Welcome to the August 8, 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."
|
|
|
New Tool Helps Users Control Which Countries Their Internet Traffic Goes Through
Princeton University
Molly Sharlach
August 2, 2018
Princeton University researchers have developed a tool that enables users to redirect their Internet traffic to avoid passing through a particular country by diverting traffic through intermediate points. The team began by assessing several countries' efforts to reduce their reliance on U.S. networks for routing their Internet traffic, looking at traffic to the 100 most popular websites in several nations. More than half of all routing paths originating in other countries passed through the U.S., with Brazil having the greatest dependence, with 84 percent of traffic traveling through the U.S. To create the tool, called Region-Aware Networking (RAN), the researchers established a network of relays using machines in 10 countries and a mechanism to forward Internet traffic through the relays. RAN was found to be more effective in bypassing some countries than others, since many popular websites are hosted only on servers in the U.S. or Europe. By decreasing international routing, tools like RAN could help avoid surveillance and censorship, boost connection speeds, and reduce costs.
|
AI May Put Private Data at Risk
Cornell Chronicle (NY)
Melanie Lefkowitz
August 2, 2018
Cornell Tech researchers have determined current models of machine learning are vulnerable to privacy leaks and other attacks. Cornell Tech's Vitaly Shmatikov has developed models that can tell whether a certain piece of data was used to train a machine learning system, with more than 90% accuracy. That knowledge could potentially be used to leak sensitive genetic or medical information, as well as detailed behavioral and location data. Shmatikov says tools that enable people to ascertain if a record was used to train an algorithm can help them determine if their data was misused. He and colleagues examined cloud services from Google and Amazon, which help customers build machine learning models from their own information. The team built "shadow models" from true or false data that accurately identified the records used to construct them, indicating that customers who use these services can easily expose their own training data.
|
It's Harder to Turn Off a Robot When It's Begging for Its Life
The Verge
James Vincent
August 2, 2018
Researchers at the University of Duisburg-Essen in Germany have found that people are very susceptible to social cues from machines, to the point that they will refrain from shutting off a robot that begs for its life. In their experiments, volunteers were instructed to complete tasks using the small humanoid Nao robot to improve its learning algorithms. After the tasks were completed, the actual test assessed how volunteers reacted when they were asked to turn Nao off, with the machine programmed to protest and even plead not to be turned off in about half of the sessions. Thirteen of the 43 volunteers who heard Nao's pleas refused to turn it off, while the remaining 30 took twice as long to comply, on average, compared to those who did not hear the pleas. A surprising finding was that Nao's exhibition of social behavior before it asked not to be deactivated had little impact on whether participants acquiesced. The researchers said, “Triggered by the objection, people tend to treat the robot rather as a real person than just a machine by following or at least considering to follow its request to stay switched on.”
|
The Smarter City: Las Vegas Tunes IT Operations With AIOps
Government Computer News
Sara Friedman
August 3, 2018
The City of Las Vegas is enhancing its information technology (IT) operations by applying machine learning and advanced analytics to data from its IT systems using artificial intelligence for IT operations (AIOps). Market research firm Gartner says AIOps platforms "combine big data and AI or machine learning functionality to enhance and partially replace a broad range of IT operations processes and tasks, including availability and performance monitoring, event correlation and analysis, IT service management, and automation." Las Vegas is using FixStream's AIOps platform to monitor critical business systems and detect IT problems before they induce downtime. The platform functions in both traditional IT environments and with the city's cloud-based help-desk platform. City official Michael Sherwood says he envisions broadening the use of AIOps to include network infrastructure, as well as smart city elements such as traffic signals.
|
Beyond 3D Printers: Additive Folding Creates Artistic Animatronics
ZDNet
Greg Nichols
August 2, 2018
Researchers from the Korea Institute of Science and Technology, the Massachusetts Institute of Technology, the University of Pennsylvania, and the University of York in the U.K. are collaborating on the creation of soft and actuated three-dimensional (3D) objects from printable two-dimensional (2D) materials that are folded. The researchers say their additive folding technique involves two-dimensional slices that are "chain-connected, accordion-folded, and stacked up to become animatronic robots of desired 3D shapes." The team developed algorithms for modeling and designing 2D fold patterns from 3D computer simulations, which then segmented the 3D objects into 2D shapes that can be folded in layers. Using a paper-crafting machine, the final robot is then fabricated from folded strips of polyester film held together by nylon strings that are pulled by motors to animate the robot.
|
Amazon Makes $525,000 Donation to Canada Learning Code
IT World Canada
Meagan Simpson
August 1, 2018
Amazon says it will donate $525,000 to Canada Learning Code to improve and develop digital literacy programs across the country. The donation “will have a transformative effect on Canadian skills development needed to prosper in the digital economy,” according to Canada Learning Code's Melissa Sariffodeen. The donation will be directed to Canada Learning Code's computer science and coding programs, and will support both a traveling mobile classroom and Canada Learning Code Week. In addition, the organization expects Amazon's donation will help create 275,000 learner experiences over the next few years. Canada Learning Code helps disadvantaged Canadians from all groups, genders, and generations develop technology skills. Said Amazon’s Alexandre Gagnon, “This donation will help provide students the skills they need to excel in one of Canada’s fastest growing job sectors.”
