Welcome to the October 29, 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."

fiber optics, illustration RMIT Touts Faster Internet Thanks to 'Twisted Light'
Asha McLean
October 25, 2018

Researchers at the Royal Melbourne Institute of Technology (RMIT) in Australia have developed a nanophotonic device that encodes more data and processes it much faster than current fiber optics by harnessing 'twisted' light. The technology carries data on light waves that have been twisted into a state of orbital angular momentum (OAM). Haoran Ren at RMIT's School of Science said, "Our miniature OAM nano-electronic detector is designed to separate different OAM light states in a continuous order and to decode the information carried by twisted light." Two years ago, RMIT's Laboratory of Artificial Intelligence Nanophotonics described how it decoded a small range of twisted light on a nanophotonic chip, but at the time, technology to detect a wide range of OAM light for optical communications was not viable. The materials used in the latest nanophotonic device are compatible with silicon-based materials use in most technology, making it easier to scale up.

Full Article
How Walmart's Sam's Club Is Following Amazon Into Cashierless Stores
David Meyer
October 29, 2018

Walmart's Sam's Club is taking a cue from Amazon and testing new digital automated-store solutions at its new cashierless Dallas outlet, Sam's Club Now. The company says the facility is a "technology lab that doubles as a live, retail club," designed to trial technologies mainly based on the Scan & Go smartphone app prior to national launch. Scan & Go lets people scan items they place in their cart, and pay for them via smartphone. At Sam's Club Now, shoppers can use voice search to view a map of the store displaying item locations on their smartphones. The app will apply machine learning to compose people's shopping lists for them, and eventually integrate with the mapping feature to plot out shoppers' ideal store routes. Sam’s Club CEO Jamie Iannone said that in addition to the new app, "We’ll test electronic shelf labels that will instantly update prices, removing the need to print and replace new item price signs."

Full Article
Japan Preschools Using Tablets to Prep Tots for Digital Age
Associated Press
Yuri Kageyama
October 27, 2018

Almost 400 kindergartens and nursery schools in Japan are employing smartphone software applications in conjunction with tablet computers to enable "digital play" among preschoolers. The KitS apps allow children to color animated birds and flowers as three-dimensional computer graphics; users also draw various creatures that, when captured as computer images, swim or float within virtual landscapes. The goal is to foster creativity, focus, and leadership skills seen as essential to the digital era. The University of Tokyo's Yuhei Yamauchi envisions practical advantages to this approach, because by the time today's preschoolers enter the workforce, computer skills are expected to be basic professional requirements. Moreover, Yamauchi says in view of Japan's shrinking populace, people may work into their 80s and change jobs several times, which will make digital skills increasingly vital.

Full Article

Sierra supercomputer at Lawrence Livermore National Laboratory Meet Sierra: Livermore's Powerful New Supercomputer
The Mercury News
Lisa M. Krieger
October 27, 2018

The Lawrence Livermore National Laboratory's new Sierra supercomputer can perform 125 quadrillion calculations each second, with the goal of fortifying the aging U.S. nuclear stockpile. Assembled by IBM and Nvidia, Sierra will support Livermore and the two other U.S. nuclear security labs, the Sandia and Los Alamos national laboratories. Livermore's Fred Streitz says the system "enables simulations 100,000 times more realistic than is possible on a desktop." Sierra's power is derived from its "heterogenous" supercomputer architecture, which makes it possible to transfer data between a high-speed connection that connects its central processing units and graphics processing units. Sierra has a forecast top-performing lifespan of five to six years; after that, experts expect exascale supercomputing that can process more than 1 quintillion calculations a second will render systems like Sierra obsolete.

Full Article
*May Require Paid Registration
New Tool Streamlines the Creation of Moving Pictures
Princeton University
Molly Sharlach
October 24, 2018

Princeton University researchers have developed a tool designed to simplify image animation. Users manually choose a subset of repeating objects, then draw motion lines and specify the frequency and speed at which the objects should move; an algorithm extracts similar objects in the image and partitions them into their own layer for animation. Said Princeton's Adam Finkelstein, "The person provides clues about what aspects of the scene they would like to animate, and the computer removes much of the difficulty and tedium that would be required to create the animation completely by hand." The tool was presented at the ACM Symposium on User Interface Software and Technology (UIST 2018) in Berlin.

Full Article

fruitflies Building a Fly Brain in a Computer
Canadian Institute for Advanced Research
Jon Farrow
October 25, 2018

Researchers at Guelph University and the University of Toronto Mississauga in Canada have built a neural network that mimics the fruit fly's visual system, and can distinguish and re-identify flies. This development could enable thousands of laboratories that use fruit flies as a model organism to conduct more longitudinal work. The researchers built their algorithm by combining insect biology and machine learning to process low-resolution videos of fruit flies over two days. The algorithm reliably identified the same fly on the third day with an F1 score of 0.75, which is only slightly lower than scores for algorithms without the limitations of fly-brain biology, while outperforming human biologists. Toronto's Joel Levine said, "The approach of pairing deep learning models with nervous systems is incredibly rich. It can tell us about the models, about how neurons communicate with each other, and it can tell us about the whole animal."

