Welcome to the September 6, 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."
|
|
Animal Motion-Capture Studio Tracks Bird Flocks, Insect Swarms
New Scientist Jeremy Hsu September 1, 2023
The SMART-BARN motion-capture studio built by researchers at Germany's Max Planck Institute of Animal Behavior (MPI-AB) and the University of Konstanz can monitor the movements and behaviors of bird flocks and insect swarms. The laboratory boasts 30 infrared cameras that can track as many as 500 individual markers attached to animals within a space 25% as large as a standard basketball court. The SMART-BARN uses six video cameras and computer vision software to track unmarked animals, while microphones capture animal noises and locate animals by sound. MPI-AB's Iain Couzin said this setup enables close observation of predator-prey interactions, and of animal group dynamics involving leadership, communication, and cooperation.
|
Analog and Digital: The Best of Both Worlds in One Energy-Efficient System
EPFL News (Switzerland) Michael David Mitchell September 6, 2023
Researchers at Switzerland's École Polytechnique Fédérale de Lausanne (EPFL) have developed a device that combines ultra-thin two-dimensional semiconductors and ferroelectric materials, with the goal of integrating digital logic and brain-inspired analog operations. The semiconductors enable ultra-efficient digital processors, while the use of ferroelectric materials allows for the simultaneous processing and storage of memory. EPFL's Adrian Ionescu said, "The research marks the first-ever co-integration of von Neumann logic circuits and neuromorphic functionalities, charting an exciting course toward the creation of innovative computing architectures characterized by exceptionally low power consumption and hitherto unexplored capabilities of building neuromorphic functions combined with digital information processing."
|
AI Rivals Human Nose When Naming Smells
Science Elizabeth Pennisi August 31, 2023
Researchers at artificial intelligence (AI) company Osmo, working with colleagues at Philadelphia’s Drexel University and the Monell Chemical Senses Center, developed a graph neural network that reliably matched human volunteers' identification of 55 odors, then predicted the smells of 500,000 additional molecules without having to produce or sniff them. The researchers fed the structures and odor descriptions of 5,000 molecules to an AI to teach it to identify patterns in the training data by correlating a molecule's odor with attributes of its underlying atoms. After calculating average human odor identification ratings, the researchers found the neural network got closer to this average than any individual in the volunteer group did in over half the cases. The AI then deduced how the 500,000 hypothetical chemical structures should smell.
*May Require Paid Registration
|
Software Means Biomedical Researchers Don't Have to Be Computer Scientists
Georgia Tech College of Engineering Joshua Stewart August 31, 2023
Scientists at the Georgia Institute of Technology and Emory University created open source software that enables any researcher with imaging data to perform sophisticated biomedical imaging analysis without the need for coding skills. The interactive Cellular assay Labeled Observation and Tracking Software (iCLOTS) program adapts current well-validated imaging algorithms, with its designers focused mainly on microfluidics platforms. The team subjected iCLOTS to repeatability and sensitivity tests by applying it to various published datasets and compared its analysis to the studies' results and to manual outcomes compiled by humans reviewing images or videos. The software exposed insights in data from projects that might otherwise have been overlooked during validation testing.
|
Cracking the Puzzle of Serpentine Motion
IEEE Spectrum Rahul Rao September 5, 2023
Researchers at the California Institute of Technology and Germany's Technical University of Berlin have developed an algorithm that can be used to simulate, animate, and predict serpentine motion. The algorithm could be used to help animators create virtual models of snakes and similar animals, and paves the way for robots that move in different ways. The algorithm looks at shapes comprised of vertices and identifies the most energy-efficient means for them to rotate or translate, taking into account both gravity's pull and the material through which a body is moving. For instance, the algorithm uses Helmholtz's principle for bodies moving through viscous fluid or water, and Euler's principle of least action for bodies moving through a vacuum or air.
|
New Route to a Quantum Internet
Princeton University Scott Lyon August 30, 2023
Princeton University researchers have developed a method for connecting quantum devices over long distances that involves sending telecom-ready light emitted from a single ion implanted in a crystal. The rare earth ion used in the new device does not require signal conversion because it emits light at an optimal infrared wavelength. The device is comprised of a calcium tungstate crystal containing erbium ions that emit light through the crystal when pulsed with a special laser, and a nanoscopic piece of silicon carved into a J-shaped channel located on top of the crystal that captures and moves individual photons encoded with information from the ion into the fiber optic cable. The researchers observed up to 80% suppression of individual photons at the interferometer output, which Princeton's Salim Ourari said demonstrated the signal was significantly above the hi-fi threshold.
|
Can We Talk to Whales?
