Welcome to the March 2, 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."
As Tanks Rolled Into Ukraine, So Did Malware. Then Microsoft Entered the War.
The New York Times
David E. Sanger; Julian E. Barnes; Kate Conger
March 1, 2022


U.S. technology companies are helping to defend Ukraine against cyberattacks orchestrated alongside the Russian invasion. Shortly before the military incursion began, Microsoft's Threat Intelligence Center responded to previously unseen "wiper" malware targeting Ukraine's government ministries and financial institutions; the center dissected the malware, informed Ukraine's cyberdefense forces, and updated Microsoft's virus detection systems to block the code within hours. Meanwhile, Meta said it had locked down Facebook accounts of Ukrainian military officials and public figures when hackers attempted to spread disinformation through them. Corporate-government partnerships are being tested in the effort to analyze and counter Russia's cyberoffensive tactics, with tech companies a primary source of actionable intelligence.

Full Article
*May Require Paid Registration
UC Berkeley Robot Navigation Could Chart New Course for Self-Driving Systems
ZDNet
Tiernan Ray
February 23, 2022


Researchers at the University of California at Berkeley (UC Berkeley) have developed a system that allows a wheeled robot to travel kilometers over suburban terrain without mapping its environment. Using heuristics from 30 hours of video of previous runs and overhead maps of the terrain, the robot creates a better schematic of the relationship of stations along the journey without a full map. The system, called ViKiNG (Vision-Based Kilometer-Scale Navigation with Geographic Hints), provides Clearpath Robotics' Jackal unmanned ground vehicle with geographic hints, which permit ViKiNG to reach goals up to 25 times further away than the furthest goal reached by its predecessor system, and up to 15 times faster. UC Berkeley's Sergey Levine said, "I would characterize ViKiNG as a kind of reinforcement learning method with a higher level planner sitting on top of it."

Full Article
AI to Help Farmers See Cloud-Covered Fields
BBC News
Jonathan Amos
February 24, 2022


An artificial intelligence technique developed by researchers at U.K. startup Aspia Space aims to help farmers better manage their crops by giving them the ability to see their fields through cloud cover. The patented algorithm relies on two EU satellites: the Sentinel-1, which uses radar to map the ground and can see the surface through clouds and in darkness, and Sentinel-2, which views the ground at optical wavelengths. The algorithm was trained to transform radar reflections into synthetic optical images, so farmers can see their fields despite cloudiness or other weather activity. Aspia's Jim Geach said, "What we do is we just mimic all of the Sentinel-2 bands (colors). And then you can use those in exactly the same way you would if you had the real imagery."

Full Article
Singapore Tech Salaries Jump 22% in Chase for Skilled Coders
Bloomberg
Low De Wei; Olivia Poh
February 24, 2022


A report by Singapore technology career site NodeFlair and Asian venture capital firm Quest Ventures found that salaries for software engineers in Singapore rose an average 22% last year. Specialist roles like lead software programmers saw average pay raises of up to 32%, according to the report, which cited wide competition-fueled variance in salaries. Also causing salaries to spike were rising living costs in Singapore, while stricter worker visas complicated the search for less-senior foreign professionals. Yu Liuqing at the Economist Intelligence Unit said this tightening will likely compound staffing shortages, adding that smaller firms and local startups may be unable to sustain aggressive salary hikes for junior positions.

Full Article

Los Angeles Kings center Anze Kopitar (11) and Vegas Golden Knights center William Karlsson (71) vie for the puck after a faceoff during the first period of an NHL hockey game in Las Vegas. NHL Launching Faceoff Probability Stat with AWS Technology
Associated Press
Stephen Whyno
March 1, 2022


The National Hockey League (NHL) is launching a faceoff probability statistic, using Amazon Web Services (AWS) to apply 10 years of data and in-game analytics. AWS software generates the odds of each player in a faceoff winning possession of the puck. The Amazon Machine Learning Solutions Laboratory's Priya Ponnapalli said the model was based on data from thousands of faceoffs from the NHL's Hockey Information and Tracking System. The NHL's Dave Lehanski said hockey, unlike baseball, does not easily allow analysis of every single maneuver as a discreet event; the project with AWS aims to "demystify those plays and identify all those little things that come together to get fans to appreciate them and understand."

Full Article
Yale's Data Analysis Tool Uncovers COVID-19 Clues
YaleNews
Bill Hathaway
February 28, 2022


Yale University researchers developed a data analysis tool that identifies immune cell types associated with higher mortality risk from COVID-19. The Multiscale PHATE machine learning tool, built on the PHATE algorithm, enables scientists to pass through all resolutions of cellular data in minutes. Yale's Manik Kuchroo used the tool to analyze 55 million blood cells sampled from 163 patients admitted to Yale New Haven Hospital with severe COVID-19. The researchers associated TH17, a helper T cell, with higher mortality when bunched with immune system cells IL-17 and IFNG. They said they measured quantities of these cells in the blood to forecast whether a patient lived or died; those predictions, they said, were accurate 83% of the time.

Full Article
Deep Learning Technique Predicts Clinical Treatment Outcomes
MIT News
Lauren Hinkel
February 24, 2022


Massachusetts Institute of Technology (MIT) and IBM researchers have developed a deep learning predictive methodology that models counterfactual, time-varying, and dynamic clinical treatment strategies, enabling doctors to select the best treatment option for patients. The G-Net technique is based on the g-computation algorithm, a causal inference method that calculates the impact of dynamic exposures in the presence of measured confounding factors. G-Net applies recurrent neural networks, which feature node connections that allow them to better simulate temporal sequences with complex and nonlinear dynamics. MIT's Li-wei Lehman described G-Net as the first deep learning g-computation-based process "that can predict both the population-level and individual-level treatment effects under dynamic and time-varying treatment strategies."

