Welcome to the May 3, 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."
|
|
'Godfather of AI' Leaves Google, Warns of Danger
The New York Times Cade Metz May 1, 2023
Artificial intelligence (AI) pioneer Geoffrey Hinton has resigned from Google, warning about the risks of generative AI-based products. With the technology already being used to produce misinformation, Hinton and others fear it could soon threaten jobs, and even humanity. The neural network technology that became Hinton's life's work earned him and two collaborators ACM's 2018 A.M. Turing Award and led to the creation of chatbots like OpenAI's ChatGPT. Hinton said the threat of AI systems escalates as their capabilities improve, and Google's decision to deploy generative AI systems in response to Microsoft enhancing its Bing search engine with a chatbot is concerning. His immediate fear is of the Internet being swamped with fake content, while in the longer term chatbots could replace professionals who perform rote tasks.
|
Sensor Enables High-Fidelity Input from Everyday Objects
University of Michigan Computer Science and Engineering April 27, 2023
The SAWSense system designed by University of Michigan (U-M) scientists can transform virtually any surface into a high-fidelity computer input device. SAWSense updates equipment from bone-conduction microphones to enable the detection of surface-acoustic waves, using machine learning to recognize inputs like taps, scratches, and swipes with 97% accuracy. The system's Voice PickUp unit (VPU) sensors are contained in a hermetically sealed chamber that blocks even loud ambient noise, while a mass-spring system conducts the surface-acoustic waves within the chamber without interacting with sounds in the surrounding environment. U-M's Yasha Iravantchi said, "VPUs do a good job of sensing activities and events happening in a well-defined area. This allows the functionality that comes with a smart object without the privacy concerns of a standard microphone that senses the whole room, for example."
|
Fish Dished Up by 3D Printers
Reuters Rami Amichay; Amir Cohen; Ari Rabinovitch May 3, 2023
Israeli food technology company Steakholder Foods worked with Singapore's Umami Meats to develop three-dimensionally (3D)-printed ready-to-cook fish fillets using laboratory-cultured animal cells. Umami excises and grows cells from grouper into muscle and fat, then Steakholder Foods adds the cells to a bio-ink for 3D printing. Production involves sliding a glass dish back and forth in the printer as the fillet's mass gradually accrues. Umami hopes to debut the fillets commercially in Singapore next year, then bring them to countries like the U.S. and Japan, depending on local regulations. Said Steakholder Foods' Arik Kaufman, "As time goes by, the complexity and level of these products will be higher, and the prices linked to producing them will decrease."
|
Optical Neural Networks Hold Promise for Image Processing
Cornell Chronicle Diane Tessaglia-Hymes April 27, 2023
An optical neural network (ONN) developed at Cornell University could help pave the way for faster, smaller, and more energy-efficient image sensors. The ONN can filter relevant data from a scene before a camera detects the visual image. Said Cornell's Tianyu Wang, "By discarding irrelevant or redundant information, an ONN can quickly sort out important information, yielding a compressed representation of the original data, which may have a higher signal-to-noise ratio per camera pixel." The researchers observed compression ratios of up to 800-to-1 with ONN pre-processors. They also demonstrated the original image could be reconstructed using data generated by ONN encoders trained only for image classification. Wang noted, "The reconstructed images retained important features, suggesting that the compressed data contained more information than just the classification."
|
Smart Glasses Dim Bright Objects
New Scientist Matthew Sparkes May 2, 2023
Camera- and liquid-crystal display (LCD) -equipped smart glasses developed by scientists at Japan's Nara Institute of Science and Technology (NAIST) can "balance" a scene by dimming bright objects without affecting dim ones. The glasses feed camera input through a computer running an algorithm that dims the brightest areas. The LCD lenses can tune the amount of light passing through each pixel of the LCDs according to computer output. The device can adjust to changing lighting conditions in 20 microseconds, enabling seamless-seeming transitions. NAIST's Xiaodan Hu said, "The general public can use them while driving or cycling for added comfort and safety."
|
Drones Can Survey Oyster Reefs More Efficiently
University of Florida News Eric Hamilton May 1, 2023
Researchers at the University of Florida (UF) and the U.K.'s University of Oxford have used drones to map oyster reefs to determine those at risk of collapse. The researchers demonstrated how drones can accurately assess reef conditions more efficiently than manual surveys through the use of light detection and ranging (LiDAR). The drones fly over and scan the reefs, then software converts the LiDAR scans into three-dimensional digital elevation maps. The team's latest analysis indicated a strong correlation between reef volume and manual oyster counts. Said UF's Michael Espriella, "With enough surveys of reefs around the state, managers could dig deeper into specific situations and hopefully prevent larger collapses."
