Welcome to the October 13, 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."
|
|
|
Israel Sees Cyber Incursions Across Digital Systems
WSJ Pro Cybersecurity
James Rundle; Kim S. Nash
October 12, 2023
Cyber aggression has escalated in Israel amid the conflict with Hamas. Check Point Software Technologies reported Israel's smart billboards were hacked to display pro-Hamas messages and images. Cyber firm Sepio reported an increase in attempts to access Israel's industrial systems. Distributed denial-of-service attacks also have impacted numerous municipal and consumer websites. Israel was already the Middle Eastern country most targeted in nation-state cyberattacks, according to Microsoft research published last week. In response, members of the country’s tech community formed the all-volunteer Israel Tech Guard to search for hostages and missing people using clues from online posts. Cyber volunteers are also working to protect key services, such as missile-alert apps, as well as databases and websites for aid coordination, applications for first responders, and location services for displaced civilians.
*May Require Paid Registration
|
Teaching Computers to Recognize Food in All Forms
University of Maryland Department of Computer Science
Maria Herd
October 9, 2023
University of Maryland (UMD) computer scientists have developed a dataset to train machine learning systems to recognize 20 different fruits and vegetables regardless of whether they are peeled, sliced, or chopped. The researchers developed Chop & Learn by filming seven different food preparation styles from four different angles to ensure comprehensive coverage. Said UMD's Abhinav Shrivastava, "Being able to recognize objects as they are undergoing different transformations is crucial for building long-term video understanding systems, as well as dealing with the long-tail problem in object recognition. We believe our dataset is a good start to making real progress on the basic crux of this problem in compositional image generation and action recognition."
|
Scrolls That Survived Vesuvius Divulge Their First Word
The New York Times
Nicholas Wade
October 12, 2023
Researchers have recovered a handful of letters and the word “porphyras,” ancient Greek for “purple,” from a papyrus scroll that was carbonized by the eruption of Mount Vesuvius in A.D. 79. The scroll, which comes from a villa thought to have been owned by the father-in-law of Julius Caesar, would fall apart if unrolled. The approach used to read the scroll, developed by Brent Seales, a computer scientist at the University of Kentucky, uses computed tomography (CT) and artificial intelligence software to help distinguish ink from papyrus. Seales released his software programs into the public domain and has offered prizes for certain milestones to accelerate efforts to retrieve text from other scrolls found at the villa. Some 1,500 people, many of them machine learning experts, are now involved. Private donors have sponsored a $700,000 prize if someone can retrieve four separate passages of at least 140 characters from the scrolls this year.
*May Require Free Registration
|
Illuminating Errors Creates New Paradigm for Quantum Computing
Princeton University
October 11, 2023
A team led by researchers at Princeton University developed a technique to pinpoint the location of errors in quantum computing to make them easier to correct by applying the erasure-error model to matter-based qubits. Using an array of 10 qubits, the researchers manipulated individual qubits, then pairs of qubits to determine the probability errors would occur. They used a different set of energy levels within an atom to store the qubits, allowing them to observe flashes of light emitted by qubits with errors in real time. The researchers were able to detect half of single-qubit errors and a third of two-qubit errors during the experiment. Said Princeton University's Jeff Thompson, "What's nice about erasure conversion is that it can be used in many different qubits and computer architectures, so it can be deployed flexibly in combination with other developments."
|
Disney Packs Big Emotion into Little Robot
IEEE Spectrum
Evan Ackerman
October 6, 2023
A team of Disney Research scientists in Zurich, Switzerland, unveiled a bipedal robotic character combining a child-size frame with stubby legs, wiggling antennae, and an expressive walk at the 2023 IEEE/Robotics Society of Japan International Conference on Intelligent Robots and Systems last week. The mostly three-dimensionally-printed robot has a four-degrees-of-freedom head and five-degrees-of-freedom legs with hip joints so it can perambulate while balancing dynamically. Disney's Michael Hopkins said the design team included an animator to help the robot walk in a manner that conveys emotions. Disney Research’s reinforcement learning-based pipeline integrates robust robotic movement with an animator's vision through simulation. Disney's Moritz Bächer said the team was able to develop the robot character in months rather than years because the pipeline can condense years of behavioral training into a few hours.
