Welcome to the July 26, 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."
|
|
|
Open Source Platform for Neural Radiance Field Development
Berkeley Engineering
Marni Ellery
July 25, 2023
University of California, Berkeley (UCB) researchers have created Nerfstudio, a Python-based open source development platform for accelerating Neural Radiance Fields (NeRF) projects. NeRF technology can convert flat images into three-dimensionally (3D)-navigable scenes, and Nerfstudio supplies plug-and-play components for deploying NeRF-based techniques. UCB's Angjoo Kanazawa explained the platform "is intended to simplify the development of custom NeRF methods, the processing of real-world data, and interacting with reconstructions." Roboticists, gaming studios, and news outlets that work with computer graphics already are producing 3D reconstructions in real-world environments with Nerfstudio.
|
Researchers Find Deliberate Backdoor in Police Radio Encryption Algorithm
Ars Technica
Kim Zetter
July 25, 2023
Researchers with Netherlands-based security consultancy Midnight Blue have uncovered a secret backdoor in technology long used for critical data and voice radio communications worldwide. The backdoor resides in an algorithm embedded within commercially sold devices that transmit encrypted data and commands, allowing users to eavesdrop on communications and potentially hijack critical infrastructure. The researchers found the backdoor and four other flaws in the European Telecommunications Standards Institute's Terrestrial Trunked Radio (TETRA) standard in 2021, but waited until radio manufacturers could develop patches and mitigations before disclosing them. The researchers also learned most police forces worldwide (excluding the U.S.) use TETRA-based radio technology.
|
Cleaning Up with TidyBot
Princeton Engineering News
Julia Schwarz
July 24, 2023
A multi-institutional team of engineers mated a mobile robotic arm to a vision model and a large language model to create the TidyBot cleaning robot. The robot's arm is attached to a wheeled base; its pincer hand can open drawers and retrieve objects, enabling it to carefully put dishes in the sink or toss clothes onto a laundry pile. TidyBot can differentiate between types of objects with its combined camera/vision model, and the large language model with which it is programmed allows it to learn how carefully or casually it can handle different items. The robot was 85% accurate in categorizing objects and where to put them.
|
How Hacking Honeybees Brings AI Closer to the Hive
IEEE Spectrum
Sarah Wells
July 21, 2023
Computer scientists at the U.K.'s University of Sheffield have developed a new form of decision-making machine intelligence by analyzing the brains of honeybees. The researchers monitored 20 bees as they probed color-coded flowers, determining their accurate and faster decision-making compared to animals and artificial systems "exhibits a level of intricacy that parallels certain aspects of decision-making seen in higher animal species," according to Sheffield's HaDi MaBouDi. The researchers created a bee-like model with acceptance and rejection decision pathways that weigh the quality of stimuli to arrive at decisions while retaining past stimuli information to recall irrelevant stimuli. The model's response rates were similar to those of bees when presented with 25 scenarios of random high-reward and low-reward stimuli.
|
Uncapping the Power of Knee Implant Placement Software
Texas A&M University Engineering
Grace Dalton
July 20, 2023
Scientists at Texas A&M University (TAMU), New York University Langone Health, and surgical navigation and robotics company Caira Surgical validated a computational knee model to help surgeons better orient implants during knee replacement. The model enables doctors to review an implant's position for optimal accuracy and evaluate potential misalignment prior to final placement, as well as simulating the implant's anatomical interaction and projecting post-surgical knee motion and possible mobility alterations. The researchers' initial validation entailed three-dimensionally printing a physical knee model, which TAMU's Aaron Henry compared to the computational model via motion capture. The researchers found the computational model reasonably matched the physical model's real-world performance.
|
Construction Company Turns to Wearable Tech to Beat the Texas Heat
The Dallas Morning News
Irving Mejia-Hilario
July 17, 2023
Dallas-based contractor Rogers-O'Brien Construction aims to counter exceedingly hot temperatures across Texas via technology intended originally for a manned space mission. Workers at a Southern Methodist University construction site wear arm sensors that continuously track various biometric indicators, alerting them and safety supervisors to anomalous readings. The sensors, installed in Polar Verity Sense armbands and Garmin watches, link to Safeguard software from Sentinel Occupational Safety. Extreme readings by the sensor cause the danger-level color of the armband’s biometric indicator to change, while sending an alert to the supervisor's phone.
|
Error Rate Reduced for Scalable Quantum Technology
Physics
July 24, 2023
A multi-institutional team of researchers demonstrated a scalable system for controlling quantum bits (qubits) with a low error rate in a step toward building practical quantum computers. The system controls conventional superconducting qubits with flux quanta, reducing the error rate by keeping the control circuits and the qubits on separate chips. The researchers have achieved more than 99% accuracy with this separation, which the University of Wisconsin-Madison's Chuan-Hong Liu said reduces interference between the flux-quanta generator and the qubits. The team found its flux generators produced a 1.2% error rate, versus a 2.1% error rate detailed by another team last year.
