Welcome to the July 13, 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."
|
|
|
Tyson Turns to Robot Butchers, Spurred by Coronavirus Outbreaks
The Wall Street Journal
Jacob Bunge; Jesse Newman
July 9, 2020
Engineers and scientists at meatpacking company Tyson Foods are developing an automated deboning system to process meats in order to avoid spreading infectious diseases like Covid-19, as well as maintain productivity amid staff shortages. Tyson CEO Noel White said the pandemic likely will accelerate company investments—currently estimated at $500 million—in technology and automation. Tyson president Dean Banks said training robots to cut and sort meat, which involves soft material and variability, is a massive operational challenge compounded by low temperatures and blood splatter at meat plants. Tyson technicians are attempting to teach machines to identify and quickly adjust to differences in meat coloration and shape. Banks said the technology is necessary to relieve plant bottlenecks due to a lack of skilled workers. Tyson designed a water-jet cutting system that carves up chicken breasts more precisely than humans.
*May Require Paid Registration
|
AI Enables Efficiencies in Quantum Information Processing
U.S. Army Research Laboratory
July 8, 2020
Researchers from the U.S. Army Combat Capabilities Development Command's Army Research Laboratory and Tulane University integrated machine learning with quantum information science, applying photon measurements to reassemble the quantum state of an unknown system. The team used quantum state tomography to define an unknown quantum system, preparing and measuring identical unknown quantum systems and employing a complex computational process to ascertain the system most consistent with the measurement results. A recent implementation reconstructed quantum states and standard, more computationally intensive methods—and outperformed those techniques in several instances. The team modeled familiar sources of measurement error to train the machine learning artificial intelligence, then refined the system when measurements were noisy or entirely absent. The actual reconstruction demands relatively modest resources, because the team can pile the expensive computation into the training process.
|
Smartwatch Hack Could Send Fake Pill Reminders to Patients
BBC News
July 9, 2020
Researchers at U.K.-based security firm Pen Test Partners found security vulnerabilities in smartwatches that aim to help elderly patients remember to take their pills or alert caregivers if the patient wanders off. The software in question, SETracker, is used in a wide range of low-cost smartwatches and has been downloaded more than 10 million times. The researchers expressed concern that hackers could send medication alerts to the watches numerous times, raising the risk of overdoses. The China-based company behind the software fixed the security flaw after being notified, but Pen Test researchers said there was no way to know whether the flaw had been exploited before being fixed.
|
Virtual Try-Ons Replacing Fitting Rooms During the Pandemic
The Washington Post
Abha Bhattarai
July 9, 2020
The Covid-19 pandemic has spurred consumers to substitute fitting rooms with augmented reality (AR) using their smartphones and computers. Jewelry brand Kendra Scott is allowing shoppers to try on earrings virtually with their iPhone browsers, while AR provider Zeekit is about to launch the world's largest virtual fitting room, letting shoppers drag hundreds of thousands of clothing items onto photos of themselves. Zeekit co-founder Yael Vizel said shoppers who use the company's artificial intelligence to try on clothing virtually are five times more likely to buy—and keep—the items. Some retailers are pursuing contactless fittings in stores, with mall operator Brookfield Properties recently announcing it will add three-dimensional body scanners that can direct consumers to the best-fitting clothing brands and sizes. Privacy experts worry virtual try-ons leave an exploitable trail of personal and biometric data, and recommend consumers check whether applications are recording this information.
*May Require Paid Registration
|
Software Suite Expedites Reproducible Computer Simulations
Vanderbilt School of Engineering
July 8, 2020
Researchers at Vanderbilt University have developed a suite of open source software tools that expedites the process of reproducing computer simulations. It can be difficult to reproduce experiments and obtain the same results in some fields, mainly because details provided in peer-reviewed publications aren't sufficient to recreate the conditions of the original experiment. The Molecular Simulation and Design Framework (MoSDeF) aims to eliminate those by automating as many research steps as possible, with one component providing validated parameters and applying forcefields (mathematical models for how molecules interact with each other) automatically. All modules and workflows developed for MoSDeF build on the scientific Python stack, which simplifies the creation, atom-typing, and simulation of complex molecular models. Vanderbilt’s Clare McCabe said, “By using freely available tools designed for collaborative code development, such as GitHub and Slack, we are creating a community-developed effort.”
|
Filter Protects Against Deepfake Photos, Videos
NextGov.com
Jeremy Schwab
July 8, 2020
Computer scientists at Boston University (BU) have created an algorithm that prevents deepfakes by adding an imperceptible filter to videos and photos before they are uploaded to the Internet. If a deep neural network attempts to alter a protected image or video, the filter leaves the media unchanged, or completely distorted and rendered unusable as a deepfake. The open source code has been made publicly available on GitHub. Said BU's Sarah Adel Bargal, "We covered what we call 'white-box' attacks in our work, where the network and its parameters are known to the disruptor. A very important next step is to develop methods for 'black-box' attacks that can disrupt deepfake networks [in ways] inaccessible to the disruptor."
