Welcome to the November 30, 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."
|
|
|
Reproducing Paintings That Make an Impression
MIT News
Rachel Gordon
November 29, 2018
Massachusetts Institute of Technology researchers have combined deep learning and three-dimensional (3D) printing into a system that reproduces paintings authentically, irrespective of lighting conditions or placement. The researchers said the RePaint system is more than four times more accurate than state-of-the-art physical models at reproducing the precise color shades for different artworks. The team used a technique called "color-contoning," combining a 3D printer with 10 transparent inks stacked in thin layers, in conjunction with half-toning, which generates an image from many small colored dots, rather than continuous tones. The researchers then trained a deep learning model to predict the optimal stack of different inks, fed it images of paintings, and used the model to determine which hues to employ in particular areas for specific paintings. RePaint will be presented at ACM SIGGRAPH Asia 2018 next week in Tokyo.
|
The 'Smart' Baby Technology Raising Today's Children
BBC News
Nicholas Mancall-Bitel
November 29, 2018
Many millennial parents are adopting "smart" technology for their children, marking a new type of parenting that makes nurseries more akin to data centers. Such connected and data-producing products gives modern parents access to feedback from bottles, dummies, cots, prams, clothing, and more. Some of the products are meant to ease stress for new parents, while others simply automate parts of the parenting process by shifting tasks that were once intuition-based to being fully automated. As more millennial parents in developed countries decide to leave home to work, industry experts expect them increasingly to rely on digital monitoring products to stay connected to their children.
|
Inside the Futuristic Restaurant Where a Robot Has Replaced the Bartender
The Washington Post
Peter Holley
November 29, 2018
The Cyberdog wine bar in Prague, Czech Republic, is equipped with a robot bartender that serves drinks ordered via a smartphone app. Cyberdog owner Marcel Soural said the innovation was motivated by his expectation that robots will soon supplant waitstaff in bars and restaurants worldwide. The robotic bartender is programmed to select from a list of wines in a nearby cooler; after customers place orders, the robot can pour up to four glasses simultaneously, then puts them on a tray. A mechanical system in the ceiling delivers the drinks to nearby tables. Said Soural, "When [the robot bartender] has no order, he is bored and he shows that by making moves according to what he has in his programming," which include dancing.
*May Require Paid Registration
|
Brazil's Generation Z Aims for Technology Careers
ZDNet
Angelica Mari
November 28, 2018
Generation Z Brazilians want to work in roles related to the development and implementation of new technologies, with a Dell survey finding 85% of 704 polled individuals between 16 and 23 years old expressed a desire to find employment in the tech industry. Said Dell Brazil's Diego Puerta, "The study reveals the profile of connected young people who see technology as an essential part of the search of the balance between personal and professional life." About 39% of respondents cited face-to-face communication as their preferred way of working over applications like WhatsApp. Although 79% of respondents said they considered personal and corporate data security a high priority, only 38% said they are making maximum effort to shield information. Nearly all of the polled Generation Z Brazilians said they realize their social networking content can affect their future professional lives.
|
A Big Step Toward the Practical Application of 3D Holography With High-Performance Computers
R&D Magazine
November 28, 2018
Researchers at Chiba University in Japan have developed a computer that can project high-quality three-dimensional (3D) holography as a video. Chiba's Tomoyoshi Ito began working on specially designed computers for holography, called HORN (for HOlographic ReconstructioN), in 1992. The latest version, the HORN-8, which utilizes a calculation method called the "amplitude type" for adjusting the intensity of light, was recognized as the world's fastest computer for holography earlier this year. With the newly developed "phase type" HORN-8, the calculation method for adjusting the phase of light was implemented, allowing the researchers to successfully project holographic information as a 3D video with high-quality images. Said Ito, "We have been developing the high-speed computers for 3D holography by implementing the knowledge of information engineering and the technology of electrical and electronic engineering and by learning insights from computer science and optical methods."
|
Researchers Successfully Train Computers to Identify Animals in Photos
University of Wyoming
November 27, 2018
University of Wyoming (UW) researchers developed a computer model that identifies images of wild animals from camera-trap photos with high precision and efficiency. The work builds on an earlier deep learning model that analyzed 3.2 million camera images in Africa, and categorized animal images with 96.6% accuracy at a faster rate than humans. The latest model was trained on Wyoming’s Mount Moran, the site of UW's high-performance computer cluster, to classify wildlife using 3.37 million camera-trap images of 27 species obtained from five U.S. states. The team tested the model on about 375,000 animal images at a rate of approximately 2,000 images a minute on a laptop, and found it 97.6% accurate. Independent testing on 5,900 images of several species from Canada yielded 81.8% accuracy, and the model was 94% successful in removing images without any animals from a set of photos from Tanzania.
|
Amazon Opens Its Internal Machine Learning Courses to All for Free
TechCrunch
Connie Loizos
November 26, 2018
Amazon is opening the internal machine learning courses it uses to train its own engineers for the first time to people outside the company, for free. According to the company's Matt Wood, Amazon has made available 30 different courses comprising more than 45 hours of training for developers, data scientists, data platform engineers, and business professionals. Wood said every course "starts with the fundamentals, and builds on those through real-world examples and labs, allowing developers to explore machine learning through some fun problems we have had to solve at Amazon. These include predicting gift wrapping eligibility, optimizing delivery routes, or predicting entertainment award nominations using data from IMDB (the Internet Movie Database)." Wood also said the coursework helps streamline best practices, and shows trainees how to begin work on a variety of Amazon Web Services (AWS) machine learning services, such as Amazon SageMaker, AWS DeepLens, Amazon Rekognition, Amazon Lex, Amazon Polly, and Amazon Comprehend.
