Welcome to the September 25, 2023, 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."
|
|
|
Cyber Experts Set Out to Secure 2024 U.S. Election
ComputerWeekly.com
Alex Scroxton
September 22, 2023
A team of cybersecurity experts established through the U.S.-based nonprofit Information Technology–Information Sharing and Analysis Center will undertake a "clear and concerted approach" to create a protocol for collaborative election security enhancement. The Election Security Research Forum's proposal involves convening security experts, nonprofits, companies, and former state and local election officials to strengthen U.S. electoral infrastructure. The forum has recruited three researchers to assess the cyber-resilience of new technology slated for testing in 40 states next fall, including digital scanners, ballot-marking devices, and electronic pollbooks. They will mainly concentrate on technology that Americans will use to vote in the 2024 presidential election. Participating researchers and companies have pledged to comply with Coordinated Vulnerability Disclosure policies and best practice, with all sides cooperatively evaluating and addressing new vulnerabilities uncovered while researchers and manufacturers will strive to mitigate or remove the risk or severity of confirmed flaws.
|
3D Model of an Exotic Supernova Took 5 Million Hours to Make
ScienceAlert
Nancy Atkinson
September 22, 2023
An international team of scientists has created the first three-dimensional (3D) hydrodynamic model of exotic supernovae, an effort that took 5 million hours and used supercomputers at the U.S. Department of Energy's Lawrence Berkeley National Laboratory and the National Astronomical Observatory of Japan. These resources enabled the researchers to simulate the effect of internal turbulent structures. "Turbulence plays a critical role in the process of a supernova explosion, resulting from irregular fluid motion, leading to complex dynamics," the researchers wrote. "These turbulent structures mix and distort matter, influencing the release and transfer of energy, thereby affecting the supernova's brightness and appearance." The researchers said current simulations have primarily been restricted to one-dimensional models.
|
Connection Machine Joins All to All to Optimize Better
IEEE Spectrum
Samuel K. Moore
September 23, 2023
University of Minnesota (UMN) researchers have developed an Ising machine that uses standard complementary metal-oxide semiconductor circuits to encode optimization problems onto a chip. The machine's connection scheme links all 48 magnetic spins to each other, using "nature to solve the problem," said UMN's Chris Kim. A series of interconnected inverter circuits form the core of the chip, arranged into 48 horizontally and vertically oriented oscillators. The oscillations' interactions emulate an Ising model moving to a lower energy state, answering the problem in a few microseconds. The researchers engineered the test chip with a 65-nanometer process, and the device consumed 105 milliwatts for the most densely connected problem it tackled while expending as little as 16 milliwatts on problems with meager links.
|
GPUs Are Here for Quantum
The Wall Street Journal
Isabelle Bousquette
September 21, 2023
Some companies are running quantum hardware on graphics processing units (GPUs) and other advanced chips used to power artificial intelligence (AI). Jack Hidary of the quantum software firm SandboxAQ said, "We don't have to wait for a quantum computer. We're not using a quantum computer, but we're using quantum equations, quantum software on GPUs. And that's a big breakthrough." GPUs can handle quantum algorithms due to their high bandwidth memory and dense mathematics capabilities. Nvidia's Timothy Costa explained, "It's a workload which is a great fit for GPUs for the same reasons that AI is a great fit for GPUs." Although quantum algorithms can be run on existing small-scale quantum computers, GPUs are viewed as an attractive alternative given that quantum computers are not yet advanced and have high error rates. Even when quantum hardware matures, Costa said GPUs could be used for error correction.
*May Require Paid Registration
|
Advances in Eye Scans, Protein Structure Win 2023 Lasker Awards
The New York Times
Noah Weiland; Cade Metz
September 21, 2023
This year's Lasker Awards recipients were honored for advancing eye disease diagnosis and cellular protein structure prediction. The awards are closely watched by researchers in biomedical fields and often foreshadow Nobel Prizes. Demis Hassabis and John Jumper at U.K. artificial intelligence laboratory DeepMind received the Albert Lasker Basic Medical Research Award for developing the AlphaFold algorithm to predict protein shapes from amino acid strings in minutes. DeepMind open-sourced AlphaFold for scientists worldwide, and by 2022 had released predictions for nearly every protein known to science, which exceeded 200 million proteins across a million species. The Lasker-DeBakey Clinical Medical Research Award, meanwhile, was given to a team of three scientists who helped invent optical coherence tomography, which can detect conditions like macular degeneration and diabetic retinopathy earlier than previous methods.