|
Helping Drug-Delivering Particles Squeeze Through a Syringe
MIT News
Anne Trafton
July 8, 2020
Researchers at the Massachusetts Institute of Technology (MIT) developed a computational model that could improve the injectability of drug-delivering microparticles and prevent them from clogging a syringe. The model resulted in a six-fold increase in the percentage of successfully injected microparticles, according to the researchers. The researchers experimented with the size and shape of microparticles, the viscosity of the solution in which they are suspended, and the size and shape of the syringe and needle used to deliver them. The data was used to train a neural network to predict how each of these factors impact injectability. Said MIT's Morteza Sarmadi, "We hope that our work can improve the clinical translation of novel and advanced controlled-release drug formulations."
|
Breakthrough Machine Learning Approach Produces 50X Higher-Resolution Climate Data
National Renewable Energy Laboratory
July 7, 2020
At the U.S. Department of Energy's National Renewable Energy Laboratory (NREL), researchers developed a machine learning technique to improve the resolution of wind velocity data 50 times, and that of solar irradiance data 25 times. Using adversarial training, the team trained two types of neural networks to respectively recognize the physical characteristics of high-resolution solar irradiance and wind velocity data, and to incorporate those properties within the coarse data. The networks produced more realistic data over time and improved their differentiation between real and fake inputs, enabling the researchers to add 2,500 pixels for every original pixel. NREL's Karen Stengel said, "By using adversarial training—as opposed to the traditional numerical approach to climate forecasts, which can involve solving many physics equations—it saves computing time, data storage costs, and makes high-resolution climate data more accessible."
|
Technology Bridges the Gap to Better Sight
The New York Times
Janet Morrissey
July 7, 2020
IrisVision aims to help people with low vision see and read clearly using a smartphone, virtual reality (VR) headset, and algorithms. The company expects the wearable device bearing its name to be used to diagnose conditions and test and treat patients remotely within the next two years. In people with visual acuity as low as 20/1000 and with conditions like macular degeneration, diabetic retinopathy, and glaucoma, IrisVision helps the brain use parts of the eye that still function properly. The VR headset and algorithms enhance images captured by the smartphone's camera by filling in the gaps and remapping the scene to provide a complete picture. Voice control, video streaming, Alexa, and other interactive features have been added to the headset. Forrester Research's Kjell Carlsson said the device's size would have to be reduced, along with its price tag ($2,950), before it could go mainstream.
|
Survey Finds Americans Skeptical of Contact Tracing Apps
IEEE Spectrum
Jeremy Hsu
July 7, 2020
A recent survey by researchers at Cornell University found that only 42% of respondents support the use of contact tracing apps to notify those potentially exposed to Covid-19. Cornell's Sarah Kreps said GPS-based apps likely are more effective from a public health standpoint, but GPS location data is intrusive from a privacy standpoint because it can reveal behavioral and lifestyle insights about users. Survey respondents viewed apps differently based on whether they use GPS location data or Bluetooth. Respondents also were more willing to download apps with decentralized data storage, rather than centralized systems that might be overseen by a company or government agency.
|
Tool Identifies Cancer-Causing Mutations Lurking in 98% of the Genome
News-Medical
Emily Henderson
July 6, 2020
Scientists at St. Jude Children's Research Hospital, working with colleagues at China's Soochow University Children's Hospital and Shanghai Children's Medical Center, have developed a computational tool that can identify tumor-inducing mutations in most of the human genome. The cis-expression (cis-X) algorithm identifies abnormal pathogenic variants and oncogenes activated by such variants in regulatory noncoding DNA of patient tumors, which constitutes 98% of the genome. Cis-X looks for genes with altered expression, using a novel mathematical model to find genes expressed on just one chromosome at aberrantly high levels, then seek the cause of the abnormal expression by looking for alterations in regulatory regions of noncoding DNA within a three-dimensional genome architecture. St. Jude's Jinghui Zhang said, "Cis-X offers a powerful new approach for investigating the functional role of noncoding variants in cancer, which may expand the scope of precision medicine to treat cancer caused by such variants."
|
Home Router Warning: They're Riddled With Known Flaws and Run Ancient, Unpatched Linux
ZDNet
Liam Tung
July 6, 2020
Researchers at the Fraunhofer Institute for Communication (FKIE) in Germany studied 127 home routers from seven brands and found that 46 had not had a security update within the past year. The study also revealed that many routers have hundreds of known vulnerabilities, and that vendors are shipping firmware updates without fixing known vulnerabilities. The researchers further found that German router manufacturer AVM was the only one that didn't publish private cryptographic keys in its router firmware. A Linux operating system was used by about 90% of the routers studied, but researchers found that manufacturers were not updating it. Said FKIE's Johannes vom Dorp, "Really, all the manufacturers would have to do is install the latest software, but they do not integrate it to the extent that they could and should."
|
Robotic Lab Assistant Is 1,000 Times Faster at Conducting Research
The Verge
James Vincent
July 8, 2020
At the University of Liverpool in the U.K., researchers have developed a robotic lab assistant that can conduct scientific experiments 1,000 times faster than a human. Although the robot needs to be programmed with the location of lab equipment and is unable to design its own experiments, it can work 22 hours a day, allowing scientists to automate time-consuming and tedious research. Said Liverpool’s Andy Cooper, "The idea is not to do things we would do faster, but to do bigger, more ambitious things we wouldn't otherwise try to tackle." The robot uses LIDAR technology to navigate the lab, which enables it to operate in the dark, and manipulates the lab equipment using an industrial arm. The ability to reprogram the robot enables it to handle a variety of tasks. Said Cooper, "The idea was to automate the researcher, rather than the instrument."
|
|
