of noise, and the researchers leveraged this noise in their hack to deduce the private information of individuals in the database. They deployed an intersection attack by posing two queries that requested the number of individuals in the dataset who met certain conditions. The queries differed by only one condition, so that by calculating the difference between the results they could neutralize part of the noise. The team was able to obtain five pairs of these results, then performed a likelihood ratio test to ascertain specific information about an individual.

Full Article
A Surprising New Superconductor
Cooperative Institute for Research in Environmental Sciences
May 1, 2018


Don David at the Cooperative Institute for Research in Environmental Sciences (CIRES) in Boulder, CO, along with colleagues at the U.S. National Institute of Standards and Technology (NIST), have described a process for producing an ultrathin layer of rhenium sandwiched between gold layers, each measuring 1/1000th the diameter of a human hair, that can superconduct at critical temperatures higher than 6 degrees Kelvin (-449 degrees Fahrenheit). The electroplated rhenium is ideal for use in circuit boards for ultrafast, next-generation computing applications: superconducting at higher, easier-to-achieve critical temperatures; easy to work with mechanically; non-toxic, and having a high melting point. This follows last year's discovery of the combination by CIRES and NIST researchers. The achievement represents a major step forward in the creation of the high-performance, superconducting computers of the future.

Full Article

Pure iron ore on a wooden plank Mining for Gold With a Computer
Texas A&M Engineering News
Elizabeth Thomson
May 1, 2018


Engineers from Texas A&M University and Virginia Polytechnic Institute and State University have used image-analysis software to gain new insights into nanoporous gold (NPG) by mining existing information on the material. NPG has applications in a growing number of areas, including energy storage and biomedical devices. Using 150 peer-reviewed papers, the software analyzed photographs of NPG, measuring key features that were correlated with written descriptions of how the samples were prepared. This helped the team learn how to make NPG with specific characteristics by changing processing times and temperatures. The team also identified a new parameter related to NPG that could be used to adjust the material for specific applications. Due to NPG's complex structure, measuring its features has been extremely difficult, but the new software simplifies the task. The researchers were able to look at about 80 data points, while earlier attempts relied on very small data sets of five or six data points.

Full Article
Research Removes Obstacle to Lasers and Computers Working at Speed of Light
The Newsstand (SC)
Hannah Halusker
April 30, 2018


Clemson University researchers have developed an optical diode that enables light to move in one direction, a breakthrough that could be used in high-energy lasers and in optical computers capable of processing data at the speed of light. Conventional computers use the movement of electrons to process data--but the speed at which computers can perform calculations and process information is limited by how quickly electrons can navigate within solid devices such as diodes and transistors. The Clemson optical diode solves this problem by placing two different nanomaterials with different optical properties side by side, one a saturable absorber and the other a reverse saturable absorber. The researchers used hydrofluoric acid to selectively pull aluminum atoms out of a titanium-aluminum-carbon mixture, which resulted in a two-dimensional titanium carbide-dubbed MXene that can withstand the high energies of light and be manufactured into different thicknesses, adding durability.

Full Article

White boat sailing in the open blue sea Computer Scientists Find the Longest Straight Line You Could Sail Without Hitting Land
Technology Review
April 30, 2018


Researchers at the United Technologies Research Center in Ireland and IBM Research India have developed an algorithm for calculating the longest straight-line path around the world on land or sea. The algorithm considers potential solutions as branches on a tree, checking one branch after another. Each branch contains a subset of potential solutions, of which one is the optimal solution. Another technique, called "bounding," measures the properties of the subsets to determine whether the solution is closer to the optimal value. When it finds a closer solution, the algorithm considers this the best subset until a better subset on a better branch is identified. The researchers used this method to determine that the longest straight-line path over water stretches 32,089.7 kilometers from Sonmiani, Balochistan, Pakistan to the Karaginsky District, Kamchatka Krai, in Russia. The longest path over land runs 11,241.1 kilometers from Jinjiang, Fujian in China to Sagres in Portugal.

Full Article
The Continuing Arms Race: Code-Reuse Attacks and Defenses
 
ACM Publications
 

Association for Computing Machinery

2 Penn Plaza, Suite 701
New York, NY 10121-0701
1-800-342-6626
(U.S./Canada)



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]