Welcome to the March 15, 2021 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."

Awaiting vaccination shots in Richmond, VA. Faulty Software Snarls Vaccine Sign-Ups
The New York Times
Kellen Browning
March 12, 2021


Persistent flaws in software for setting up Covid-19 vaccination appointments online threaten to slow the U.S. vaccine rollout, with many states switching software providers originally recommended by the U.S. Centers for Disease Control and Prevention (CDC), with little improvement. The CDC recommended multinational Deloitte's Vaccine Administration Management System (VAMS) software, which Deloitte said was originally intended for smaller groups at early stages of state rollouts. Health experts blamed glitches on multiple impediments, including developers condensing work that would normally take years into weeks, localities' individual eligibility requirements, and the inability for different scheduling systems to communicate with each other. Some frustrated states switched from VAMS to Maryland's PrepMod software, which has been undermined by myriad shortcomings, including failures to reserve appointment slots as people filled out their information. PrepMod's Tiffany Tate has faulted healthcare workers, and not the software itself, for such difficulties.

Full Article
*May Require Paid Registration

Researchers demonstrate the experimental setup they used to observe how group participants are prompted to speak up in an activity led by a robot. Robots Can Use Eye Contact to Draw Out Reluctant Participants in Group
KTH Royal Institute of Technology (Sweden)
David Callahan
March 10, 2021


Researchers at Sweden's KTH Royal Institute of Technology have demonstrated that robots can encourage participation in group settings by making eye contact with reluctant participants. The study involved robots leading a Swedish word game with a pair of participants, one fluent in Swedish and one learning the language. The game required the players to give clues in Swedish to the robot, whose face was an animated projection on a plastic mask. The robot would redirect its gaze to the less-fluent player to encourage their participation. KTH's Ronald Cumbal said the study’s results demonstrate "that even when there is an imbalance in skills required for the activity, the gaze of a robot can still influence how the participants contribute."

Full Article
New Approach Found for Energy-Efficient AI Applications
Graz University of Technology (Austria)
Christoph Pelzl
March 11, 2021


Researchers at Austria's Graz University of Technology (TU Graz) have demonstrated a new approach to energy-efficient artificial intelligence that needs very few signals to function, and that assigns meaning to pauses between signals. TU Graz's Wolfgang Maass and Christoph Stöckl optimized artificial neural networks in computer models for image classification so the neurons only have to transmit extremely simple signals occasionally, achieving an accuracy similar to that of state-of-the-art tools. The data transmission model relies not only on how many spikes a neuron sends out, but also on when the neuron transmits the spikes. Maass said, "With just a few spikes—an average of two in our simulations—as much information can be conveyed between processors as in more energy-intensive hardware."

Full Article

Illustration of a water purification membrane with computationally designed molecular-scale patterning of surface functional groups. Computing Clean Water
The Current (University of California, Santa Barbara)
Andrew Masuda
March 11, 2021


Researchers at the University of California, Santa Barbara (UCSB), the University of Texas at Austin, and the U.S. Department of Energy's Lawrence Berkeley National Laboratory have computationally modeled affinity between solutes and membrane surfaces, to characterize their effects on water purification. The research team developed a genetic algorithm to repattern surfaces by reconfiguring surface chemical groups to minimize or maximize a given solute's affinity for the surface, or to maximize a solute's surface affinity relative to that of another. Simulations demonstrated a stronger link between surface affinity and how molecules near a surface or a solute reorganize in response. Former UCSB researcher Jacob Monroe said, "This work tackles the grand challenge of understanding how to design next-generation membranes that can handle huge yearly volumes of highly contaminated water sources."

Full Article
NTSB Asks NHTSA for More Self-Driving Car Rules, Citing Tesla's 'Full Self-Driving' Beta
CNet
Kyle Hyatt
March 12, 2021


The U.S. National Transportation Safety Board (NTSB) has requested tougher rules for advanced driver-assistance systems (ADAS) and self-driving car development and testing on public roads from the National Highway Traffic Safety Administration (NHTSA). NTSB's Robert Sumwalt cited electric-vehicle company Tesla's beta test of its Level 2 ADAS on public roads with regular drivers, called Full Self-Driving, as "having limited oversight or reporting requirements." Sumwalt said, "NHTSA's hands-off approach to oversight of [autonomous vehicle] testing poses a potential risk to motorists and other road users." Ideally, NHTSA would not only impose stronger rules for ADAS testing on public roads, but also provide clear terminology that improves customer understanding of ADAS' capabilities, and ban manufacturers' use of marketing to exaggerate those capabilities.

Full Article
Deep Learning Enables Real-Time 3D Holograms on Smartphone
IEEE Spectrum
Charles Q. Choi
March 10, 2021


Massachusetts Institute of Technology (MIT) scientists can generate photorealistic three-dimensional (3D) holograms in color on a smartphone, in real time. The technique utilizes a deep learning convolutional neural network (CNN) to generate holograms without requiring separate calculations for how each chunk of the image appears. The MIT researchers compiled a database of 4,000 computer-generated images, each with color and depth information for each pixel, as well as a corresponding 3D hologram for each image. The CNN tapped this data to calculate an optimal hologram generation process, then produced new holograms from images with depth information calculated from a multi-camera setup or from LiDAR sensors included in certain iPhones. The system can generate 60 color 3D holograms per second with a resolution of 1,920 by 1,080 pixels on one consumer-grade graphics processing unit (GPU), using less than 620 kilobytes of memory.

