Welcome to the July 19, 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).

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."
Researchers Develop New Computational Method to Make Data-Driven 3D Modeling Easier
July 18, 2017

Researchers at the National University of Defense Technology (NUDT) in China, along with researchers from Adobe Research, IIT Bombay, Simon Fraser University, and Stanford University, have developed GRASS, a generative model based on deep neural networks that enables the automatic creation of plausible, novel three-dimensional (3D) shapes. GRASS uses machine learning techniques and artificial intelligence to eliminate the time-consuming process of generating multiple 3D shapes by hand. The researchers say GRASS eventually could transform the video game, film, computer-aided design, and virtual reality industries. They say GRASS learns to code an arbitrarily complex 3D shape into a fixed set of parameters, and to regenerate it from those parameters. "Our work is a data-driven automatic shape generation computational method," says NUDT researcher Kai Xu. "Given a set of example 3D shapes, our task is to generate multiple shapes of one object class, automatically." The researchers will demonstrate GRASS at the upcoming ACM SIGGRAPH 2017 conference in Los Angeles.

Full Article
Beauty Spot or Landscape Blot? Computer Trained to Judge Scenery
The Guardian
Ian Sample
July 19, 2017

Researchers at the University of Warwick and the Alan Turing Institute in the U.K. have developed software that can differentiate scenic views from blots on the landscape, a breakthrough they say could help with decisions over what land to protect, and how to better design new towns and cities. The researchers trained the program on more than 200,000 photographs of places in Great Britain that had been rated for beauty on the Scenic-or-not website. The software then linked the ratings to features in the images and developed a list of what makes one scene beautiful compared to others. The researchers found the program identified lakes, valleys, coastlines, and mountains as beautiful, while derelict industrial sites, motorway intersections, and construction zones all scored the lowest. "We want to understand what beautiful places are composed of because there is a connection between beautiful places and people's well-being," says Warwick researcher Chanuki Seresinhe.

Full Article

A student using Minecraft coding tutorial Women and Minorities Shatter Records as Thousands Take AP Computer Science Exam
Chelsey Ballarte
July 18, 2017

The number of women taking Advanced Placement (AP) computer science exams rose by 135 percent from 2016 to 2017, while the number of underrepresented minorities taking the exam rose 170 percent over the same time period, according to the College Board. Although lack of diversity in the tech industry has become a major issue in recent years, these new figures indicate a positive trend, according to Code.org founder Hadi Partovi. "This is a problem I think our country has been grappling with, and to see such strong participation by women and minority students is great." However, many schools still do not offer AP computer science as a course. Code.org has taught more than 60,000 teachers how to bring computer science into the classroom and recently expanded into more diverse, urban communities.

Full Article
The 2017 Top Programming Languages
IEEE Spectrum
Stephen Cass
July 18, 2017

Python is now the top programming language, according to IEEE Spectrum's recently released fourth interactive ranking of the leading languages. Python was ranked third last year, but this year passed C and Java to claim the top stop. To determine the most popular programming languages, IEEE enlisted the help of data journalist Nick Diakopoulos, who mined and combined 12 metrics from 10 online sources to rank 48 languages. In addition, IEEE Spectrum allows users to choose how those metrics are weighted when they are combined, resulting in a level of personalization not seen in other rankings. For the second consecutive year, no new languages entered the rankings, indicating a period of consolidation in coding as programmers digest the tools created to cater to the proliferation of cloud, mobile, and big data applications. The top four languages each received better than a 97 ranking out of a 100-point scale.

Full Article
At Cybersecurity Camps, Teen Girls Learn About Protecting Nation, Breaking Barriers
The Wall Street Journal
Leslie Brody
July 19, 2017

A growing number of cybersecurity camps are helping to prepare young women to work in this field, as women remain underrepresented in the information security workforce worldwide. Women make up 11 percent of the information security workforce, even as a shortage of 1.8 million cybersecurity professionals is expected to exist globally by 2022, according to a report from Frost & Sullivan. Both public- and private-sector experts have called for a stronger pipeline of cybersecurity professionals, and the U.S. National Security Agency and National Science Foundation, for example, are jointly sponsoring 131 "GenCyber" camps for students and teachers in kindergarten through 12th grade across the U.S. this summer, up from seven camps in 2014. Many of them are focusing on attracting women, such as a free three-week camp this month at New York University's Tandon School of Engineering, which attracted 45 teenage girls.

Full Article
Researchers Prove the Security of the Vector Stream Cipher
Kyoto University
July 17, 2017

Researchers at Kyoto University in Japan have developed the Vector Stream Cipher (VSC), which they say is the first example of a 128-bit key chaotic cipher with provable security. The researchers conducted a range of tests, including a method to evaluate the lock's randomness. "Before evaluating the security of VSC with randomness tests, we found a way to make it significantly more reliable and sensitive," says Kyoto researcher Ken Umeno. The research confirms that VSC is secure, structurally simple, and low on memory usage compared with existing technologies. These features could make the new technology useful for high-density data transmission applications such as in 5G mobile networks and 4k television broadcasts. "Chaotic ciphers have been in use for about 30 years, but before this study we had not expected to find proof of security," Umeno says. "We hope that our work will be studied widely and applied throughout our digital world."

Full Article

An evening city skyline Technion and Toronto Researchers Aim to See Night in a New Light
American Technion Society
Kevin Hattori
July 13, 2017

Researchers at the Technion-Israel Institute of Technology and the University of Toronto in Canada have developed a technology for producing a new understanding of the nighttime landscape based on the flicker of electric lights. Light emitted from all lamps connected to the electricity grid is constantly changing, but because of the high speed of this effect, humans do not sense this flickering. The researchers have developed a new way to produce information from the flicker patterns of lighted scenes. Their approach combines various fields of research, including optics, computer vision, image processing, and electrical grid engineering. The researchers developed a system that extracts information from a passive video of the desired scene. The analysis of the information obtained through this process determines how the scene would look if the light coming from some of the bulbs was changed.