|
App Developed for/with People with Intellectual/Developmental Disabilities
Rhody Today
July 12, 2023
Researchers at the University of Rhode Island and the Massachusetts Disabled Persons Protection Commission developed an application that trains adults with intellectual and developmental disabilities to identify and report abuse. The researchers collaborated with intellectually and developmentally disabled consultants to design the free Recognize, Report, and Respond (R3) app, which is available via the Apple and Amazon app stores for smartphones and tablets. The consultants' insights prompted the developers to shift R3's focus from reporting abuse to education, based on training programs created by the protection commission. R3 integrates simple text, images, videos, and interactive follow-up activities to help users recognize, report, and respond to abuse.
|
New Cryptocurrency Offers Users Tokens for Scanning Their Eyeballs
The Guardian (U.K.)
Hibaq Farah
July 25, 2023
OpenAI CEO Sam Altman has launched the Worldcoin cryptocurrency scheme to differentiate "verified humans" from artificial intelligence (AI) systems via biometric scanning. Participants launching a new account have their iris scanned in exchange for a "genesis grant" of 25 tokens, equivalent to roughly £40 ($51.66); the World ID they receive will verify they are a "real and unique person," according to the scheme. Users also will be able to make payments, purchases, and transfers globally using digital assets and traditional currencies via the World application. Two million users from 33 countries, mainly in Europe, India, and southern Africa, have enrolled and been scanned on the service, which officially launched this week.
|
Wearable Sensor Sets Record for Solar Power Efficiency
Caltech News
Cynthia Eller
July 20, 2023
Researchers at the California University of Technology (Caltech) and Austria's Johannes Kepler University Linz have developed wearable solar-powered biosensors with record-breaking efficiency. The sweat sensor's flexible perovskite solar cell (FPSC)'s power conversion efficiency tops 31% under indoor illumination, compared to silicon solar cells' 26% to 27%. Caltech's Wei Gao designed the sensors to be worn for 12 hours daily and continuously track pH, salt, glucose, and temperature, while monitoring sweat rate every five to 10 minutes. The device's components allocate electricity collected by the FPSC, induce sweating and electrochemically measure the sweat's contents, and process data and communicate wirelessly with a cellphone application to display readings.
|
Impersonation Detection Feature Safeguards Brands, Personalities from Fake Accounts on Social Media
Viral Nation
July 25, 2023
Canada-based social media marketing transformation, talent, and technology company Viral Nation announced the release of an impersonation protection feature for its Viral Nation_Secure social media monitoring platform. The enhancement can shield brands, personalities, and society at large from fake accounts. The feature can detect bogus accounts with unmatched accuracy using proprietary multi-artificial intelligence analysis that includes recognition and detection cognition, and deep analysis. It can perform continuous, concurrent searches across various platforms, labeling and uncovering false accounts while disclosing takedowns to the platforms to protect personal brands and companies' online presence. The feature also can identify unsafe social media content like profanity, hate speech, and violence across images, videos, and text, and detect audio copyright infringement.
|
Physics-Informed Supervised Learning Framework Could Make Computational Imaging Faster
Optica
July 25, 2023
A new physics-informed variational autoencoder (P-VAE) framework developed by Swarthmore College researchers could accelerate computational imaging via supervised learning. The framework uses sparse measurements to capture and jointly reconstruct each light source in an image, deducing prior and posterior distributions by pooling information from dataset-spanning measurements and incorporating established data about the forward physics of imaging. The researchers enhanced light-emitting diode (LED) array microscopy with P-VAE, shortening the acquisition time by reducing the number of images required per object. They also used only sparse measurements to reconstruct objects imaged by computed tomography through the framework.
|
Robotically Manufactured Timber Construction as a Model for Resource-Efficient Construction
University of Freiburg (Germany)
July 18, 2023
Scientists at Germany's universities of Freiburg (Uni Freiburg) and Stuttgart built a pavilion from lightweight timber using new computer-based planning techniques, robotic fabrication and construction processes, and human-machine interaction processes. Uni Freiburg's livMatS Biomimetic Shell @ FIT incorporates the construction principles of the sea urchin’s skeleton, with lightness and strength derived from individually arranged plates. German construction company müllerblaustein HolzBauWerke fabricated the pavilion's components on a transportable robotic platform, using augmented reality to combine manual partial assembly steps for special elements. Stuttgart's Achim Menges said this approach "enables the effective, digitally controlled production of complex components with a high degree of precision." The pavilion's skylight incorporates a weather-sensitive "solar gate" built from four-dimensionally-printed materials to regulate interior temperatures by passively adapting to solar conditions.
|
|
