Welcome to the March 4, 2022, 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."

Two people consulting a laptop computer. Ukraine's Vital Tech Industry Carries on Amid Russian Invasion
The Wall Street Journal
Sam Schechner
March 2, 2022


Many software developers in Ukraine continue to produce code for overseas clients amid the Russian invasion. Many also are volunteering for the ad hoc hacking army launching cyberattacks against Russia. Some Ukrainian technology companies are relocating employees to the west, donating money to the war effort, or offering office space as refugee housing, among other things. Said Tufts University's Bhaskar Chakravorti, "There is a serious talent crunch in IT, especially at the higher end where Ukraine was increasingly going. It's hard to imagine there will be too many other places for clients to go." Stepan Veselovskyi of the Lviv IT Cluster trade group said most tech companies in the city are working. Veselovskyi explained, "It's important for businesses with international clients to be alive and pay taxes and pay salaries to people in a time of war."

Full Article
*May Require Paid Registration

Biometric scanning devices at a newly opened Whole Foods Market. Here Comes the Full Amazonification of Whole Foods
The New York Times
Cecilia Kang
February 28, 2022


Amazon has almost completely automated a Whole Foods store in Washington, DC's Glover Park neighborhood. The store incorporates Just Walk Out technology, a network of cameras, sensors, and deep learning software that analyzes shopping habits. Shoppers can activate virtual shopping by scanning their palms at kiosks or by scanning quick response codes in the Amazon phone app. Just Walk Out detects when shoppers lift sensor-affixed products, itemizes their picks, and charges their Amazon account when they exit the store, skipping checkout lines. Amazon, which has tested such automation for over four years, plans to open a second prototype automated Whole Foods store in Los Angeles this year.

Full Article
*May Require Paid Registration
Surgeon General Demands Data on COVID-19 Misinformation From Major Tech Firms
The Hill
Brad Dress
March 3, 2022


U.S. Surgeon General Vivek Murthy reportedly has asked major technology companies to disclose data on COVID-19 misinformation. He asked for information about the prevalence and scale of the problem on the firms' Websites, and on social networks, search engines, crowdsourced and e-commerce platforms, and instant messaging systems. Murthy specified that the data should detail demographics impacted by misinformation, misinformation sources, and "exactly how many users saw or may have been exposed to instances of COVID misinformation." Said Murthy, "Technology companies now have the opportunity to be open and transparent with the American people about the misinformation on their platforms. This is about protecting the nation's health."

Full Article

Illustration of an eyeball surrounded by various computing devices. The Benefits of Peripheral Vision for Machines
MIT News
Adam Zewe
March 2, 2022


Massachusetts Institute of Technology (MIT) scientists have demonstrated a computer-vision model that perceives visual representations similar to human peripheral vision. The researchers investigating why adversarially trained neural networks are robust against image manipulation tasked study participants with distinguishing original images from noise-synthesized versions generated by a "normal" machine learning model, an adversarially robust model, and one called Texforms that accounts for certain aspects of human peripheral processing. MIT's Arturo Deza said the inability to tell original images apart from the adversarially robust model or the Texforms model implies that the former captures some peripheral vision aspects. Thomas Wallis at Germany's Technical University of Darmstadt said the researchers "propose that the same mechanism of learning to ignore some visual input changes in the periphery may be why robust images look the way they do, and why training on robust images reduces adversarial susceptibility."

Full Article
Move Over Candy Bars, New York Vending Machine Now Sells NFT Art
Reuters
Daniel Fastenberg
March 2, 2022


The first in-person non-fungible token (NFT) vending machine has been installed in New York City by digital art collecting platform Neon. The "NFT ATM," located in a small storefront in Lower Manhattan's financial district, sells QR codes connected to pieces of online art ranging in price from $5.99 to $420.49. Customers do not know which piece of digital art they have purchased until they scan the QR code, which allows them to display the art on any smartphone, laptop, or tablet. Neon's Kyle Zappitell said the target customer is "the crypto curious, the people who tried to buy cryptocurrency or they were interested in buying an NFT, but they just hit too many barriers."

Full Article

University of Illinois researchers Weichen Li, left, and Shelly Zhang. New Approach to Flexible Robotics, Metamaterials Design Mimics Nature, Encourages Sustainability
University of Illinois at Urbana-Champaign News Bureau
Lois Yoksoulian
February 28, 2022


Researchers at the University of Illinois Urbana-Champaign (UIUC) and the Technical University of Denmark have assembled multimaterial structures free of human intuition or trial-and-error to produce nature-mimicking flexible robots. They applied optimization theory and topology optimization, an algorithmic design process, to build composite structures that can support sophisticated prescribed mechanical responses. UIUC's Shelly Zhang and colleagues designed macroscale structures capable of swift stiffening, large-scale deformation buckling, multiphase stability, and long-lasting force plateaus. The resulting model devices, fashioned from two polydimethylsiloxane elastomers, are structured in a geometry that mimics a frog's legs. "We have designed reusable and fully recoverable energy dissipators, which is aligned with today's demand for sustainable devices that are good for the environment," said Zhang.

