Welcome to the October 7, 2022, edition of ACM TechNews, providing timely information for IT professionals three times a week.
Please note: In observance of the U.S. Columbus Day holiday, TechNews will not be published on Monday, October 10. Publication will resume Wednesday, October 12.
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."
|
|
Leading Makers Pledge Not to Weaponize Their Robots
NPR Joe Hernandez October 6, 2022
Six major robot manufacturers have signed a letter promising not to weaponize their products. Boston Dynamics, Agility Robotics, ANYbotics, Clearpath Robotics, Open Robotics, and Unitree pledged against weaponizing their "advanced-mobility general-purpose robots" or their underlying software, while also vowing to ensure their customers do not weaponize them either. The companies also said they do not oppose "existing technologies" used by governments to "defend themselves and uphold their laws." Boston Dynamics says police and fire departments are using the company's canine-like robot Spot to assess hazardous situations, but the firm notes Spot is not designed for surveillance or as a substitute for police officers.
|
DeepMind Invents Faster Algorithms to Solve Tough Math Puzzles
Nature Matthew Hutson October 5, 2022
Researchers at artificial intelligence (AI) laboratory DeepMind have created an algorithm that can solve tough mathematical calculations with improved computing efficiency. The researchers designed the AlphaTensor algorithm to execute matrix multiplication, which entails multiplying numbers arranged in grids that might represent data. AlphaTensor incorporates reinforcement learning as well as tree search, a game-playing approach in which the AI probes the outcomes of branching possibilities while planning its next action. The researchers tested AlphaTensor on input matrices up to 5 x 5. In some cases it rediscovered shortcuts previously formulated by mathematicians, while in others it found new shortcuts of its own.
|
Researchers Nanoprint Electrodes for Customized Treatments of Disease
Carnegie Mellon University October 5, 2022
Carnegie Mellon University (CMU) researchers have produced a microelectrode array (MEA) for brain-computer interface (BCI) platforms to enable customized therapies for neurological disorders. The researchers three-dimensionally (3D)-printed the nanoscale MEA via Aerosol Jet 3D printing. CMU's Rahul Panat said the microfabrication technique allows users to tailor MEAs for specific needs, enables the arrays to function three-dimensionally in the brain, and increases their density and robustness. MEA-based BCIs link neurons with external electronics to monitor or stimulate neural activity, and the addition of the third dimension boosts the arrays' sampling ability. Custom-made MEAs also provide more accurate and higher-fidelity readings. The researchers say their CMU Array is the densest BCI to date.
|
Nutrient-Measuring Device Could Make Microwaves That Count Calories
New Scientist Jeremy Hsu October 5, 2022
Researchers at the Ohio State University found that electromagnetic radiation from microwave ovens can be used to estimate the nutrients and calories of the food being cooked. They demonstrated that these electromagnetic radiation signals, which provide information on the vibration of different food molecules and the rates at which they heat up, can be detected using off-the-shelf radio receivers. Tests of a prototype system consisting of a small radio antenna receiver and software for analysis found that it could determine the percentages of water, protein, carbohydrates, and fat in over 150 foods with 81% accuracy. Although the system must factor in details about the microwave containers, such as materials and surface area, and the mass of the food, the researchers found it to be more accurate than calculating nutritional facts based on images. Ohio State's Avishek Banerjee noted that "our work is dealing with the intrinsic properties of the food."
*May Require Paid Registration
|
Algorithm May Help Prevent Power Blackouts from Ransomware Attacks
Purdue University October 4, 2022
Ransomware-rigged power blackouts may be prevented by an algorithm developed by Purdue University researchers to map out areas of the power grid where utilities should prioritize security. The algorithm would incentivize each security decision-maker to apportion security investments in order to limit the cumulative damage caused by a ransomware attack. The researchers evaluated the algorithm in the context of different types of critical infrastructure in addition to the power industry. They tested the tool in models of previously reported hacks of a smart grid, an industrial control system, an e-commerce platform, and a Web-based telecommunications network. The algorithm enabled the most optimal allocation of security investments for mitigating attacks, the researchers said.
|
How Did the Moon Form? A Supercomputer May Have Just Found the Answer
Live Science Ben Turner October 5, 2022
A hypothesis based on supercomputer simulations by researchers at the U.K.'s Durham University suggests that the moon may have formed within hours after a chunk of Earth was broken off following a collision with an ancient protoplanet known as Theia. This is in contrast to the long-standing belief that the moon formed over thousands of years after the collision. The researchers used Durham's COSMA supercomputer to run the SWIFT (SPH With Inter-dependent Fine-grained Tasking) program, which simulated hundreds of collisions between Earth and Theia at different angles, spins, and speeds. The researchers modeled up to 100 million particles, higher than previous simulation resolutions. Durham's Jacob Kegerreis said, "Even more samples from the surface of the moon could be extremely helpful for making new and more confident discoveries about the moon's composition and evolution, which we can then trace back to model simulations like ours."
|
NIST's Superconducting Hardware Could Scale Up Brain-Inspired Computing
U.S. National Institute of Standards and Technology October 6, 2022
Researchers at the U.S. National Institute of Standards and Technology have developed a circuit that mimics a biological synapse while transmitting and receiving signals with single photons. Computations occur where a single-photon detector meets a superconducting Josephson junction, and researchers can set different synaptic weights by altering the bias to the Josephson junctions. This allows for a longer-term memory that can be used to render networks programmable, enabling the same network to solve many different problems. The demonstration marks a key advance toward realizing superconducting optoelectronic networks.
