Welcome to the August 21, 2020 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."
|
|
|
Ann Syrdal, Who Helped Give Computers a Female Voice, Dies at 74
The New York Times
Cade Metz
August 20, 2020
Computer science researcher Ann Syrdal, who helped to create synthetic female voices and establish a platform for modern digital assistants, has died at 74. At AT&T she joined a community of scientists who began developing synthetic speech systems in the mid-1980s, and helped engineer a system called Natural Voices, described as "the first truly high-quality female synthetic voice." Syrdal and colleagues developed methods for stitching together snippets of recorded human speech to form new words and new sentences on the fly, rather than producing sounds from scratch. Natural Voices used a female voice when it topped an international speech synthesizer competition in 1998. Syrdal was named a fellow of the Acoustical Society of America in 2008, to recognize her role in the advent of female speech synthesis.
*May Require Paid Registration
|
Covid-19 Crisis Accelerates U.K. Military's Push Into Virtual War Gaming
Financial Times
Helen Warrell
August 19, 2020
The U.K. Ministry of Defence (MoD) is looking to fast-track new virtual reality technology from software developer Improbable that would create a digital replica of the country. The technology could be used to test resilience to future pandemics, natural disasters, and attacks by hostile states. Known as a "single synthetic environment," the technology layers maps of geographical terrain and critical infrastructure with details of fuel, power, and water supplies, telecom networks, weather patterns, and more. This virtual "twin" uses artificial intelligence to simulate future scenarios and offers the ability to "war game" responses. Ultimately, the MoD hopes to build simulations of likely conflict zones. Said Improbable's Joe Robinson, "What we're aiming for in the longer term is ... to enable governments to test ideas and test choices of action in a virtual world before implementing them in the real world."
*May Require Paid Registration
|
'This Is Our Future': Fairfax Tests Region's First Self-Driving Shuttle for Public Transit
The Washington Post
Luz Lazo
August 16, 2020
Fairfax County, VA, and Dominion Energy are testing an autonomous electric shuttle that could start ferrying passengers between the Dunn Loring-Merrifield Metro station and the Mosaic District as early as this fall. The driverless Relay bus will travel a mile-long stretch without steering wheel or pedals, with sensors keeping it on a predefined route and detecting and evading obstacles. The bidirectional vehicle is currently learning its fixed route, undergoing assessment, and awaiting approval from the U.S. National Highway Traffic Safety Administration. The Relay is mapped to the route, and uses global-positioning system technology and LiDAR to continuously scan its environment for obstacles, slowing down if any are detected. Officials and academics from state universities will observe tests and examine ridership, rider demographics, and other issues.
*May Require Paid Registration
|
Hackers Can Remotely Hijack Enterprise, Healthcare Temi Robots
ZDNet
Charlie Osborne
August 6, 2020
At the Black Hat USA event, McAfee's Advanced Threat Research team unveiled an exploit that could remotely hijack "personal robots" used in hospitals, care homes, and enterprises. The team focused on Robotemi Global's Temi robots, which can be controlled using a mobile device to scan their quick response code. The researchers discovered four vulnerabilities, including use of hard-coded credentials, an origin validation error, missing authentication for critical functions, and an authentication bypass. "Together, these vulnerabilities could be used by a malicious actor to spy on Temi's video calls, intercept calls intended for another user, and even remotely operate Temi—all with zero authentication," the team said. Robotemi Global quickly patched the bugs after McAfee reported them in early March.
|
Tech Firms Offer Carbon Tracking Software to Meet Oil-Investor Demands
Reuters
Liz Hampton; Jennifer Hiller
August 18, 2020
Tech companies from SAP and Salesforce.com to Arizona-based startup Persefoni are rolling out software to help oil and gas producers measure their carbon emissions. These products have been released in response to calls from top investors for energy producers to reduce greenhouse gases. Developers say tracking their emissions can show investors they are serious about efforts to curb them. Persefoni's product uses artificial intelligence built around the Greenhouse Gas Protocol accounting method, but it eventually will meet all reporting frameworks and standards. The system standardizes data to calculate a carbon footprint. Said Deloitte's Kat Lampen, "[Companies] are realizing, like financial data, this needs to be collected in the system with the right validation and checks on it."
|
New Approach Takes Quantum Key Distribution Further
The Optical Society
August 19, 2020
Researchers from China's Nanjing University of Posts and Telecommunications have demonstrated secure measurement-device-independent quantum key distribution (MDI-QKD) transmission across 170 kilometers (106 miles). The protocol uses photons with three characterized quantum states to encode data, which delivers immunity to attacks targeting detection devices that measure the quantum properties of individual photons. The researchers utilized a state-of-the-art experimental installation for encoding and detection and showed that this approach can transmit keys over longer distances and at higher rates than similar MDI-QKD protocols. Theoretical calculations indicated secure transmission could be possible over distances of up to 200 kilometers (124 miles).
|
What if You Could Better Control What Mobile Apps Do With Your Data?
