ACM TechNews

Welcome to the March 19, 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."

research team Origami-Inspired 'Inspector Gadget Robotic Arm'
Seoul National University
March 16, 2018

Researchers at Seoul National University's Soft Robotics Research Center in South Korea have developed a foldable origami-inspired robotic arm that can self-fold while also being highly rigid. The machine makes it possible to change the shape with a single wire, thus implying practical application of the origami structure. The lightweight arm can fold flat and extend in the manner of an automatic umbrella and even becomes instantly stiff. A collapsible locker enables the device to withstand external forces and be easily actuated, while perpendicular folding drives the variable stiffness mechanism. In addition, the lockers can be easily unlocked and the structure is folded flat by pulling a single wire with a small force. The arm's advantages can be maximized when it is attached to drones where the weight and the size constraints are the most extreme. The proposed variable stiffness mechanism is applicable to other types of robots and structures in extreme environments.

Full Article
How Accurate Is Your AI?
Kyoto University
March 16, 2018

A researcher at Kyoto University in Japan has developed a new technique that evaluates artificial intelligence's (AI) performance based solely on the input data. In typical AI development, a performance evaluation is trusted if there is an equal number of positive and negative results, and data biased toward either value means the current system of evaluation will distort the system's ability. "The novelty of this technique is that it doesn't depend on any one type of AI technology, such as deep learning," says Kyoto's J.B. Brown. "It can help develop new evaluation metrics by looking at how a metric interplays with the balance in predicted data. We can then tell if the resulting metrics could be biased." Brown's work breaks down the AI utilization and analyzes the nature of the statistics used for reporting an AI's ability, while also producing a probability of the performance level, given evaluation data.

Full Article

beams emitted along symetry axis, illustration Supercomputer Simulation Opens Prospects for Obtaining Ultra-Dense Electron-Positron Plasmas
Denis Bedoya
March 16, 2018

Researchers from the Russian Academy of Sciences, the Chalmers University of Technology in Sweden, and Lobachevsky University in Russia have developed a software tool called PICADOR for numerical modeling of laser plasmas on modern supercomputers. PICADOR is a parallel deployment of the particle-in-cell technique that has been optimized for modern heterogeneous cluster systems. The Lobachevsky team has formulated the conditions under which the avalanche-like production of electrons and positrons in the focus of a high-power laser pulse generates electron-positron plasma of record density, making it possible to understand processes occurring in astrophysical objects and to analyze elementary particle production processes. Large-scale numerical modeling of the electron-positron avalanche development in a tightly focused laser field demonstrates the quasistationary states of dense electron-positron plasma. With a total number of particles of the order of 1,011, the density surpasses the value of 1,026 particles for each cubic centimeter, limited only by the resolution of numerical simulation.

Full Article
Code.Org Is Giving Kids a Chance to Code By Bringing Computer Science to Schools
Fast Company
Aaron Hurst
March 15, 2018 founder Hadi Partovi says his nonprofit has partnered with 180 of the largest U.S. school districts to add computer science to the curriculum. These districts teach nearly 10 percent of all U.S. students and 15 percent of Hispanic and African-American students. Hartovi notes more than 800,000 teachers worldwide have enrolled to teach the introductory CS courses on's platform. One of Hartovi's chief goals in launching has been to get more women and minorities to study CS, and measures diversity as a key metric across its programs.'s annual 2017 report found 45 percent of the students in its CS classrooms are female and 48 percent are from underrepresented minorities. "If only 1 percent of these girls continue to study CS in university, that would outnumber the gender gap that exists today," Hartovi says.