|
Smartphone App Could Help Advance Early Detection of Parkinson's Disease, Severe COVID-19
News-Medical Life Sciences October 5, 2022
An app developed by a team led by researchers at Australia's RMIT University could improve the management of Parkinson's disease and severe COVID-19 through early detection. The screening app uses voice recordings to determine within 10 seconds whether the user should be referred to a health care provider for diagnosis and treatment. The app can detect changes in a person's voice due to rigidity, tremor, and slowness, which occurs with Parkinson's disease, and changes in a person's voice due to lung infections from COVID-19. The tool, trained using voice recordings from people with Parkinson's and pulmonary disease and a group of healthy individuals, was found to be faster and more accurate than similar methods. Said RMIT's Dinesh Kumar, "This research will allow a non-contact, easy-to-use and low-cost test that can be performed routinely anywhere in the world, where the clinicians can monitor their patients remotely."
|
Malaysia's Palm Planters Eye Robots, Drones to Combat Labor Crunch
Reuters Mei Mei Chu October 7, 2022
Palm oil plantations in Malaysia hope to address a labor shortage with technology including wearable exoskeletons and drones. Plantation owner Hamidon Salleh said he and colleagues at Malaysia's University of Technology have partnered with palm oil producer Sime Darby Plantation to test the exoskeleton, which enables one harvester, rather than two, to harvest 10 metric tons of palm fruit. Other producers are accelerating deployment of drones to spray crops with fertilizer and pesticide, map estate holdings, and monitor trees. Drone services provider Meraque has expanded its fleet from three drones to 62 since 2018 in order to meet pandemic-fueled demand. The company's Razalee Ismail said one artificial intelligence-equipped drone can do the work of six people.
|
Device Can Recognize 'Millions of Colors' in Machine Vision Advance
Northeastern University Tanner Stening October 5, 2022
Northeastern University researchers have constructed a device that uses artificial intelligence techniques to perceive "millions of colors" in what they call a massive step for machine vision technology. The researchers said the A-Eye device can analyze and process color with greater accuracy than existing machines. The team produced two-dimensional material whose quantum properties can process diverse colors when incorporated within an optical window. The material forms "transmissive windows" that substitute for filters. The researchers said A-Eye also can "accurately recognize and reproduce 'seen' colors with zero deviation from their original spectra" through machine learning algorithms. "A-Eye can continuously improve color estimation by adding any corrected guesses to its training database," they said.
|
Advanced Control of a Six-Qubit Quantum Processor in Silicon
QuTech (Netherlands) September 29, 2022
Researchers at QuTech, a collaboration between the Netherlands' Delft University of Technology and TNO, the Netherlands Organization for Applied Scientific Research, have developed a chip with six silicon-based spin qubits that operates with a low-error rate. The research is seen paving the way for the development of scalable silicon-based quantum computers. The researchers created an array of six "quantum dots" made of individual electrons spaced 90 nanometers apart. They controlled and measured the spin of the individual electrons and facilitated their interaction using microwave radiation, magnetic fields, and electric potentials. Said QuTech's Lieven Vandersypen, "In this research we push the envelope of the number of qubits in silicon, and achieve high initialization fidelities, high readout fidelities, high single-qubit gate fidelities, and high two-qubit state fidelities. What really stands out though is that we demonstrate all these characteristics together in one single experiment on a record number of qubits."
|
Hello, Fellow Car. We've Got a Problem. Let's Talk
The New York Times Eric A. Taub October 5, 2022
Car manufacturers are developing cellular vehicle-to-everything (C-V2X) technology so vehicles can communicate with each other, as well as with people and road infrastructure systems. V2X technology is expected to enable cars to notify drivers of obstructions; alert highway workers to oncoming vehicles; and warn cyclists, pedestrians, and others of collisions and other potential hazards. V2X technology could appear in vehicles as soon as the 2025 model year, and the industry will initially use Long-Term Evolution cellular technology to affect communication. "I can't imagine future vehicles without V2X technology," said Ram Iyer at vehicle technology supplier Harman International. "Because the human attention span is decreasing, you'll look for these technologies to save your life."
*May Require Paid Registration
|
Rethinking the Computer Chip in the Age of AI
Penn Engineering Today Devorah Fischler September 29, 2022
A team of researchers from the University of Pennsylvania (Penn), Sandia National Laboratories, and Brookhaven National Laboratory has unveiled a computing architecture suited for artificial intelligence (AI). The researchers developed a transistor-free compute-in-memory (CIM) architecture where processing and storage happen in the same place, removing transfer time and minimizing energy consumption. The architecture, which builds on earlier work on a ferroelectric switching scandium-alloyed aluminum nitride semiconductor, could potentially perform up to 100 times faster than a conventional computing architecture. The design also performs on-chip storage, parallel search, and matrix multiplication acceleration. Penn's Xiwen Liu said the work "proves that we can rely on memory technology to develop chips that integrate multiple AI data applications in a way that truly challenges conventional computing technologies."
|
|