Carnegie Mellon University
Daniel Tkacik
August 14, 2020
Researchers from Carnegie Mellon University's CyLab have developed a digital privacy assistant that uses machine learning to allow mobile app users to better control their data without being overwhelmed by a large number of privacy decisions. Users currently cannot selectively control how their data will be used. The privacy assistant asks users five to six questions about their privacy preferences to help it infer many of the privacy decisions they would want to make. If the machine learning does not have enough data to make a prediction, it will ask a couple of follow-up questions. Said CyLab's Daniel Smullen, "The surprising finding here is that when you include permissions that take the purpose of data collection into account, you actually end up with more powerful predictive models than if you ignored purpose information."
|
Facebook AI Takes MRI Images Four Times Faster by 'Recreating Missing Parts'—Experts Can't Tell the Difference
The Daily Mail
Ryan Morrison
August 18, 2020
Facebook and experts from New York University have developed an open source artificial intelligence (AI) model for magnetic resonance imaging (MRI) scanners that can reconstruct missing sections in less detailed scans. The developers trained the fast MRI model on thousands of full scans, and tests with six radiographers showed the method was accurate and produced higher-quality images four times faster than normal. The researchers said the model could "make MRI as fast as or faster than an X-ray," but with much more data available for diagnosis and without radiation. "We are using AI to create complete images from far less raw data,” Facebook said. “Since collecting that data is what makes MRIs so slow, this has the potential to speed up the scanning process significantly."
|
Virtual Assistant Will Screen Your Phone Calls to Block Spammers
New Scientist
Chris Stokel-Walker
August 17, 2020
Georgia Institute of Technology researchers have developed a virtual assistant that screens phone calls to block robocallers. Numbers not on a phone's contact list are diverted to the assistant, which asks the caller to state the recipient's name; if the caller responds, the assistant interrupts with, "Sorry I didn't get that. Can you please say who you are trying to reach again?" If the caller is live, they will stop speaking, and the assistant will pass on the call and a transcribed name through an application. The system blocked 100% of 8,000 recorded robocalls and 97.8% of 21 human callers who engaged with the assistant but occasionally used the incorrect name. David Schlangen at Germany's University of Potsdam said, "It's like a spam filter that checks that the sender has at least minimal knowledge about the addressee and is flexible enough to produce that knowledge on request."
|
Soldiers Could Teach Future Robots How to Outperform Humans
U.S. Army Research Laboratory
August 12, 2020
Researchers at the U.S. Army Combat Capabilities Development Command's Army Research Laboratory (ARL) and the University of Texas at Austin developed an algorithm that enables autonomous ground vehicles to enhance their navigation systems by observing human drivers. The APPLD (adaptive planner parameter learning from demonstration) algorithm tunes existing systems to act more like human demonstrations, adding flexibility and adaptability while preserving classical navigation functions. "APPLD tunes the system automatically by watching a human drive the system—something that anyone can do if they have experience with a video game controller," ARL's Jonathan Fink said. Researchers expect the algorithm will allow soldiers to train future robots to operate next-generation field combat vehicles autonomously in off-road deployments.
|
Companies Step Up Distribution Automation Under Pandemic Strains
The Wall Street Journal
Jennifer Smith
August 12, 2020
More companies are considering automating distribution and fulfillment during the coronavirus pandemic as consumers do more shopping online and social distancing practices in warehouses worsen strains on logistic networks. American Eagle Outfitters is rolling out another 26 robots at its main U.S. distribution centers. The robots from Kindred Systems Inc. sort through apparel using mechanical arms, computer vision, and artificial intelligence. Said the company's Shekar Natarajan, "During non-Covid times, if demand grew by 50% I would go hire 300 more people. You cannot actually bring in 1,000 to 2,000 untrained people into the distribution facility and maintain safe working conditions." A survey of warehouse operators by Honeywell Intelligrated found that more than half were willing to invest in automation in response to the pandemic. However, Interact Analysis predicts that the pandemic will reduce global industrial robot shipments this year.
*May Require Paid Registration
|
Mirai Botnet Is Targeting RCE Vulnerability in F5 BIG-IP Software
Computing
Dev Kundaliya
August 4, 2020
Researchers at cybersecurity company Trend Micro warn that the Mirai botnet is attempting to exploit a remote code execution vulnerability in F5 BIG-IP software discovered in July, with an Internet of Things (IoT) botnet downloader that can be added to new malware variants. The downloader looks for exposed BIG-IP devices and sends a malicious payload to target systems with the bug. The flaw exists in devices' Traffic Management User Interface (TMUI), and attackers must deliver a specially crafted HTTP request to the TMUI-hosting server. The researchers said a successful exploit would enable attackers "to create or delete files, disable services, intercept information, run arbitrary system commands and Java code, completely compromise the system, and pursue further targets, such as the internal network." FT Networks has patched the flaw, and the researchers recommended that system administrators constantly monitor manufacturers' releases to guarantee their IoT devices' firmware runs on the latest versions.
|
Google Uses Crowdsourced Photos to Recreate Landmarks in 3D for AR/VR
VentureBeat
Jeremy Horwitz
August 12, 2020
Researchers at Google have created a method to automate and improve the three-dimensional (3-D) modeling process using a neural network. Google's NeRF-W, or neural radiance fields in the wild, system uses crowdsourced photos of a location to convincingly recreate landmarks and lighting in 3-D for users of virtual reality and augmented reality devices. Standard NeRF aims to determine where light rays terminate in order to extract 3-D depth data from two-dimensional images. NeRF-W goes further by using "in-the-wild photo collections" as inputs to broaden the computer's ability to view landmarks from multiple angles. It also can separate out photographic and environmental variations and shot-to-shot object differences, and recreate scenes with radiance guidance for scene lighting and shadowing. A comparison of standard NeRF and NeRF-W results shows that the new system allows users to experience complex architecture as it actually appears, with time-of-day and weather variations.
|
|