Full Article

Nanyang Technological University’s Matteo Seita utilizes the setup to check the strength and toughness of a 3D-printed alloy sample. System Could Simplify Inspection of 3D-Printed Metal Objects
New Atlas
Ben Coxworth
February 24, 2022


Scientists at Singapore's Nanyang Technological University (NTU) have developed a system for inspecting three-dimensionally (3D)-printed metal objects for structural flaws more quickly and affordably. The researchers integrated an optical camera, a flashlight-like light source, and a notebook computer with machine learning-based software. Users chemically etch the surface of a 3D-printed item to expose the alloy's microstructure, then the camera photographs the object with the light source illuminating it from different angles. The software reads the microcrystals' shape, size, and orientation based on how they scatter the reflected light in the photos, a process that takes about 15 minutes. "Using our inexpensive and fast-imaging method, we can easily tell good 3D-printed metal parts from the faulty ones," said NTU's Matteo Seita.

Full Article
Fingertip Sensitivity for Robots
Max Planck Institute for Intelligent Systems (Germany)
February 24, 2022


A thumb-shaped sensor with a camera inside, developed by researchers at Germany's Max Planck Institute for Intelligent Systems (MPI-IS), uses computer vision and a deep neural network to infer haptic contact information. The Insight system generates a three-dimensional force map when the thumb is touched, estimating where objects come into contact with the sensor and the amount of force applied. Said MPI-IS's Huanbo Sun, "Our unique hybrid structure of a soft shell enclosing a stiff skeleton ensures high sensitivity and robustness. Our camera can detect even the slightest deformations of the surface from one single image."

Full Article
TU Delft Develops Car That Can 'Look into the Future' with Smart Eco Mode
TU Delft (Netherlands)
February 24, 2022


Researchers at the Netherlands' Delft University of Technology (TU Delft) collaborated with French automobile manufacturer Groupe Renault to develop a system to help drivers reach desired speeds more quickly, while maintaining energy efficiency. The Proactive Eco Mode system uses a simple algorithm to predict driver behavior. Said TU Delft's Timo Melman, "Our system allows your car to look into the near future while you are driving. This is possible because we make predictions about when a driver will need a lot of power and when not, and we change the car's engine settings if necessary. This does not require complicated algorithms: we simply do it by collecting data about how the driver and other road users drive. Thanks to this data, the car 'knows' when you want to accelerate." Groupe Renault indicated interest in implementing the system in future vehicles.

Full Article
AI Tool Detects Often-Overlooked Heart Diseases
Cedars-Sinai
February 23, 2022


Scientists at Cedars-Sinai's Smidt Heart Institute have developed an artificial intelligence tool that can detect and distinguish between two often overlooked heart conditions. The researchers used the algorithm to analyze roughly 34,000 cardiac ultrasound videos from Cedars-Sinai and Stanford Healthcare, and identified characteristics indicating certain patients might be exhibiting hypertrophic cardiomyopathy or cardiac amyloidosis. Smidt's David Ouyang said the algorithm spotted high-risk patients with greater accuracy than clinical experts, "because the algorithm picks up subtle cues on ultrasound videos that distinguish between heart conditions that can often look very similar to more benign conditions, as well as to each other, on initial review."

Full Article

A view of the 3D model of Red Hook Farm in Brooklyn, NY, featuring a processing shed and high tunnels for vegetable production. VR Farm Tour Expands Access to Urban Agriculture
Cornell Chronicle
Krishna Ramanujan
February 28, 2022


Cornell University researchers have created a virtual reality (VR) tour of New York City's Red Hook Farms. They created the virtual model using drones and photogrammetry. Users with a VR headset, a computer, or mobile phone, can "walk" around the farm and enter areas where farm managers lead demonstration and instructional videos on topics like cultivation, composting, and weeding. The experience is enhanced with ambient sound. Said Cornell's Tapan Parikh, "We're looking for some kind of happy medium between a situated embodied experience and the ability to convene people from across the state. We hope that's what [VR] provides us."

Full Article

David Boggs in 2021 holding a device from the 1980s that connected computers to a network. David Boggs, Co-Inventor of Ethernet, Dies at 71
The New York Times
Cade Metz
February 28, 2022


David Boggs, who helped invent Ethernet, the computer networking technology that links devices to the Internet in offices and homes, has died at 71. An ACM Fellow, Boggs co-designed the first version of Ethernet with Bob Metcalfe at the Xerox PARC research laboratory in the early 1970s. Ethernet stood out from other networking technologies like Arpanet, by connecting devices in proximity, rather than over long distances; after the technology evolved into an industry standard in the 1980s and 1990s, it became the reigning protocol for assembling networks in corporate offices. Modern wireless networks also borrow elements from the Ethernet protocol and frequently connect to Ethernet hardware. "Whatever you are reading this on—if you are not reading it on paper—you are probably using Ethernet," said Marc Weber at California's Computer History Museum.

Full Article
*May Require Paid Registration
ACM Digital Government: Research and Practice
 
ACM Online Books and Courses
 

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]