|
Apple, Google Partner to Combat Creepy Tracking Tactics
Associated Press May 2, 2023
Apple and Google have jointly proposed to the Internet Engineering Steering Group new standards for combatting clandestine surveillance through Bluetooth object-tracking devices like Apple's AirTag. Apple and AirTag hope to have a plan to foil stealth tracking ready by year's end, with the solution to be circulated via iPhone and Android phone updates. Erica Olsen at the National Network to End Domestic Violence's Safety Net Project said she believes the initiative will help protect abuse survivors and others targeted by stealth technology. Olsen said the new standards “will minimize opportunities for abuse of this technology and decrease the burden on survivors in detecting unwanted trackers."
|
IW3C2 Selects “Landmark” Paper for 2023 Seoul Test of Time Award
International World Wide Web Conference Committee May 2, 2023
The International World Wide Web Conference Committee (IW3C2) selected the authors of a 2010 paper to receive the 2023 Seoul Test of Time Award at the 32nd ACM Web Conference in Austin, TX. Seoul Test of Time Award Committee Chair Ryen White said the paper, “A Contextual-Bandit Approach to Personalized News Article Recommendation,” offered "a new approach for personalized recommendation using contextual bandit algorithms" that "addressed fundamental challenges in real-world recommendation systems via computationally efficient algorithms grounded in learning theory." Paper co-author and Amazon senior principal scientist Lihong Li said basing the contextual bandit model on the fact that recommendation was basically a decision-making challenge framed as reinforcement learning enabled the researchers to consider online exploration and offline evaluation on an Internet scale.
|
Scientists Take Step Towards Using Quantum Computers to Advance Materials Science
Ames Laboratory April 28, 2023
Researchers at the U.S. Department of Energy's Ames National Laboratory (Ames Lab) created an adaptive algorithm for simulating materials to expand quantum computing's role in materials science. The work focuses on finding alternatives for rare earth materials, whose complex electronic structures challenge simulation, according to Ames Lab's Yongxin Yao. The researchers based their approach on models that describe magnetic impurities in materials and capture electronic properties by accounting for how those impurities interact with the rest of the material. The process also applies quantum embedding methods systematically to simulate the bulk materials, reducing computational resources while preserving accuracy.
|
Highly Dexterous Robot Hand Can Operate in the Dark
Columbia Engineering News Holly Evarts April 28, 2023
Columbia Engineering scientists melded advanced tactile sense with motor learning algorithms into a dexterous robot hand that does not rely on vision. The hand was able to arbitrarily rotate an unevenly shaped grasped object in a stable, secure grip without visual feedback. The hand features five fingers outfitted with touch-sensing technology and 15 independently actuated joints, and uses deep reinforcement learning to learn new tasks via practical application. Physics simulators and highly parallel processors enabled the robot to learn roughly one year of practice in just hours, which translated into the desired dexterity with the actual hand. Columbia Engineering's Matei Ciocarlie said, "Once we also add visual feedback into the mix along with touch, we hope to be able to achieve even more dexterity, and one day start approaching the replication of the human hand."
|
Android-Based Application for Photoacoustic Tomography Image Reconstruction
SPIE Newsroom April 27, 2023
Researchers from Singapore and the U.S. have developed an application for photoacoustic tomography (PAT) image reconstruction based on the Python coding language. The app facilitates image reconstruction on the Kivy cross-platform Python 3.9.5 framework using a single-element ultrasound transducer-based delay-and-sum beamformer algorithm. The researchers utilized simulated and experimental PAT datasets to confirm the app's performance on different mobile phones. Iowa State University's Manojit Pramanik explained, "The developed application can successfully reconstruct the PAT data into high-quality PAT images with signal-to-noise ratio values above 30 decibels." The code for the Android-based application has been made freely available on GitHub.
|
Open Source Platform Simulates Wildlife for Soft Robotics Designers
MIT News Alex Shipps May 2, 2023
A multi-institutional team of scientists developed the open source SoftZoo platform to permit engineers to co-design soft robots, using three-dimensional models of animals to simulate the performance of tasks. The framework optimizes algorithms to model the robot's configuration and motion-facilitating control to enhance automatic production of robots' outlines. A multiphysics engine enables SoftZoo to simulate movement that responds to the physical properties of various biomes. Users can design and manipulate soft robots with more refined algorithms because the engine's differentiability reduces the number of costly simulations needed to solve computational control and design challenges. The Massachusetts Institute of Technology's Chuang Gan explained, "SoftZoo provides open source simulation for soft robot designers, helping them build real-world robots much more easily and flexibly while accelerating the machines' locomotion capabilities in diverse environments."
|
|