Welcome to the October 19, 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."
|
|
|
ML-Based Solution Could Help Firefighters Circumvent Deadly Backdrafts
U.S. National Institute of Standards and Technology
October 17, 2022
Scientists at the U.S. National Institute of Standards and Technology have formulated a machine learning-based model to predict potentially deadly backdrafts. The researchers based the model on data from hundreds of laboratory-engineered backdrafts, with the hope firefighters will deploy it to avoid or adjust to hazardous circumstances. The team initially fed the model data on gas levels, fuel richness, and temperature measured at one location in the lab chamber before a door was opened, to calculate the odds of a backdraft occurring. The model predicted backdrafts correctly in 70.8% of the experiments; adding measurements at a second location boosted its accuracy to 82.4%.
|
South Korea Aims to Boost Economy with Digital IDs on Blockchain
Bloomberg
Sam Kim
October 16, 2022
South Korea intends to offer its citizens a blockchain-secured digital identity in order to improve economic growth. The nation will launch digital IDs in 2024, expecting 45 million citizens to be using them within two years. The IDs will be embedded into mobile devices like smartphones. Suh Bo Ram, director-general of South Korea's digital-government bureau, said a decentralized identity framework will prevent the government from accessing data on individual phones, including whose digital IDs are used, how they are used, and where. Hwang Seogwon at South Korea's Science and Technology Policy Institute said although digital IDs can be highly beneficial economically, "there has to be more risk assessment technologically to make sure the danger doesn't outweigh the benefits."
*May Require Paid Registration
|
Would You Like a QR Code in That Cookie?
AlphaGalileo
October 17, 2022
A three-dimensional (3D) printing method developed by scientists at Japan's Osaka University can embed edible quick response (QR) codes inside food. "Many foods can now be produced using 3D printers," explained Osaka University's Yamato Miyatake. "We realized that the insides of edible objects such as cookies could be printed to contain patterns of empty spaces so that, when you shine a light from behind the cookie, a QR code becomes visible and can be read using a cellphone." The "interiqr" technique produces a QR tag made from the cookie itself, keeping taste, flavor, and outward appearance intact.
|
Improving Access to Electric Vehicle Charging Stations
NC State University News
Matt Shipman
October 17, 2022
North Carolina State University (NC State) researchers aim to make electric vehicles more appealing to drivers by improving access to charging stations via a game theory-based computational tool. The tool uses a game theory framework to balance factors such as the length of time required to reach a charging station, the cost of using it, wait times for access, and the potential for fines for exceeding charging time limits. The method helps users find the nearest charging station, and incorporates a dynamic system that charging station operators can use to ascertain how long vehicles can spend charging before they must make way for the next vehicle.
|
Deep Learning Makes X-Ray CT Inspection of 3D-Printed Parts Faster, More Accurate
Oak Ridge National Laboratory
S. Heather Duncan
October 14, 2022
Researchers at the U.S. Department of Energy (DOE) Oak Ridge National Laboratory (ORNL) have developed a deep-learning framework that uses X-ray computed tomography (CT) to improve the speed and accuracy of inspecting three-dimensionally (3D) printed metal parts. The framework is being incorporated into software commercial partner ZEISS uses in its equipment at DOE’s Manufacturing Demonstration Facility at ORNL, where companies get help perfecting their 3D printing. ZEISS's Pradeep Bhattad said, "With this, we can inspect every single part coming out of 3D-printing machines. Currently CT is limited to prototyping. But this one tool can propel additive manufacturing (3D printing) toward industrialization."
|
Targeted Billboard Ads Are a Privacy Nightmare
Gizmodo
Mack DeGeurin
October 13, 2022
A report from the U.K. civil liberties group Big Brother Watch highlights privacy concerns related to targeted digital advertising on physical billboards. These billboards use facial recognition software to generate ads personalized to passers-by. Big Brother Watch's Jake Hurfurt calling the practice "some of the most intrusive advertising surveillance we've ever seen in the U.K." The report raised concerns about the mass collection of user data, including precise GPS location, gender and age demographics, and behavioral data, which it said "is being gathered not just to work out if an ad campaign was successful, but to alter how people experience reality without their explicit consent, all in an attempt to make more sale."
|
Goalkeeping Robot Dog Tends Net Like a Pro
IEEE Spectrum
Evan Ackerman
October 16, 2022
The Mini Cheetah robotic goalie developed by scientists in the University of California, Berkeley's Hybrid Robotics Lab can save 87.5% of shots taken on goal, compared to almost 80% achieved by players in the English Premier League. The researchers designed the quadrupedal robot to integrate a locomotion controller with a planner for the end-effector trajectory to find the best way to maneuver itself in front of the ball, in less than a second. They trained Mini Cheetah on goalkeeping skills such as sidestepping for intercepts nearby and close to the ground; diving to reach the goal's lower corners, and jumping to cover the top of the goal and the upper corners.
