Welcome to the June 21, 2019 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."
|
|
|
Florida City Agrees to Pay Hackers $600,000
The New York Times
Patricia Mazzei
June 19, 2019
The city council of Riviera Beach, FL, submitted to hackers' demands by agreeing to pay almost $600,000 in bitcoin to ransom back their computer systems, held hostage by malware for the past three weeks. The incident is the latest in a string of high-profile ransomware attacks, one of which cost Baltimore $18 million to repair. Jason Rebholz with the Moxfive technology advisory firm said the relatively high price for restoring control of Riviera Beach's systems suggests extortionists' growing technical sophistication is encouraging them to target more government agencies. A 2016 survey of U.S. local government chief information officers found more than a third were using dated technologies, increasing their vulnerability to cyberattacks.
*May Require Paid Registration
|
Denoising Method Generates Sharper Photorealistic Images Faster
SIGGRAPH 2019
June 11, 2019
Researchers at the Massachusetts Institute of Technology, Adobe, and Finland's Aalto University have created a new technique for generating higher-quality images and scene designs faster, using a deep learning denoising methodology. The end results are sharper images that render fine details from sample features, like shadowing, indirect lighting, motion blur, and depth of field. Adobe's Michael Gharbi said, "Our algorithm can produce clean images from noisy input images with very few samples, and could be useful for producing quick rendered previews while iterating on scene design." The technique uses a convolutional network that learns to denoise renderings directly from a raw set of Monte Carlo samples, instead of from reduced, pixel-based representations; an innovative kernel-predicting computational framework "splats" individual samples onto nearby pixels to refine image composition.
|
Investors Urge AI Startups to Inject Early Dose of Ethics
The Wall Street Journal
Jared Council
June 16, 2019
Artificial intelligence (AI) startup investors are urging companies to improve their products from an ethical perspective, using a code of ethics to guide operations, a tool to explain how software makes decisions, and best practices that feature consistent, open communication and immediate feedback about algorithmic output. For example, Analytics Ventures' startups employ a tool called Klear to forensically analyze why AI systems arrive at decisions. Analytics Ventures' Andreas Roell said, "I see explainability as a core component of having an ethical guardrail around AI." Meanwhile, a tech accelerator run by Innovation Works unveiled a voluntary ethics component to its program for startups, in partnership with Carnegie Mellon University. The program targets issues like bias and data privacy, and asks each startup’s founders to craft an ethical values statement.
|
Discovery of a “Holy Grail” with the Invention of Universal Computer Memory
Lancaster University
June 20, 2019
Researchers at the University of Lancaster in the U.K. have developed a new type of computer memory that could solve the digital technology energy crisis. The new electronic memory device can immediately reduce peak power consumption in datacenters by 20%, and could lead to the development of computers that do not need to boot up and could instantaneously enter energy-saving sleep mode. The new device could replace Dynamic Random Access Memory (DRAM), the "working memory" of conventional computers, as well as the solid-state memory in flash drives. Said Lancaster researcher Manus Hayne, "Our device has an intrinsic data storage time that is predicted to exceed the age of the universe, yet it can record or delete data using 100 times less energy than DRAM."
|
Walmart Kickstarting $1-Trillion Driverless Delivery Market
Bloomberg
Keith Naughton; Matthew Boyle
June 19, 2019
Walmart is testing a fleet of "robo-vans" to deliver packages at least part of the way to consumers. The retail giant hopes to exploit this "middle mile" market, which could potentially be worth $1 trillion, using self-driving technology developed by Silicon Valley startup Gatik. The robo-vans follow fixed routes, to reduce the likelihood of accidents. Many of these routes are already mapped out using human drivers, making it unnecessary to build new infrastructure to load and receive the goods. Gatik CEO Gautam Narang said, "This middle mile is the most expensive part of the whole supply chain; it's a huge pain point. This fills a big gap in the market."
*May Require Paid Registration
|
Software to Protect World's Most Endangered Species
EPFL (Switzerland)
Sandrine Perroud
June 20, 2019
Researchers at the Swiss Federal Institute of Technology, Lausanne (EPFL) are integrating genetic and environmental databases to help identify species threatened by climate change more accurately, in order to design conservation strategies. The R.Samada software developed in EPFL's Laboratory of Geographic Information Systems assists evolutionary biologists by offering them concurrent direct access to these databases. R.Samada evaluates how genetic and environmental information relate to each other, and generates graphs and maps for rapid data visualization. EPFL's Stephane Joost said, "The software identifies the genes involved in the process through which a species evolves to adapt to weather conditions." R.Samada will inform researchers' efforts to conserve endangered species by gauging genetic variants' fitness for survival in certain eco-climatic zones.