Full Article
Efforts to Acknowledge the Risks of New AI Technology
The New York Times
Cade Metz
October 22, 2018

A coalition of 46 scientists is calling on the research community to reconsider how it shares new artificial intelligence (AI) technology by specifying its potentially negative societal impacts as well as its benefits. The researchers, collectively known as the ACM Future of Computing Academy, are urging peer-reviewed journals to reject papers that do not explore such hazards, a suggestion many scientists oppose because it will not eliminate AI's potential dangers. Concerned U.S. and British scientists and policymakers published a report on weaponized AI in February, while others are building technologies to demonstrate how the technology can go awry. The Massachusetts Institute of Technology's Matt Groh, for example, constructed a system that deletes objects and people from photos, to illustrate AI's potential usage for disinformation. Another recent example is a joint Google Brain/DeepMind system that outperforms professional lip-readers, with ramifications for surveillance.

Full Article
*May Require Paid Registration

Penn St. graduate students NSF Announces $78.2 Million to Support Frontiers of Cybersecurity, Privacy Research
National Science Foundation
Linda A. McBrearty
October 24, 2018

The U.S. National Science Foundation (NSF) Secure and Trustworthy Cyberspace (SaTC) program has announced new support for a diverse, $78.2-million portfolio of more than 225 new projects in 32 states spanning a wide range of research and education topics, including artificial intelligence, cryptography, network security, privacy, and usability. A prominent recipient is the Center for Trustworthy Machine Learning (CTML), which will address challenges in cybersecurity science and engineering. The $10-million, five-year CTML award will allow the center to focus on methods to defend a trained model from adversarial inputs, rigorously grounded measures of model and training data robustness, and ways adversaries may abuse generative machine learning models and developing countermeasures for defending against such attacks. Said NSF's Jim Kurose, "Our goal is to identify fundamentally new ways to design, build, and operate secure cyber systems at both the systems and application levels, protect critical infrastructure, and motivate and educate individuals about security and privacy."

Full Article
Origami Cushions Protect Drones From Collisions
Imperial College London
Caroline Brogan
October 22, 2018

Researchers in the Aerial Robotics Laboratory of Imperial College London in the U.K. have outfitted flying drones with lightweight, impact-absorbent cushioning inspired by origami to shield them from collisions. Imperial College's Pooya Sareh says the Rotary Origami Protective System (Rotorigami) offers a way to "let miniature flying robots navigate in confined or cluttered spaces safely and efficiently." Sareh's team folded a thin, lightweight plastic sheet into simple origami pattern, and built it into a protective structure around a rotating inner frame. This structure safeguarded propellers from side impacts and helped keep the drone airborne during and after an impact. The team's comparison of the Rotorigami-equipped drones with existing designs determined the protective structure reduced the force of an impact, and helped keep drones from spinning out of control after collisions. Their next step is to design ways to ameliorate damage from collisions with the tops and bottoms of drones.

Full Article

Jason Papin and Greg Medlock A Simple Way to Solve Complex Mysteries of the Microbiome
Charlottesville Daily Progress (VA)
Ruth Serven Smith
October 25, 2018

University of Virginia researchers have developed a computational model for how six species of bacteria interact and create an ecosystem. The team identified rules of behavior and translated the findings into parameters for the model, which allowed them to simplify a complex system and discover new behaviors of a bacterium. The researchers focused on the interactions among 15 pairs of bacteria with the goal of identifying the small proteins, or metabolites, that each pair of organisms produced as they started growing. The model helped the researchers to find additional variables, and to be more confident in relationships between data. Going forward, the team aims to conduct trials on mice and determine how the community of bacteria responds to inputs such as sugar or a virus.

Full Article

U.S. flag comprised of code Tech Industry Lobby Proposes Data Privacy Laws; Critics Call Them Weak
Bernadette Tansey
October 22, 2018

The Information Technology Industry (ITI) Council has proposed a conceptual framework for congressional legislation that it says "advances the privacy rights of consumers and defines the responsibilities of companies in using personal data while continuing to enable the innovations that transform our lives." The framework would purportedly help consumers better understand how their data is collected and used, support controls over personal data use, and uphold regulations that instill accountability for violations. Consumers would have "the right to access, correct, port, delete, consent, and object to the use of personal data about themselves." Critics say the guidelines are designed to insulate technology firms from even stricter regulations that could be imposed following major corporate data breaches.

Full Article
Army Scientists Develop Computational Model to Predict Human Behavior
R&D Magazine
October 24, 2018

U.S. Army Research Laboratory (ARL) scientists have developed an analytic model to show how groups of people influence individual behavior. The model quantitatively incorporates the dynamic behavior of the group, showing the analytic solution to this kind of equation coincides with predictions of the large-scale computer simulations of group dynamics. The model is comprised of many interacting individuals that have a 'yes' or 'no' decision to make. When individuals are alone, they often cannot make up their minds, switching back and forth between the two options; but when they start talking to other individuals, the information exchange results in the numerical calculation using the computer model to find people hold their opinions for a significantly longer time. ARL's Bruce West says this research could lead to a new area of study dovetailing network science and fractional calculus.

Full Article

translation app, illustration Baidu's AI Can Do Simultaneous Translation Between Any Two Languages
IEEE Spectrum
Eliza Strickland
October 23, 2018

Researchers at Baidu in China have developed an artificial intelligence-powered tool that uses natural language processing to predict and anticipate the words a speaker is about to say a few seconds in the future, before translating those words into a different language. The new Simultaneous Translation with Anticipation and Controllable Latency (STACL) tool improves on other translation tools by predicting the verb to come in a sentence, based on all the sentences it has seen previously. The researchers trained STACL on newswire articles that appeared in multiple languages. The team demonstrated the system's ability to translate from Chinese to English—two languages with significant differences in word order. Baidu’s Liang Huang says the goal of the research is to make instant translation services more readily accessible and affordable to the general public.

Full Article
ACM Transactions on Internet of Things (TIOT)
ACM Publications

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]