|
AI Language Models Show Bias Against People with Disabilities, Study Finds
Penn State News
Jessica Hallman
October 13, 2022
Pennsylvania State University (Penn State) researchers found that natural language processing models often are biased against people with disabilities. The researchers studied 13 popular machine learning models trained to generate sequences of words, and tested over 15,000 unique sentences on each model to produce word associations for over 600 adjectives that could be associated with individuals with or without disabilities. The researchers assessed the sentiment of each adjective generated as positive, negative, or neutral, finding that sentences with disability-related words scored more negatively than sentences lacking them. Penn State's Pranav Venkit said the work demonstrates "that people need to care about what sort of models they are using and what the repercussions are that could affect real people in their everyday lives."
|
Sensor System for IoT Devices Integrates Processing, Computing to Save Energy, Protect Data
University of Virginia Engineering
Karen Walker
October 11, 2022
The University of Virginia's Kyusang Lee and colleagues are exploring a field Lee calls the artificial intelligence of things with the development of a smart sensor that stores and processes data. The smart sensor will sit at the edge of a device, which itself sits at the outer reaches of a wireless network. It can detect and process diverse signals that emulate human biology, including imagery for vision; ultrasound for hearing; pressure and strain linked to motion and touch; and virus detection via chemical sensing. Said the University of Virginia’s Kyusang Lee, “I believe our sensor will be especially useful in robotics that rely on combined sensory inputs and real-time integrated processing.”
|
Announcing ICU4X 1.0
The Unicode Consortium
September 29, 2022
The Unicode Consortium recently published ICU4X 1.0, a lightweight, portable, secure Unicode library to help devices, use cases, and programming languages tap internationalized software. The organization describes ICU4X as the first library to enable static data slicing and dynamic data loading, to support multiple programming languages out of the box, and to ensure memory- and thread-safety via a Rust-facilitated type system and ownership framework. The consortium created ICU4X to replace many non-Unicode i18n libraries designed to incorporate the industry-standard repository into new coding languages and resource-limited environments, but which are rarely maintained and often incomplete. The developers made ICU4X lightweight by implementing small binary size, low memory usage, and deliberate performance optimizations.
|
Researchers Take Step Toward Quantum Computing
University of Bristol News (U.K.)
October 11, 2022
Researchers at the U.K.'s University of Bristol, its quantum startup Phasecraft, and Google Quantum AI have developed the first truly scalable algorithm for observing ground-state properties of the Fermi-Hubbard model on a quantum computer. The researchers used the algorithm and improved error-mitigation methods to run an experiment that is four times larger and uses 10 times more quantum gates than any previously recorded efforts. Bristol's Ashley Montanaro described the experiment as "particularly exciting because it suggests that we will be able to scale our methods in order to leverage more powerful quantum computers as the hardware improves."
|
Nuclear Crossing Guard
Texas Advanced Computing Center
Jorge Salazar
October 12, 2022
University of Illinois at Urbana-Champaign (UIUC) researchers used the Texas Advanced Computing Center's Frontera and Stampede2 supercomputers to develop a computer model of nuclear pore complexes in the nuclei of living cells. The model simulates apertures in a cell's nuclear membrane that regulate the passage of biomolecules into and out of the nucleus. The researchers applied brute force simulations to analyze the kinetics of nuclear pore transport over tens of milliseconds. Former UIUC researcher David Winogradoff said the model showed the nuclear pore's mesh-like interior structure hardens or softens in response to protein size. He envisions this research providing guidelines for the development of therapeutics that deliver drugs to cell nuclei.
|
Smelling in VR is Possible with Gaming Technology
Stockholm University (Sweden)
October 12, 2022
Scientists at Sweden's Stockholm and Malmö universities have created a three-dimensionally (3D)-printable "olfactometer" that can generate odors in virtual reality via a gaming controller. The researchers deployed the scent machine in a game where players guess types of wines in a virtual cellar by their aromas. When the players lift virtual glasses, the olfactometer attached to the controller discharges an odor. The olfactometer has four valves that are each connected to a channel, which under the player's control open to a different extent and release unique fragrance mixtures. Said Malmö University's Simon Niedenthal, "The possibility to move on from a passive to a more active sense of smell in the game world paves the way for the development of completely new smell-based game mechanics based on the players' movements and judgments."
|
|