Full Article
Zero Field Switching (ZFS) Effect in a Nanomagnetic Device
National Institute of Standards and Technology
March 16, 2018

Researchers at Johns Hopkins University and the U.S. National Institute of Standards and Technology have discovered an unexpected zero field switching (ZFS) effect in a nanomagnetic device that could lead to smaller, lower-power memory and computing devices. The team found they could flip cobalt-iron-boron (CoFeB) magnetization in a stable manner between 0 and 1 states by transmitting only electric "spin" current through metallic layers of platinum and tungsten adjacent to the nanomagnet, without a magnetic field. In the absence of the spin current, the CoFeB magnetization is stable against any fluctuations in current and temperature. This ZFS effect presents new issues to theorists about the underlying mechanism of the observed spin-orbit torque-induced switching phenomenon. The researchers are currently investigating a way to identify other prospective materials that enable ZFS of a single perpendicular nanomagnet, determine how the ZFS behavior changes for nanomagnets with smaller lateral sizes, and develop the theoretical platform for this phenomenon.

Full Article

Richard Brooks and Lu Yu Can Bitcoin Technology Improve Research Data Integrity? This Clemson Professor Thinks So
Greenville News (South Carolina)
Liv Osby
March 16, 2018

Researchers at Clemson University have developed a technique to secure raw data using blockchain. "The chain is built by everyone signing their own data, using cryptography, to prove the data comes from them," says Clemson professor Richard Brooks. "Then all this information is interweaved together into blocks and that to modify one of the things becomes very difficult and infeasible. This provides an audit trail so it can be verified that nobody's modified the data." Brooks' team worked with drug trial data, based on an earlier project that used blockchain to guarantee academic integrity by making data forgery tougher. Brooks believes in addition to providing the data trail, the new system could potentially provide other researchers with access to all trial data. Another potential advantage is helping scientists reproduce study results, which Brooks says "would be one way, if it's registered, to have a fuller record of what was done."

Full Article
Unlocking On-Package Memory's Effects on High-Performance Computing's Scientific Kernels
March 15, 2018

Researchers from the Pacific Northwest National Laboratory, Virginia Tech University, and the University of Cophenhagen in Denmark studied the impact of on-package memory (OPM) on the performance and power efficiency of important high-performance computing (HPC) scientific kernels. The study characterized and analyzed modern OPM storage to establish guidelines on tuning the memory to speed up HPC applications. The researchers tested different tuning modes of OPM and how they impacted application tuning for optimal system performance. This study allowed the team to derive an intuitive visual analytical model for complex architectural scenarios, and create a framework for future architecture optimizations and efficiency tuning. The work is expected to help advance computing systems, for example, by motivating software-architecture co-design exploration and validating models and simulations.

Full Article
FPGA Programming Made Easy
Inside HPC
March 15, 2018

Field-programmable gate arrays (FPGAs) are increasingly popular for applications that require high performance, low latency, and power efficiency. Although FPGAs can be re-configured for a specific application, traditionally it has been necessary to understand a complex programming language such as Verilog or VHDL that was designed for a specific FPGA. However, Open Computing Language (OpenCL) is an application program interface (API) for the programming of a diverse set of processors, and it is a royalty free and open standard used for a wide range of accelerators. Developers can use a familiar language such as OpenCL to become more productive sooner when deciding to use an FPGA for a specific purpose. When developing an application for an FPGA that uses OpenCL, there are different ways to create a high-performing application--emulation, offline compile, and cloud. OpenCL also includes an API for the host to communicate with the device, or for one kernel to communicate with another without host interaction.

Full Article

The PhenoCam network PhenoCam Network Harnesses 'Big Data' to Predict Impact of Warmer Climate on Ecosystem Productivity and Carbon Cycling
Northern Arizona University
Kerry Bennett; Diane Hope
March 13, 2018

Northern Arizona University professor Andrew Richardson has developed a vast network of digital cameras to capture millions of images of seasonal vegetation changes across North America. The PhenoCam network is the result of a 10-year collaboration between Richardson and scientists from the University of New Hampshire and Boston University to observe phenological phenomena on a continental scale. The team needed a big data solution to handle approximately 15 million images collected through the end of 2015, representing 750 years' worth of data that requires 6 terabytes of disk space. PhenoCam offers a permanent record that enables the study of the phenological state of the vegetation surveyed at any point in time. The network can also be used for evaluation of satellite remote-sensing data products, benchmarking earth system models, and studies of climate change impacts on terrestrial ecosystems. The team is also using the data to refine continental-scale climate forecasts.

