Welcome to the November 7, 2022, 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."
|
|
|
Google Wants Robots to Write Their Own Python Code
ZDNet
Liam Tung
November 3, 2022
Google Research has introduced a new process for using large language models (LLMs) to permit robots to write their own code in Python, based on instructions from humans. This approach builds on Google's PaLM-SayCan model for robots to comprehend open-ended prompts and respond reasonably and safely in physical environments. Google researchers described the approach as Code as Policies, and said code-writing LLMs can be reconfigured to compose robot policy code in response to natural language commands. They said the process is able to handle values like velocities, based on vague natural-language descriptions like "faster" or "to the left," and can incorporate instructions from other languages than English.
|
Web Inventor Tim Berners-Lee Wants Us to 'Ignore' Web3
CNBC News
Ryan Browne
November 4, 2022
ACM A.M. Turing Award recipient Tim Berners-Lee, credited with inventing the World Wide Web, considers Web3 nonviable for building the next iteration of the Internet. At the Web Summit in Lisbon, Portugal, Berners-Lee called Web3 a vague term to describe a theoretical Internet that is more decentralized than the current Web, incorporating technologies like blockchain, cryptocurrencies, and nonfungible tokens. Berners-Lee described blockchain protocols as "too slow, too expensive, and too public." He said people frequently confuse Web3 with his Web 3.0 framework for reconfiguring the Internet. His new Inrupt startup intends to allow users to control their own data via a global single sign-on feature for universal logins, login IDs that let users exchange data, and a "common universal application programming interface."
|
Smart Mask Tracks Respiratory Sounds for Disease Identification
City University of Hong Kong
November 2, 2022
A smart face mask developed by researchers at the City University of Hong Kong (CityU) can detect and recognize respiratory sounds using an integrated, high-sensitivity soundwave sensor and machine learning algorithms. The ultra-thin, lightweight sensor, comprised of carbon nanotube and polydimethylsiloxane materials, can be used with both rigid masks and deformable non-woven fabric masks. The researchers found the acoustic wave sensor can measure both static and dynamic pressure and sense air movement, including directional flow and vibration. Said CityU's Wang Jianping, "Advanced artificial intelligence technology enables the integrated mask to recognize different coughing and breathing patterns automatically, indicating its potential use to diagnose respiratory-related diseases in the future."
|
Technique Simultaneously Locates Multiple Defects on Microchip Circuits
NIST News
November 3, 2022
National Institute of Standards and Technology (NIST) researchers developed a method to locate individual electric flaws in multiple microcircuits on the same chip simultaneously by modifying an existing defect identification technique. The standard technique involves applying alternating current (AC) voltage to the ultrasharp tip of an atomic force microscope while it scans individual wires under the surface of a silicon chip, with an abrupt change in the tip's vibration indicating a wire break or defect. However, voltages from neighboring wires interfere with the scanning. The researchers overcame this by using an external generator to apply specific AC voltages to individual neighboring wires, ensuring the voltage to the neighboring wires was at its lowest value when the voltage to the wire being scanned was at its highest. NIST's Joseph Kopanski said, "The method does not require a new instrument and could be easily implemented by the semiconductor industry."
|
XPRIZE Winner Is Closest You Can Get to Teleportation
Forbes
Craig S. Smith
November 6, 2022
Japan's All Nippon Airways (ANA) named a team of German scientists and engineers recipients of the ANA Avatar XPRIZE for demonstrating wireless "teleportation" that connects a user’s vision, hearing, and tactile perception to that of a humanoid robot. The contest challenged teams to develop human-operated robotic avatars that can conduct tasks and remotely reproduce a operator's senses, actions, and presence in real time. The German team earned the $5-million XPRIZE for tele-operating the wheeled NimbRo robot through an obstacle course. The operator used haptics to manipulate NimbRo's hands so it could pick up a power drill and unscrew a bolt, then reach out of view and feel various objects to determine the object with the roughest texture.
|
Detecting Nuclear Threats with Artificial Reasoning
Pacific Northwest National Laboratory
Sarah Wong
November 3, 2022
Scientists at the U.S. Department of Energy's Pacific Northwest National Laboratory (PNNL) are employing machine learning (ML), data analytics, and artificial reasoning to better detect and analyze nuclear threats. PNNL's Benjamin Wilson and partners at Sandia National Laboratories built a virtual duplicate of a nuclear reprocessing facility, then educated an ML model to detect and process data patterns reflecting the diversion of nuclear materials. Wilson and other PNNL researchers also are working on the development of Artificial Judgement Assistance to analyze texts for signs of nuclear proliferation, and to answer direct queries from analysts. Another PNNL project with multi-institutional collaborators aims to facilitate ML-based forensic analysis of nuclear materials to determine their origin.