|
Roboticist at OSU Uses Performing Arts to Make Robots More Social
Corvallis Gazette-Times (OR)
Lillian Schrock
July 30, 2018
Oregon State University's (OSU) Heather Knight, director of the Collaborative Humans and Robots: Interaction, Sociability, Machine Learning, and Art (CHARISMA) Research Laboratory, is applying performing arts principles to the creation of more socially intelligent robots. Said Knight, "To make functional machines, you need to think about how to get them to do their job, but also how to add value to that shared experience." One of the earliest ways she tried to reinforce social relationships between robots and humans was by having a robot perform stand-up comedy. More recently, Knight and her students programmed a remote-controlled robot to function as a health coach by visiting OSU professors and students to inquire how they were feeling. The robot would then count out exercises for them to perform, and offered them healthy snacks. Knight said the idea that robots can be socially intelligent is “exciting because we get to be paving a path toward that kind of future.”
|
Research Identifies Key Weakness in Modern Computer-Vision Systems
Brown University
Kevin Stacey
July 30, 2018
Brown University researchers have demonstrated a flaw in modern computer vision algorithms that makes them consistently poor at differentiating between two objects in an image. Brown's Thomas Serre and colleagues used cutting-edge computer vision algorithms to analyze monochrome images containing two or more randomly produced shapes to identify a same-or-different relationship. The researchers determined the algorithms' recognition of this relationship did not improve after many training examples, but by removing object-individuation tasks from the computer, the algorithms could easily learn this relationship, provided they did not have to view the two objects in the same image. Serre says the object-individuation flaw is embedded in the framework of convolutional neural networks driving the algorithms, which are exclusively fed information in one direction, unlike the workings of the human brain's virtual systems. Serre suggests making computer vision smarter will require neural networks that can better approximate the recurrent nature of human visual processing.
|
Five Questions With Will Sentance
The New York Times
Olivier De Meulder
August 2, 2018
In an interview, Codesmith's Will Sentance discussed how he started the organization and why JavaScript is a useful programming language. Codesmith is an organization that teaches software engineering, computer science, and machine learning in New York and Los Angeles, as well as via an online platform to serve students around the world. Sentance says he started Codesmith to give people a chance to develop technology skills in a setting that would bring together people from every background, empower them to become autonomous individuals, “and accelerate them to leadership in software engineering and technology.” Sentance says Codesmith relies on pair-programming, which works by separating the responsibilities of solving a coding challenge between two people, with one partner verbalizing an approach to the challenge and the other implementing the actual code. He adds, “Teaching students to explain their approach to a pair-programming partner improves the quality of technical communication.”
|
A New Complex Network-Based Approach to Topic Modeling
Tech Xplore
Ingrid Fadelli
July 30, 2018
Researchers at the University of Bath in the U.K., the University of Sydney in Australia, and Northwestern University have developed a novel network approach to topic models, machine learning strategies that can find abstract topics and semantic structures within text documents. The most commonly used method for topic models, latent Dirichlet allocation, has several flaws in the way it represents text, such as the inability to choose the number of topics, discrepancies with statistical properties of real texts, and a lack of justification for the Bayesian prior. The new approach to topic modeling, which was found to be more versatile and principled than other topic models, is based on the method used to find communities in complex networks. Northwestern University's Martin Gerlach says the work shows "how to formally relate methods from community detection and topic modeling, opening the possibility of cross-fertilization between these two fields."
|
On-Chip Optical Filter Processes Wide Range of Light Wavelengths
MIT News
Rob Matheson
August 1, 2018
Massachusetts Institute of Technology (MIT) researchers have developed an optical filter on a chip that can process optical signals from across an extremely wide spectrum of light at once, which could offer greater precision and flexibility for designing optical communication and sensor systems, studying photons and other particles through ultrafast techniques, and other applications. The filter matches the broadband coverage and precision of conventional filters, but can be manufactured using traditional silicon-chip fabrication methods. The new chip uses waveguides to precisely guide the light input to the corresponding signal outputs. One section of the filter contains an array of three waveguides, while the other section contains one waveguide that is slightly wider than any of the other three. The work could lead to more effective "optical combs," which measure light emitted or reflected by objects for various applications, including powering "optical clocks" for GPS satellites.
|
'Blurred Face' News Anonymity Gets an AI Spin
Simon Fraser University
August 1, 2018
Researchers at Canada's Simon Fraser University (SFU) School of Interactive Arts and Technology (SIAT) have developed a method of employing artificial intelligence (AI) to better obscure the faces of anonymous individuals in news reports. SIAT's Steve DiPaola says the method involves an updated model of pixelization incorporating “more than 1,000 years of artistic technique,” which is used to teach the AI "to lower the outer resemblance and keep as high as possible the subject's inner resemblance—in other words, what they are conveying and how they are feeling." The system utilizes five levels of AI processing to simulate a painter that uses art abstraction to "repaint" each frame of video. DiPaola says the system’s ability to anonymize faces “while retaining a strong degree of emotional connection or resonance should result in better final product for anonymized video.”
|
|