|
Python vs. R, Salaries: Top Data Science Job Trends
ZDNet
Steve Ranger
June 19, 2019
U.K. technology recruitment firm Harnham cited growing demand for data science specialists as the reason behind major wage increases as companies increasingly try to lure those in-demand professionals to change jobs. Entry-level data scientists can earn £45,000 ($57,000) a year in London and £34,000 ($43,100) in other U.K. cities, while senior director-level data scientists can expect to earn an average annual salary of £140,000 ($177,200) in London. Given the increasing demand, data science specialists tend to stay in one job for less than three years, according to Harnham, which also named Python the top programming language for data scientists, "with R falling firmly into second place."
|
CMU Researchers Use Computer Vision to See Around Corners
TechCrunch
Darrell Etherington
June 19, 2019
Researchers at Carnegie Mellon University, the University of Toronto in Canada, and the U.K.'s University College London have developed a technique for imaging objects outside a device's line of sight. The technique combines special sources of light, in conjunction with sensors and computer vision processing, to deduce or reconstruct highly detailed imagery without having previously captured the image or "viewed" it directly. The system uses ultrafast laser light, bounced off a wall, to illuminate objects hidden around corners. The reflected light is captured by sensors, and the team measures the time it takes for the reflected light to return to its origin point. Using multiple measurements and information on the target object’s geometry allows the team to reconstruct objects around corners with great accuracy.
|
Oil Group Hopes Supercomputer Will Help it Find Oil
Reuters
Bate Felix
June 18, 2019
Energy company Total said its upgraded IBM Pangea III supercomputer will help geologists locate oil faster and at reduced expense. The company said the system will assist in seismic-data processing as it searches for hydrocarbons 10 times faster than it was previously capable of doing. The Pangea III's computing power has grown from 6.7 petaflops in 2016 to 31.7 petaflops today, earning it the top ranking in the oil and gas sector and the 11th highest rating in the overall TOP500 ranking of the 500 most powerful non-distributed computer systems in the world. Total's Kevin McLachlan said Pangea III will use new algorithms that can process vast datasets with greater accuracy and at a higher resolution, so "What used to take a week, now takes us a day to process."
|
First 'Quantum Drone' Takes Off
IEEE Spectrum
Charles Q. Choi
June 18, 2019
Researchers at Nanjing University in China have developed a quantum drone to serve as an airborne node in a future quantum network. The eight-rotor, 35-kg. (77-lb.) octocopter is equipped with an onboard quantum communication system. During testing, the researchers demonstrated that the quantum drone can operate while hovering in midair for 40 minutes at a time. The device is capable of maintaining two air-to-ground links, each about 100 meters long, and can receive and transmit entangled photons during the daytime, on a clear night, and even on a rainy night. The researchers believe they can scale down the quantum communication system to fit in small consumer drones for on-demand local-area quantum networks, or scale it up for high-altitude drones that would serve as nodes in wide-area networks spanning hundreds of kilometers.
|
Your Phone Can Become a Robot That Does the Boring Work
Purdue University News
Kayla Wiles
June 17, 2019
Purdue University researchers have developed a smartphone app that allows a user to easily program any robot to perform a basic activity like lifting objects from one area and carrying them to another. The embedded app, VRa, utilizes augmented reality to let the user either walk where the robot should go to perform tasks, or draw a workflow directly into real space. The app offers options for how those tasks should be performed, such as within a specific amount of time, in repetition, or following the completion of a job by another machine. After programming, the user places the phone into a dock attached to the robot; the phone then serves as both eyes and brain for the robot, controlling its navigation and activities. Said Purdue researcher Karthik Ramani, "Our goal is for everyone to be able to program robots, and for humans and robots to collaborate with each other."
|
AI to Manage Road Infrastructure Via Google Street View
Royal Melbourne Institute of Technology (Australia)
Michael Quin
June 19, 2019
Researchers at the Royal Melbourne Institute of Technology (RMIT) in Australia have developed an automated system that can identify and confirm the locations of street signs. Trained in object detection using Google Street View images, the researchers said the system detected traffic signs with almost 96% accuracy and identified the type of sign with almost 98% accuracy. The system was trained to see "stop" and "give way" signs, but could be altered to identify many other inputs, and is scalable for use by local governments and traffic authorities. Said RMIT’s Andrew Campbell, "Our system, once set up, can be used by any spatial analyst; you just tell the system which area you want to monitor, and it looks after it for you."
|
|