Full Article

Agricultural and biological engineering professor Girish Chowdhary with TerraSentia, a crop phenotyping robot. Ag Robot Speeds Data Collection, Analyses of Crops as They Grow
Illinois News Bureau
Sharita Forrest
March 12, 2018

Scientists at the University of Illinois have developed a lightweight, low-cost agricultural robot that could improve data collection and field scouting for agronomists, seed companies, and farmers. The TerraSentia crop phenotyping robot moves autonomously between crop rows to measure the traits of individual plants using a variety of sensors. The robot sends the data in real-time to the operator's phone or laptop computer. Using virtual reality and GPS, the operator can steer TerraSentia, which is customizable and teachable. The researchers are developing machine learning algorithms to train the robot to identify common diseases and to measure various traits, such as plant height. By automating data collection and analytics, the robot has the potential to improve the breeding pipeline by revealing why plant varieties respond differently to environmental conditions, says University of Illinois professor Carl Bernacchi.

Full Article

The LGBQ flag STEM is Losing Male LGBQ Undergrads
Science Magazine
Katie Langin
March 14, 2018

Montana State University professor Bryce Hughes recently conducted a study providing quantitative evidence that science, technology, engineering, and math (STEM) fields have a problem retaining men who identify as part of the LGBQ community, while LGBQ women are actually more likely to persist in STEM fields than their heterosexual peers. The study examined a 2015 survey of 4,162 college seniors at 78 U.S. institutions, about 8 percent of whom identified as LGBQ. Hughes says all of the students had declared an intention to stay in STEM, but only 71 percent of heterosexual students and 64 percent of LGBQ students remained in the field. Deeper investigation by Hughes determined heterosexual men were 17 percent more likely to stay in STEM than their LGBQ male counterparts. However, the reverse was true for women, as LGBQ women were 18 percent more likely than heterosexual women to remain in STEM.

Full Article
A Game Changer: Metagenomic Clustering Powered by HPC
Berkeley Lab News Center
Linda Vu
March 12, 2018

Researchers from the Lawrence Berkeley National Laboratory and Joint Genome Institute (JGI) have modified the Markov Clustering (MCL) algorithm to operate rapidly, efficiently, and at scale on distributed-memory supercomputers. Testing showed their HipMCL algorithm successfully clustered a large biological network containing about 70 million nodes and 68 billion edges in a few hours, using approximately 140,000 processor cores on the National Energy Research Scientific Computing Center's Cori supercomputer. "The real benefit of HipMCL is its ability to cluster massive biological networks that were impossible to cluster with the existing MCL software, thus allowing us to identify and characterize the novel functional space present in the microbial communities," says JGI's Nikos Kyrpides. "Moreover we can do that without sacrificing any of the sensitivity or accuracy of the original method, which is always the biggest challenge in these sort of scaling efforts."

Full Article
Field-Data Study Finds No Evidence of Racial Bias in Predictive Policing
Indiana University
March 12, 2018

Indiana University–Purdue University Indianapolis (IUPUI) research suggests predictive policing does not lead police to make discriminatory arrests. George Mohler, associate professor of computer and information science in the School of Science at IUPUI, worked with researchers at the University of California-Los Angeles and Louisiana State University to conduct the study in conjunction with the Los Angeles Police Department. Working with real-time field data, both a human analyst and an algorithm made predictions on where officers would patrol each day. A random determination decided which set was used daily by officers, and the researchers measured arrest rates of ethnic groups. "When we looked at the data, the differences in arrest rates by ethnic group between predictive policing and standard patrol practices were not statistically significant," Mohler says. However, predictive policing is a nascent field and Mohler says police departments should continue to monitor the ethnic impact of algorithms to check for racial bias.

Full Article
Shared-Memory Parallelism can be Simple, Fast, and Scalable
ACM Discounts

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.