Full Article

Radar sensors for autonomous driving. Robust Radar: AI Sensor Technology for Autonomous Driving
Graz University of Technology (Austria)
Christoph Pelzl
February 23, 2022


An artificial intelligence (AI) system for automotive radar sensors developed by researchers at Austria's Graz University of Technology (TU Graz) filters out interfering signals from other radar sensors to improve object detection. The researchers built model architectures for automatic noise suppression based on convolutional neural networks (CNNs). To make them more efficient, the researchers trained the neural networks with noisy data and desired output values, then compressed the most efficient models further by reducing bit widths, resulting in an AI model with high filter performance and low energy consumption. Said TU Graz's Franz Pernkopf, "We want to make CNNs’ behavior a bit more explainable. We are not only interested in the output result, but also in its range of variation. The smaller the variance, the more certain the network is.”

Full Article

Logo of the Conti Ransomware Group. Conti Ransomware Source Code Leaked by Ukrainian Researcher
BleepingComputer
Lawrence Abrams
March 1, 2022


A Ukrainian researcher has exposed a wealth of content on the Conti cybercrime gang, including their ransomware's source code, after they sided with Russia on the Ukraine incursion. Known on Twitter as @ContiLeaks, the researcher leaked 393 JavaScript Object Notation files containing roughly 60,000 internal messages from the Conti and Ryuk ransomware group's private Extensible Messaging and Presence Protocol chat server. ContiLeaks then released more damaging material: the most exciting disclosure was a password-protected archive featuring the source code for the Conti ransomware encryptor, decryptor, and builder. Another researcher cracked the password, making the ransomware source code accessible to everyone.

Full Article

Greg Huff (left) and Deanna Sessions, who co-led a research team to create a methodology for diagnosing 3D printing errors. Researchers Establish Framework to Diagnose 3D-Printing Errors
Penn State News
Sarah Small
February 28, 2022


A multi-institutional team of scientists has developed the first framework for real-time diagnosis of three-dimensional (3D) printing errors through the use of machine learning. The researchers compiled images captured during the 3D printing process into a curated dataset that they could integrate with an algorithm to classify printing error types. Former Pennsylvania State University researcher Deanna Sessions said, "We're using this information—from cheap optical images—to predict electromagnetic performance without having to do simulations during the manufacturing process. If we have images, we can say whether a certain element is going to be a problem." The framework can spot errors in process when applied to the print, bringing real-time error correction closer to realization.

Full Article

A doctor consults with a patient. Preparing for the Next Pandemic
University of Waterloo News (Canada)
Wendy Philpott
February 28, 2022


An interdisciplinary research team at GoodLabs Studio and Canada's University of Waterloo is working on the Syndrome Anomaly Detection System (SADS) to provide real-time, data-driven insights that could help health officials prevent a future pandemic. The system uses machine learning (ML) and artificial intelligence to conduct widespread disease monitoring, with the goal of identifying patterns of atypical disease across communities to allow for quick action. SADS features an app that uses natural language processing (NLP) to capture symptoms described by patients in conversation with their doctors. The use of NLP within the app protects patients' personal information. Deep language ML is then used to aggregate and categorize the data—symptoms, age, gender, and location—to detect patterns. SADS tracks atypical symptoms over time to show how a novel disease could be spreading in a community.

Full Article

Hotspot maps use empirical data to assess protein binding sites. 'Hotspot Mapping' Accelerates Early-Phase Drug Design
Cambridge Crystallographic Data Centre (U.K.)
March 2, 2022


An automated, quantitative method developed by U.K. scientists harnesses "ensemble hotspot maps" to identify structural distinctions that can inform drug compound selectivity across protein families. Hotspot maps use empirical data to evaluate protein binding sites, yielding a three-dimensional grid to help rate and prioritize compounds. The script employed to produce the hotspot maps is a Python package that taps the Cambridge Structural Database through the Cambridge Crystallographic Data Center's IsoStar interaction library. Oxford University's Mihaela D. Smilova said hotspot maps enable researchers to quickly explore the chemical space, and can accelerate drug design "by summarizing the information and presenting it in a way that is both interpretable by medicinal chemists and can be used in further computational analyses."

Full Article

Encryption on a laptop. Researchers Can Steal Data During Homomorphic Encryption
NC State University News
Matt Shipman
March 2, 2022


Researchers at North Carolina State University (NC State) and Turkey's Dokuz Eylul University have cracked next-generation homomorphic encryption via side-channel attacks. Homomorphic encryption renders data unreadable to third parties, while still permitting third parties and third-party technologies to perform operations using the data. NC State's Aydin Aysu said the process consumes much computing power, and the researchers were able to read data during encryption by monitoring power consumption in the data encoder using Microsoft's SEAL Homomorphic Encryption Library. "We were able to do this with a single power measurement," Aysu noted, and the team confirmed the flaw in the library up through least version 3.6.

Full Article
Robot Photographer Has Eye for Aesthetics
Cornell Chronicle
Louis DiPietro
March 2, 2022


Cornell University researchers have developed AutoPhoto, a robotic photography system that can automatically rove an indoor space and shoot aesthetically appealing photos. The researchers said AutoPhoto employs a learned aesthetic machine learning (ML) model, and trailblazes new computer vision and autonomous photography applications by integrating current ML models with customized deep learning models. The system is based on a learned aesthetic estimation model trained on over 1 million human-ranked photos. Cornell's Hubert Lin said, "To guide the robot, we trained a separate model to move around in an environment and find a place that looks good." After scanning dozens of three-dimensional photos of indoor scenes and correctly selecting the best compositional angles, AutoPhoto explored a common space on a Clearpath Jackal robot; the team said the system roamed the space and shot three quality photos in a few minutes.

Full Article
The Handbook of Multimodal-Multisensor Interfaces - Volume III
 
ACM Discounts and Special Offers Program
 

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]