Full Article

License-plate scanners, a key tool for local police as well as the Justice Department. License-Plate Scans Aid Crime-Solving But Spur Little Privacy Debate
The Wall Street Journal
Byron Tau
March 10, 2021


Law enforcement agencies increasingly are using data gathered by the vast network of automated license-plate scanners to solve crimes. The scanners initially were placed on telephone poles, police cars, toll plazas, bridges, and in parking lots but now can be found on tow trucks and municipal garbage trucks as well. License-plate scans were instrumental in the arrests of several suspected rioters at the U.S. Capitol. However, there are concerns about abuse, misidentification, and the scope of data collection, given that, for instance, some systems read a plate's number but not its state. Electronic Frontier Foundation's Dave Maass said, "License-plate readers are mass surveillance technology. They are collecting data on everyone regardless of whether there is a connection to a crime, and they are storing that data for long periods of time."

Full Article
*May Require Paid Registration
Hackers Act Differently if Accessing Male or Female Facebook Profiles
New Scientist
Chris Stokel-Walker
March 10, 2021


University of Vermont and Facebook researchers found that hackers on the social media platform display different behavior depending on the age and gender listed on the hacked Facebook account. The researchers created 1,008 realistic Facebook accounts and leaked the login details for 672 of them on websites used by hackers to trade compromised credentials. They used the other accounts to populate the friendship groups of the leaked accounts to monitor them over a six-month period. The researchers found that 46% of the leaked accounts were accessed 322 times combined. They also determined that hackers messaged the friends of younger profiles more than those of older profiles, and that in many cases male accounts—but never female accounts—were vandalized.

Full Article
*May Require Paid Registration
ML Gets Scientific Smarts in Leap for Battery Research
Stanford Linear Accelerator Center
Glennda Chui
March 8, 2021


Researchers at Stanford University, the U.S. Department of Energy's Stanford Linear Accelerator Center, the Massachusetts Institute of Technology, and the Toyota Research Institute have accelerated battery design via machine learning. The research combined advances that expedited both battery testing and the process of weeding out possible charging methods to find the best one, with a method for exposing the underlying physics or chemistry of longer-lasting batteries. Stanford's Will Chueh said, "We are teaching the machine how to learn the physics of a new type of failure mechanism that could help us design better and safer fast-charging batteries." Stanford's Stephen Dongmin Kang said the new approach disclosed a very different view of how lithium moves within a battery than previously believed. Chueh hopes these insights can solve the problem of fast charging in batteries.

Full Article
Cloud Computing Could Prevent 1B Metric Tons of CO2 Emissions by 2024
ZDNet
Stephanie Condon
March 8, 2021


A new report from market research firm IDC found that shifting to cloud computing should reduce carbon dioxide (CO2) emissions from 2021 to 2024 by at least 629 million metric tons. IDC projects 60% of datacenters will implement "smarter" sustainability practices by 2024, preventing more than 1 billion metric tons of CO2 emissions. Many tech giants have pledged to lower their own carbon emissions, with IBM and Facebook promising to achieve net-zero CO2 emissions by 2030, and Amazon by 2040. Microsoft expects its net carbon emissions to be negative by 2030. Apple plans to be carbon-neutral by 2030, and Google, which has been carbon-neutral since 2007, plans for its entire business to be carbon-free by 2030.

Full Article

A student on her laptop using a tool that allows students to work with it, online, on their own, to assess the credibility of digital imagery. Tool Makes Students Better at Detecting Fake Imagery, Videos
Uppsala University (Sweden)
Linda Koffmar
March 9, 2021


Researchers at Sweden's Uppsala University have devised a digital self-test that trains users to assess news items, images, and videos posted on social media for credibility. Uppsala's Thomas Nygren said, "The aim is for the students to get better at uncovering what isn't true, but also improve their understanding of what may be true even if it seems unlikely at first." The News Evaluator is organized so students can use it online on their own, displaying news items in a social-media format, with pictures or videos. Learners also are encouraged to double-check the news to confirm the stories' authenticity, then receive feedback on their performance. A scientific review of the News Evaluator verified the tool improved students' ability to apply critical thinking to digital sources; those who had received guidance and feedback had significantly better results compared to similar initiatives.

Full Article
A Computational Guide to Leading Cells Down Desired Differentiation Paths
Wyss Institute at Harvard University
Benjamin Boettner
March 12, 2021


An international team of researchers has developed a computer-guided design tool that substantially boosts the efficiency of cell conversions by predicting effective combinations of cell type-specific transcription factors (TFs). The investigators combined the Integrative gene Regulatory Network model (IRENE) with a genomic integration system that enables robust expression of selected TFs in human-induced pluripotent stem cells. The approach generated more natural killer cells and melanocytes than other techniques, and produced breast mammary epithelial cells, which are highly desirable for the restoration of surgically excised mammary tissue. Donald Ingber at Harvard University's Wyss Institute for Biologically Inspired Engineering said, "This novel computational approach could greatly facilitate a range of cell and tissue engineering efforts at the Wyss Institute and many other sites around the world. This advance should greatly expand our toolbox as we strive to develop new approaches in regenerative medicine to improve patients' lives."

Full Article
Competitive Programming in Python
 
ACM Chapters
 

Association for Computing Machinery

1601 Broadway, 10th Floor
New York, NY 10019-7434
1-800-342-6626
(U.S./Canada)



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: [email protected]