*May Require Paid Registration
|
Using Machine Learning to Close Canada's Digital Divide
Waterloo News (Canada)
September 20, 2023
Researchers at Canada's University of Waterloo and the National Research Council developed the machine learning-based Multivariate Variance-based Genetic Ensemble Learning Method to anticipate potential satellite problems to ensure uninterrupted Internet access for rural and remote Canadians. The method combines several artificial intelligence-driven models to identify anomalies in satellites and satellite networks before they can escalate. The researchers tested their model on the publicly available Soil Moisture Active Passive, Mars Science Laboratory Rover, and Server Machine datasets. They found the model's accuracy, precision, and recall surpassed that of existing models. Waterloo's Peng Hu said, "This research will help us to design more reliable, resilient, and secure satellite systems."
|
Why NASA Is Sending National Secrets to the Moon
BBC Science Focus
Noa Leach
September 22, 2023
The U.S. National Aeronautics and Space Administration (NASA) will work with Florida-based computing startup Lonestar and the self-governing British Crown Dependency Isle of Man to send a data payload to the Moon next February to assess lunar-based backup storage as part of the Artemis program. The collaborators hope to ensure the data's security and protection from tampering using blockchain while also demonstrating the stored information's authenticity. After landing, the researchers intend to digitally "frank" the data cube's payload on Lonestar's datacenter to prove its lunar provenance, then have it transmitted back to Earth and compiled into a blockchain to signal its verification. The information to be digitalized and launched on the data cube is stamps chosen by the Isle of Man's post office.
|
Toyota Conceives of More Efficient Method to Train Robots
Interesting Engineering
Loukia Papadopoulos
September 19, 2023
The Toyota Research Institute (TRI) has introduced a generative artificial intelligence (AI) method for training robots to perform more dexterous behaviors more efficiently. The robot behavior model incorporates haptic teacher demonstrations and spoken descriptions of objectives, enabling new behaviors inferred from many demonstrations to be introduced independently. This strategy produces reliable, reproducible, and efficient results rapidly. TRI has already used the new approach to train robots to perform more than 60 dexterous tasks simply by providing them new information. The researchers hope this model will enhance human-robot cooperation.
|
Software Detects Money Laundering Faster Than Ever Before
King's College London (U.K.)
September 22, 2023
Computer scientists at the U.K.'s King's College London (KCL) and CWI and Vrije Universiteit Amsterdam in the Netherlands have developed software that can detect money laundering faster and with greater accuracy than previously. The algorithms identify when criminals are "smurfing," or splitting up large sums of money into smaller transactions between numerous bank accounts. The algorithms process data derived from several accounts that represent nodes on a graph, with the software focusing on the graph area where suspicious activity is concentrated. KCL's Grigorios Loukides said the software "can find the best possible solution for detecting common classifications of smurfing attacks across millions of amounts of data, on average 3.2 times more effectively than the state-of-the-art methods currently used." Loukides added that the tool's rapid data analysis capability can enable money laundering experts to zero in on bad actors efficiently.
|
Technique for 3D Printing Metals at Nanoscale Reveals Surprise Benefit
Caltech News
Emily Velasco
September 20, 2023
California Institute of Technology (Caltech) researchers have repurposed a method for three-dimensionally (3D) printing thick metal components for fabricating 150-nanometer-sized parts. Although these materials would be unusable at large scale because of disordered atomic arrangements, the researchers found this same property confers nanoscale objects like metal pillars up to five times as much strength as more orderly configured structures of similar size. Caltech's Julia R. Greer said the defects in the atomic structure inhibit the propagation of catastrophic failure across the nanoscale object's grain boundaries, distributing deformations more uniformly throughout the material. Greer thinks this process could find use in building practical elements, including catalysts for hydrogen and storage electrodes for carbon-free ammonia.
|
Expanding the VR Immersion Comfort Zone
SPIE
September 19, 2023
Researchers at Taiwan's Innolux Corp. demonstrated the use of high-resolution liquid crystal displays (LCDs) to address Vergence-Accommodation-Conflict, which can occur during immersive virtual reality (VR) experiences. The researchers used a 3.1-inch 3k3k LCD to overcome light field resolution issues, such as constrained viewing angles and screen window effects. They broadened the binocular field of view in the new INNOLUX LCD with a 15-degree tilt between panels and reproduced high-quality images by employing the Modulation Transfer Function across the image field. Additionally, the researchers introduced "corrected eyebox mapping," a ray tracing-based graphical process that accounts for spherical power, cylinder power, and cylinder axis to help correct myopia, hyperopia, and astigmatism. Said Innolux's Yung-Hsun Wu, "By utilizing light field technology, both vision correction and the expansion of the eyebox are achieved, thereby elevating the overall [VR] experience and enhancing user comfort."
|
Quantum Algorithm Offers Faster Way to Hack Internet Encryption
Science
Anna Kramer
September 19, 2023
New York University's Oded Regev has proposed a quantum algorithm designed to crack Internet encryption faster than Massachusetts Institute of Technology (MIT) mathematician Peter Shor's algorithm. Shor’s algorithm has endured as an example of the promise of quantum computers for 30 years. Regev's scheme can presumably factor very large numbers using significantly fewer gates, enabling a smaller quantum system to exfiltrate secret encryption keys or a larger system to decrypt them faster. Shor's algorithm looks for prime factors by elevating a single number to high powers, with results yielded only by multiplying big numbers together; Regev's model keeps these numbers from becoming as large before yielding results by multiplying several numbers in different dimensions. Regev found only n1.5 gates are needed to factor an n-bit integer, which MIT's Vinod Vaikuntanathan calls the first significant improvement on Shor's algorithm in three decades.
*May Require Paid Registration
|
|