|
The Race to Build Megafire Prediction Tools
Bloomberg BusinessWeek
Adam Popescu
November 26, 2018
New technology is aiding the on-the-ground battle against wildfires, but increasingly dangerous weather conditions present escalating challenges. Researchers at New Mexico's Los Alamos National Laboratory have spent nearly 30 years collaborating with the U.S. Forest Service on a three-dimensional modeling program to predict wildfires' paths. The Firetec project has evolved over the years from a supercomputer into software that can guide officials in deploying people and equipment and setting controlled fires to ensure nothing in a fire-prone area is left to trigger new conflagrations. The Forest Service is allocating more of its $2.6-billion fire-suppression budget to predictive modeling that is more practical for fighting wildfires on a day-to-day basis.
|
How the U.K. Plans to Teach Computer Science to Every Child
Financial Times
Madhumita Murgia
November 26, 2018
The Raspberry Pi startup will train 40,000 British teachers to teach computer science (CS) under a government-funded project, with the ultimate goal of teaching CS to all schoolchildren in the U.K. Said the Raspberry Pi Foundation's Philip Colligan, "The combination of emphasis on teacher training, commitment to the subject, and making resources available to get teachers out of the classroom—that's an incredible cocktail of investment." The Raspberry Pi team, working with the British Computer Society and STEM Learning, the self-described “largest provider of education and careers support in science, technology, engineering and mathematics (STEM),” will create a program for primary and secondary educators that includes free educational resources, nationwide training, and certification. The partners will provide face-to-face workshops across 40 schools over the next four years, and those schools will later be paid to deliver training to neighboring schools. Colligan said this arrangement will help schools that cannot afford to send their educators for training.
*May Require Paid Registration
|
Computer Graphics Researchers Make Drone Navigation Easier
Purdue University News
Brian L. Huchel
November 27, 2018
Purdue University researchers have developed a touchscreen method to navigate and take pictures with drones. The FlyCam system is based on the concept of combining drone and camera movements so the user does not have to think about multiple controls; it relies on one- and two-finger drags across a smartphone or tablet to control the drone as it accelerates or turns and captures images. The drone moves forward or backward along the camera's axis with single or double taps to the screen. As part of the research, the team compared the ability of licensed drone pilots with a remote control with that of people who piloted a drone for the first time using FlyCam; they found the novice drone pilots using FlyCam were equal to, or demonstrated greater skill than, the licensed drone pilots.
|
Smart Headlights Inch Closer to American Roads
The New York Times
Eric A. Taub
November 23, 2018
Adaptive driving beam (A.D.B.) headlights use sensors and cameras to continuously shape a vehicle's high beams to illuminate only areas without oncoming traffic, while sending light elsewhere down the road. Car manufacturers such as Audi, BMW, Mercedes, and Toyota already offer this type of lighting, but ADB lamps currently are illegal in the U.S. The U.S. National Highway Traffic Safety Administration (NHTSA) currently requires vehicles to have distinct high and low beams, disallowing lights that can dynamically adjust. However, in October the NHTSA issued a notice of proposed rule-making that, if approved, would allow these headlamps in the U.S. In anticipation of this, Audi is already selling cars in the U.S. that feature “matrix-designed” LED headlamps, which need only a software upgrade to operate in an adaptive way.
*May Require Paid Registration
|
North Sea Deployment Shows How Quadruped Robots Can Be Commercially Useful
IEEE Spectrum
Evan Ackerman
November 27, 2018
Earlier this year, Switzerland’s ANYbotics brought one of its industrial quadruped robots to an offshore power distribution platform in the North Sea. The ANYmal autonomous mobile robot could offer support to humans in remote locations through regular and automated inspections of machinery and infrastructure. ANYmal is equipped with onboard vision and lidar that allow the robot to localize itself in its environment and plan a safe path for itself over and around obstacles. The robot’s inspection payload includes a zoom camera, thermal cameras, and microphones to provide data for detailed situational awareness. Said ANYbotics’ Peter Fankhauser, "We are still far away from replicating the human capability to understand context and manipulate objects and tools. So Instead of trying to replace human workers, our goal is to extend the workforce with automated robotic inspection."
|
Fearful of Bias, Google Blocks Gender-Based Pronouns From New AI Tool
Reuters
Paresh Dave
November 27, 2018
In May, Google introduced a new feature for Gmail that automatically completes sentences for users as they type, but the tool will not suggest gender-based pronouns because the risk is too high that the system might predict someone's sex or gender identity incorrectly and offend users. Google is taking extra precautions because it wants to bill itself as better understanding the nuances of artificial intelligence (AI) when compared to its competitors. Gmail has 1.5 billion users, and its "Smart Compose" feature assists on 11% of messages worldwide sent from Gmail.com. Google's decision to be cautious on gender follows some high-profile embarrassments for the company's predictive technologies. The policy of banning gendered pronouns also affects the list of possible responses in Google's Smart Reply service, which allows users to respond instantly to text messages and emails with short phrases such as "sounds good."
|
|
