Welcome to the March 9, 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."
|
|
|
Gender Bias in Recruitment: AI Hiring Tools Are Hindering Women's Careers
EuroNews
Natalie Huet
March 8, 2022
New research found artificial intelligence (AI) is hurting women's careers, because of algorithms built on biased data. The United Nations Educational, Scientific, and Cultural Organization (UNESCO), the Organization for Economic Co-operation and Development, and the Inter-American Development Bank determined that job platforms generate fewer ads for high-paying jobs for users who select the female gender. UNESCO's Marielza Oliveira said recruitment and HR recommendation algorithms are based on historical data with baked-in bias against women. AI also tends to pair women with jobs demanding soft skills, which are not compensated as highly as those requiring hard skills, maintaining the gender pay gap. New systems are trying to address this by reconsidering language used in job ads. "We cannot have innovation by depreciating and disengaging half of the brains of the world," said Oliveira. "If we want AI to work, we really need to make it inclusive."
|
These Campaigns Hope 'Deepfake' Candidates Help Get Out the Vote
The Wall Street Journal
Timothy W. Martin; Dasl Yoon
March 8, 2022
South Korean presidential candidate Yoon Suk-yeol is reaching out to younger voters in online videos featuring a digital avatar (AI Yoon) that is more personable and relatable. AI Yoon, based on hours of video and audio recorded by the candidate along with his expressions and facial mannerisms, is controlled by his campaign team; Yoon's political opponents have adapted similar tactics with an avatar of rival candidate Lee Jae-myung, produced from real footage of his actual comments. Seoul-based artificial intelligence company DeepBrain AI provides the technology behind AI Yoon by synthesizing voice and video into avatars that hold down real-time dialogue. Some avatar videos can be produced in a half hour, while others take half a day, depending on what knowledge the avatar must express. Lee Jun-seok, head of Yoon's People Power Party, said Yoon has adapted his campaigning style based on the popularity of his avatar's videos.
*May Require Paid Registration
|
Tool Helps to Better Understand High-Mutating Viruses, Including COVID-19
Los Alamos Reporter
March 8, 2022
Scientists at the U.S. Department of Energy's Los Alamos National Laboratory (LANL) have developed FEVER, or Fast Evaluation of Viral Emerging Risks, a computational tool for detecting and investigating high-mutating viruses. Researchers use FEVER to design flexible measurement assays that concurrently identify whole classes of viruses for bio-surveillance, accurately diagnose an outbreak strain, and type mutations to find variants impacting public health. "We applied FEVER to COVID-19 and showed that we can indeed perform both highly specific SARS-CoV-2 diagnostics in 100 clinical samples while performing mutation typing for spike variants all at the same time," said LANL's Jessica Kubicek-Sutherland.
|
Scientists Translate Pig Grunts into Emotions for the First Time
The Daily Mail (U.K.)
Jonathan Chadwick
March 7, 2022
An international team of European scientists says it has translated pig grunts into emotions by training an artificial intelligence algorithm on 7,414 recordings of noises collected throughout the lives of pigs. The researchers claim the algorithm can decode whether an individual pig is experiencing positive emotions, negative emotions, or something in between. More rigorous analysis unearthed a new pattern that revealed what the animals experienced in certain situations, with clear differences in pig calls. "By training an algorithm to recognize these sounds, we can classify 92% of the calls to the correct emotion," said Elodie Briefer, associate professor at the University of Copenhagen in Denmark. Experts suggest sufficient data could be applied to better understand the emotions of other animals, which Briefer said could lead to applications that farmers could use to improve their livestock's well-being.
|
Spooky Action Could Help Boost Quantum Machine Learning
IEEE Spectrum
Charles Q. Choi
March 7, 2022
Researchers at Los Alamos National Laboratory and Louisiana State University suggest quantum entanglement, or "spooky action at a distance," might help clear a path toward quantum machine learning (ML) by overcoming the no-free-lunch theorem. The theorem posits that any ML algorithm is only as good as any other when their performance is averaged over many problems and training datasets. This implies that modeling a quantum system could require a volume of training data that must grow exponentially as the modeled system expands. Researchers suggested entangling extra or "ancilla" quantum bits with the quantum system that a quantum computer intends to model, enabling the quantum ML circuit to interact with many quantum states in the training data concurrently. They used quantum-hardware startup Rigetti's Aspen-4 quantum computer to confirm these findings.
|
Maps Show Airplane Contrails over U.S. Dropped in 2020
MIT News
Jennifer Chu
March 7, 2022
Maps of airplane contrails in 2020 charted by Massachusetts Institute of Technology (MIT) engineers showed an approximately 20% reduction in coverage area from 2018 and 2019, corresponding with reduced flights during the pandemic. Researchers used satellite images to compile a dataset of contrail-depicting pictures, then trained a computer-vision algorithm to differentiate contrails from other features. They ran the algorithm on about 100,000 satellite images, comprising some six trillion pixels that each represented a roughly 2-square-kilometer (1.2-square-mile) area and spanned the contiguous U.S. and parts of Canada and Mexico. MIT's Steven Barrett said the algorithm generated maps that mirrored identifiable flight routes, absent any knowledge of plane trajectories. "This kind of technology can help divert planes to prevent contrails, in real time," he suggested.