|
Billions Spent in Metaverse ‘Land’ Grab
BBC News
Joe Tidy
November 3, 2022
Research by DappRadar indicates that over the past year, people and companies have spent $1.93 billion in cryptocurrency to purchase virtual “real estate” in the metaverse. In Decentraland, parcels of “land” can sell for millions of dollars, and are being bought by companies like Samsung, UPS, and Sotheby's to build virtual shops. Adidas, Atari, Ubisoft, Binance, Warner Music, and Gucci have purchased virtual property in Sandbox, while Gucci also has created a town in Roblox. Said Amber Jae Slooten of The Fabricant, a digital design house, "There will be for sure a mass market in this because if you think about the younger generation, they already play games. For them there's no distinction between virtual and real. But it still needs to be built."
|
Global Analysis Shows Where Fishing Vessels Turn Off ID Devices
University of California, Santa Cruz
Tim Stephens
November 2, 2022
Researchers at the University of California, Santa Cruz (UC Santa Cruz), Global Fishing Watch, and NOAA Fisheries have produced the first global dataset on the disabling of Automatic Identification System devices by commercial fisheries. The researchers also developed a machine learning method to distinguish intentional disabling from legitimate disabling. The more than 55,000 suspected intentional disabling events identified between 2017 and 2019 concealed almost 5 million hours of fishing vessel activity, with over 40% of those hours occurring in four hotspots, three of which are known for illegal fishing. High levels of disabling events also were seen in areas with high transshipment activity. Legitimate reasons for disabling included hiding good fishing locations from competitors and protection from piracy.
|
Unpiloted Military Helicopter Flies 134 km in Simulated Mission
New Scientist
Jeremy Hsu
November 2, 2022
As part of the U.S. Defense Advanced Research Projects Agency's ALIAS program, the U.S. Army flew an autonomous Black Hawk Helicopter at 185 kph (115 mph) as low as 60 meters (197 feet) above the ground for 134 kilometers (83 miles) to deliver 230 kilograms (507 lbs.) of simulated and real blood. The helicopter flew between mountains using onboard sensors to avoid a collision. The program is testing Sikorsky Innovations' MATRIX technology, a drop-in kit that can transform helicopters into autonomous aircraft. Sikorsky's Igor Cherepinsky said, "It's a truly autonomous system. Once you explain what the mission is, you don't even need a data link–the aircraft makes all the decisions."
*May Require Paid Registration
|
Algorithm for 2D-to-3D Engineering Integrates Art, Nature, Science
Penn Engineering Today
Devorah Fischler
October 31, 2022
Researchers at the University of Pennsylvania School of Engineering and Applied Science (Penn Engineering) and the U.S. Army Combat Capabilities Development Command’s Army Research Laboratory have developed what they are calling a universal algorithm that permits two-dimensional materials to retain their lightness and durability when converted into three-dimensional (3D) structures. The algorithm allows hard materials to keep their mechanical strength after cutting by mimicking the structure of nacre, mollusks' natural shell coating. The algorithm can generate a computational map of cuts that are optimized for stacking, ensuring they never overlap with one another to compensate for necessary defects; fortifying tabs further bolster mechanical strength.
|
Synthetic Data Can Offer ML Performance Improvements
MIT News
Adam Zewe
November 3, 2022
A team of researchers at the Massachusetts Institute of Technology (MIT), the MIT-IBM Watson AI Laboratory, and Boston University found synthetic data can improve machine learning (ML) model performance. The researchers amassed the Synthetic Action Pre-training and Transfer dataset of 150,000 video clips to train three ML models on a wide range of human actions. They found the three models outperformed models trained with real video clips on four of six datasets, yielding the highest accuracy for datasets featuring clips with "low scene-object bias." MIT-IBM Watson AI Lab's Rogerio Feris said, "The ultimate goal of our research is to replace real data pretraining with synthetic data pretraining." Feris said that while there is a cost in creating an action in synthetic data, “Once that is done, then you can generate an unlimited number of images or videos by changing the pose, the lighting, etc."
|
|