The New Yorker Elizabeth Kolbert September 4, 2023
Researchers with the Cetacean Translation Initiative (CETI) are leveraging machine learning to help decipher whale codas, the series of clicks that they use to talk to each other, and maybe allow humans to speak with them as well. The researchers plan to attach recording devices to sperm whales near Dominica to collect data to train machine learning algorithms. They also plan to record codas using three "listening stations" tethered on the floor of the Caribbean Sea. About 25 codas have been detected among the sperm whales around Dominica, differing in the number and rhythm of clicks. Shane Gero of Canada's Carleton University has assembled an archive of sperm-whale codas containing around 100,000 clicks, but the CETI researchers estimate that about 4 billion clicks ultimately will be needed.
|
Origami-Inspired Strain Sensors for Stretchable Soft Robotics
USC Viterbi School of Engineering Matilda Bathurst August 29, 2023
The University of Southern California's Hangbo Zhao and colleagues created a strain sensor for stretchable soft robots using origami-inspired three-dimensional (3D) electrodes capable of measuring a strain range up to three times higher than typical sensors. Users can attach the sensors to soft bodies to monitor shape-change and appropriate functioning as they move, without cameras. Zhao said the electrodes' 3D structures apply origami principles to enable repeated use of the sensors by converting their stretch and release to a process of unfolding and folding, "and to give precise readings even when measuring large and dynamic deformations of soft bodies." He added that integrating the electrodes with a stretchable substrate "allows us to measure a very large deformation, as much as 200% strain, with an ultra-low hysteresis of around 1.2%" while supporting a response within 22 milliseconds.
|
Wanted: Skilled Workers to Combat Rise in Cyber Crime
Financial Times Hannah Murphy September 4, 2023
The International Information System Security Certification Consortium (ISC2) estimated there were roughly 4.7 million people in the global cybersecurity workforce last year, but CEO Clar Rosso cited a shortfall of 3 million to 4 million. Cybersecurity job market website Cyberseek found just 69% of cyber roles are filled in the U.S., while a 2023 U.K. government report observed a basic cyber skills gap among half of U.K. businesses. Roy Zur at cybersecurity and digital skills provider ThriveDX called the shortage a "self-inflicted problem" as companies pursue applicants with a strict minimum level of expertise, compounded by a dearth of specialized and accelerated training programs.
*May Require Paid Registration
|
Seismologists Use Deep Learning to Forecast Earthquakes
UC Santa Cruz Newscenter Erin Malsbury August 31, 2023
Researchers at the University of California, Santa Cruz and Germany's Technical University of Munich developed the Recurrent Earthquake foreCAST (RECAST) model to use deep learning to predict earthquake aftershocks. The researchers found RECAST slightly outperformed the Epidemic Type Aftershock Sequence model on quake catalogs of roughly 10,000 seismic events and greater, especially as the amount of data expanded. Using RECAST also significantly improved the computational time and effort for larger catalogs. The deep learning model's greater flexibility and scalability could unlock new earthquake forecasting possibilities, potentially incorporating data from multiple regions simultaneously to make better predictions about poorly investigated areas.
|
Innovation Paves Way for Driverless Cars, Drone Fleets, Faster Broadband
University College London News (U.K.) August 28, 2023
New frequency referenced multiplexing technology developed by researchers in the U.K. and China could enable driverless cars and drone fleets. The researchers believe the technology could provide more than 20 times the capacity of the best available full fiber broadband networks and surpass the speed of typical U.K. home broadband 65 times, while ensuring continued connectivity and low latency. They deployed a proof-of-concept frequency referenced multiplexing system to supply as many as 64 users with connections providing speeds of up to 4.3 gigabits/second per user, for a cumulative speed of 240 gigabits/second. Zhixin Liu at the U.K.'s University College London said, "In the short term, the technology has the potential to provide a much better home broadband service at a low infrastructure cost."
|
Carmakers Fail Privacy Test, Give Owners Little Control of Collected Personal Data
Associated Press Frank Bajak September 6, 2023
A survey by the nonprofit Mozilla Foundation found most automakers admitted to possibly selling car owners' personal data. Albert Fox Cahn at Harvard's Carr Center for Human Rights Policy warned, "The electronics that drivers pay more and more money to install are collecting more and more data on them and their passengers." Mozilla assigned cars the worst privacy score among more than a dozen product categories the organization has reviewed since 2017. None of the 25 car brands whose privacy notices Mozilla vetted this year met the nonprofit's minimum privacy standards, which include encrypting all personal information, compared to 37% of reviewed mental health applications. Nineteen car brands said they can sell owners' personal information, with half also willing to share it with the government or law enforcement without a court order.
|
Air Leak Detection System Visualizes Building Drafts
Oak Ridge National Laboratory August 31, 2023
Scientists at the U.S. Department of Energy's Oak Ridge National Laboratory (ORNL) have developed a system that uses a camera to detect air leakage from buildings in real time. The researchers based the system on background-oriented schlieren photography, which ORNL's Philip Boudreaux said can visualize leaking air that is a different temperature than the surrounding air as indicated by background changes in a sequence of images. Boudreaux said a digital video camera can capture this mirage, which "looks just like wavy patterns you might see rising up from the pavement on a hot day or in the hot exhaust of a car tailpipe." He said the team created algorithms that measure velocity and flow rate for real-time leak visualization, and has visualized air leakage through brick, vinyl siding, and concrete masonry blocks under sunny and cloudy lighting.
|
|