Full Article
Improving Disease Prediction With Big Data Analytics
Lehigh University
Lori Friedman
July 17, 2017

Lehigh University researchers have developed an approach that uses a large dataset to demonstrate an improved disease prediction model that combines data cleaning and careful feature selection with machine learning techniques. The researchers utilized the PRO-ACT database to predict which patients would fall into the three clusters of Amyotrophic Lateral Sclerosis (ALS) progression. The researchers developed a model that cleaned up the data and improved the accuracy rate in predicting the rate of patients' ALS progression. The new method achieved a 58.3-percent accuracy rate, compared to 40.5 percent for the winning team in the 2012 Prize4Life competition, which asked teams to develop a system that could predict how long ALS patients would survive from the date of diagnosis. The effort was led by Lehigh professor Mooi Choo Chuah, technical co-chair of the IEEE/ACM Conference on Connected Health: Applications, Systems and Engineering Technologies (CHASE 2017), which is taking place this week in Philadelphia.

Full Article

Row of old books Fused Imaging Reveals Sixth-Century Writing Hidden Inside Bookbinding
Northwestern University Newscenter
Amanda Morris
July 17, 2017

Northwestern University researchers say they have developed a new, non-destructive technology that reveals medieval texts hidden inside of ancient bookbindings. Between the 15th and 18th centuries, bookbinders reused the bindings from medieval parchments to create new binding materials for printed books. The Northwestern breakthrough provides a means to read these texts for the first time in centuries. The researchers originally used a visible light hyperspectral imaging technique to view the writing, but it yielded poor results due to the parchment's irregular degradation. They then tried x-ray fluorescence imaging using a portable instrument, but the text was still unreadable due to poor spatial resolutions. Finally, the researchers used a machine learning algorithm to determine that a fusion of multiple imaging techniques would produce the best results. The team combined visible hyperspectral imaging with x-ray fluorescence imaging and were able to determine the relative contribution of each modality to produce the best image.

Full Article
AI Coach Will Train Hopeless Chatbots to Pass the Turing Test
New Scientist
Matt Reynolds
July 14, 2017

Researchers at McGill University in Canada have developed an artificial intelligence system that automatically rates how human-like a piece of chatbot-generated dialogue sounds. The researchers chose 1,000 short Twitter conversations and then enlisted human volunteers to add a response. The researchers then had several chatbots of varying abilities add their own responses. Finally, a different group of human volunteers were asked to rate all of the responses according to how human-like they were. McGill researcher Ryan Lowe trained a neural network on these human ratings, teaching it to differentiate between convincing and unconvincing responses. Following the training, the neural network was able to match the judgment of the human evaluators in just a fraction of a second. Lowe plans to open source the system so other researchers can use it to improve their own chatbots.

Full Article

3D models of household objects Helping Robots Learn to See in 3D
Duke Today
Robin A. Smith
July 14, 2017

Duke University researchers have developed technology that enables machines to make sense of three-dimensional (3D) objects in a more human-like way. For example, a robot that clears dishes off a table must be able to adapt to a variety of bowls, platters, and plates of different sizes and shapes, left scattered on a cluttered surface. The researchers' robot-perception algorithm can guess what a new object is and how it is oriented, without first looking at it from multiple angles. A robot equipped with this technology would not need to see every side of a teapot to know that it likely has a handle, a lid, and a spout, and whether it is sitting upright or off-kilter. "Overall, we make a mistake a little less than 25 percent of the time, and the best alternative makes a mistake almost half the time, so it is a big improvement," says Duke researcher Ben Burchfiel.

Full Article
Research Makes Robots Better at Following Spoken Instructions
News from Brown
Kevin Stacey
July 13, 2017

Researchers at Brown University have developed a system designed to help robots better follow spoken instructions. The research was led by Dilip Arumugam and Siddharth Karamcheti in the lab of computer science professor Stefanie Tellex. The system was developed using Mechanical Turk, Amazon's crowdsourcing marketplace, and a virtual task domain called Cleanup World. Volunteers were asked what instructions they would give the robot to get it to perform a task they had just viewed. The volunteers were given guidance as to the level of specificity their directions should have. The researchers used the volunteers' spoken instructions to train their system to understand what kinds of words are used in each level of abstraction. From there, the system learned to infer not only a desired action, but also the abstraction level of the command, which triggered its hierarchical planning algorithm to solve a task from the appropriate level.

Full Article

An early prototype of an automated screening kiosk. Automated Security Kiosk Could Alleviate Travel, Border Woes
Missouri S&T News
Greg Katski
July 11, 2017

Nathan Twyman at the Missouri University of Science and Technology has created an automated screening kiosk designed to enhance safety at airports and border crossings. The kiosk uses an algorithm of "yes" or "no" questions delivered by a computer-generated avatar to quickly assess the potential threats passengers may pose to others. The kiosk has an infrared camera that scans a subject's eye movement and pupil dilation, while a video camera captures natural reactions to feeling threatened, such as body and facial rigidity. A microphone records vocal data, listening for changes in pitch that accompany uncertainty. The screening kiosk "measures various psychophysiological responses and tries to make some sort of a risk assessment outcome," Twyman says. "It's an automated risk assessment, instead of a seat-of-your-pants risk assessment." After several field tests, Twyman's research group is in talks with the government of Singapore to implement the kiosks at one of its border crossings.

Full Article
ACM Ambassador Program

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: technews@hq.acm.org