|
Mozilla Fixes Two Critical Firefox Flaws that Are Being Actively Exploited
ZDNet
Liam Tung
March 7, 2022
Mozilla has released security fixes for two critical flaws in its Firefox browser that are being exploited. The company said both CVE-2022-26485 and CVE-2022-26486 are critical use-after-free memory-related flaws, and the latter also could result in an exploitable sandbox escape. "Removing an XSLT parameter during processing could have led to an exploitable use-after-free,” the organization said. “We have had reports of attacks in the wild abusing this flaw. An unexpected message in the WebGPU IPC framework could lead to a use-after-free and exploitable sandbox escape." The company credits researchers at the Chinese security firm Qihoo 360 ATA with identifying the bugs. The security fixes were released with Firefox 97.0.2, Firefox ESR 91.6.1, Firefox for Android 97.3.0, and Focus 97.3.0, and fixed in Thunderbird 91.6.2.
|
Group Designs Rock-Tapping Remote-Control Robot to Detect Potential Slides
UNM Newsroom
Melanie Furber Fudge
March 2, 2022
Researchers at the University of New Mexico (UNM) have designed Brutus, a rock-tapping remote-controlled robot, to detect potential rockslides. Brutus uses a microphone and sound collector to locate damaged and destabilized rocks for roadside safety inspectors. An algorithm analyzes sounds generated by the tapping mechanism, then pinpoints instabilities; inspectors examine the algorithm's output to find and remove fall-prone rocks. Laboratory tests yielded the distinct properties of sounds produced by intact and cracked rocks, which were used to train and test the algorithm. Brutus has been successfully deployed in multiple field tests, including along a roadside in Tijeras, NM, where researchers "were able to classify, or distinguish, between the different regions [Brutus] hit on the roadside," said UNM graduate Roya Nasimi.
|
Bio-FlatScope Dives Deep for Useful Data
Rice University News
Mike Williams
March 3, 2022
The Bio-FlatScope developed by Rice University researchers is a small, lens-less camera for monitoring cellular-level biological details, and it can capture images beyond the capacity of microscopes. The device, which builds on the FlatCam precursor, channels light through a phase mask directly onto a camera sensor. Light filtered through the mask appears as a point spread function, which can be used to obtain details about objects below the diffraction limit, which few microscopes can visualize. Software reinterprets the data into an image that can be refocused at will to expose three-dimensional details, while the device's field of view eliminates distortion. Rice's Jacob Robinson said the Bio-FlatScope could eventually search for indications of cancer or sepsis or become an endoscopic tool.
|
Researchers Develop Tool to Make Irrigation More Efficient
Nebraska Today
Geitner Simmons
March 8, 2022
University of Nebraska-Lincoln (NU) researchers have devised a new irrigation methodology that promises to raise water efficiency and crop growers' profitability. The irrigation-scheduling algorithm factors in the level of atmospheric water concentration, and NU's Trenton Franz said this approach weighs both water supply and demand. The researchers tested the algorithm at 12 locations across Nebraska and found the tool could slash irrigation water use by 24% while maintaining crop yields, and boost profits by 11.2% and water productivity by 25.2%. The team's paper suggests this approach "could significantly improve water sustainability.”
|
Simulated Human Eye Movement Aims to Train Metaverse Platforms
Duke University Pratt School of Engineering
Ken Kingery
March 7, 2022
Virtual eyes developed by computer engineers at Duke University could be used to train virtual reality and augmented reality programs for the metaverse. The program, called EyeSyn, accurately simulates how humans look at the world while also protecting user data. The researchers relied on cognitive science literature on how humans see the world and process visual information to develop the program. Tests showed that EyeSyn closely matched the distinct patterns of actual gaze signals and simulated the different ways people's eyes react. EyeSyn eliminates the need for companies building platforms and software in the metaverse to collect data on how peoples' eyes move during various activities. "If you give EyeSyn a lot of different inputs and run it enough times, you'll create a data set of synthetic eye movements that is large enough to train a (machine learning) classifier for a new program," said Duke's Maria Gorlatova.
|
Project Icarus Is Creating Living Map of Earth's Animals
Popular Science
Charlotte Hu
March 8, 2022
Germany's Max Planck Society is leading a project to map the movements of terrestrial animals using data transmitted via a sensor network to the International Space Station (ISS). The International Cooperation for Animal Research Using Space (Icarus) project aims to generate an "Internet of animals" that can chart ecosystem changes and animals' responses to those shifts in real time. The researchers envision integrating information from the wildlife-worn sensors with other data on animal behavior across space, time, and ecosystems. The animal-affixed sensor tags will relay data whenever the ISS enters radio range, and the station's receiving computer processes and forwards the data to a ground control center, which transmits it to Icarus scientists. This information will be stored on the Movebank open online database, and the team has constructed movement profiles for certain animals in the system.
|
Can Tech Help Revive India's 'Crumbling' Health System?
BBC News
Priti Gupta; Ben Morris
March 7, 2022
India's struggling healthcare system, hit hard by the pandemic, could be at a turning point. "All aspects of access to healthcare, diagnostics, and life-sciences are moving...to low-cost and high-tech," said Invest India's Akriti Bajaj. There are an estimated 6,000-plus startups in India's healthcare sector. One startup, Dozee, has created a smart sensor that can be placed under patients' bedsheets to track micro-vibrations produced by the body and alert clinicians to unusual findings. Another firm, Niramai, has developed a small, portable screening device to detect early-stage breast cancer by measuring temperature variations in the chest using a high-resolution thermal sensor and analyzing the data using artificial intelligence. However, "…although [AI] is helping the medical fraternity in a great way, it cannot replace humans, ever," said Krishna Rajendra Hospital's Dr. Manjunth HG.
|
|