|
Optical Chip Self-Configures to Perform Various Functions
Optica
October 11, 2023
A team led by researchers at China's Huazhong University of Science and Technology developed a self-configuring optical chip that could be used in applications requiring optical neural networks. The chip features a network of Mach–Zehnder interferometers arranged in a quadrilateral pattern and can configure itself to perform optical routing, low-loss light energy splitting, and positive real matrix computations. Adjusting electrode voltages produces different light propagation paths in the quadrilateral network, allowing the chip to be reconfigured. Said Huazhong University's Jianji Dong, "Our new chip can be treated as a black box, meaning users don't need to understand its internal structure to change its function. They only need to set a training objective, and, with computer control, the chip will self-configure to achieve the desired functionality based on the input and output."
|
C# Challenges Java in Language Popularity
InfoWorld
Paul Krill
October 11, 2023
Microsoft’s C# language is catching up to Java in popularity, with the difference between the two standing at only 1.2% in the Tiobe index of programming language popularity for October. C# is anticipated to surpass Java soon under current trends. Java had long held the top spot in the index until being overtaken by C, Python, and C++ in recent years. It now stands at No. 4, with an 8.92% rating, while fifth-ranked C# has a 7.71% rating. Rounding out the top 10 most popular languages on the Tiobe ranking are JavaScript (2.91%), Visual Basic (2.13%), PHP (1.9%), SQL (1.78%), and Assembly (1.64%).
|
Cyber Algorithm Shuts Down Malicious Robotic Attack
University of South Australia
October 12, 2023
An algorithm developed by researchers at Australia’s Charles Sturt University and the University of South Australia (UniSA) was able to intercept and prevent a man-in-the-middle (MitM) eavesdropping cyberattack on an unmanned military robot within seconds. The researchers used deep learning neural networks to train the robot operating system (ROS) in a replica of a U.S. Army GVT-BOT ground vehicle to learn the signature of a MitM cyberattack. In real-time tests, the algorithm achieved a 99% success rate in preventing such attacks. UniSA's Anthony Finn said the algorithm outperforms existing cyberattack recognition techniques. Added Charles Sturt University's Fendy Santoso, "Owing to the benefits of deep learning, our intrusion detection framework is robust and highly accurate. The system can handle large datasets suitable to safeguard large-scale and real-time data-driven systems such as ROS."
|
Google Launches Flood Hub to Predict When U.S. Rivers Will Flood
Fast Company
Adele Peters
October 10, 2023
Earlier this year, Google deployed its Flood Hub early-warning flooding tool in the U.S. and Canada. The tool already had been launched in more than 80 countries since its 2018 Indian rollout. Flood Hub builds a digital model of land in a specific region from thousands of satellite images, then models river flooding in combination with weather forecasts. The tool can forecast floods two to seven days in advance, giving local residents sufficient time to evacuate. Flood Hub, whose warnings are currently limited to "riverine" floods, can notify governments and residents of flood zones though mobile-phone alert.
|
Researchers Turn to AI to Avoid Drone Collisions
Johns Hopkins University Hub
Megan Mastrola
October 9, 2023
A team of researchers from Johns Hopkins University (JHU) and the Charles Stark Draper Laboratory modeled a drone traffic-orchestrating system using artificial intelligence (AI) to substitute certain human-in-the-loop tasks with autonomous decision-making. JHU's Lanier Watkins said, "Our simulated system leverages autonomy algorithms to enhance the safety and scalability of UAS [uncrewed aircraft systems] operations below 400 feet altitude." The researchers assessed the effect of strategic deconfliction algorithms in a simulated airspace, boosting safety and nearly eliminating collisions. The simulator includes "noisy sensors" to improve adaptability to unanticipated conditions and a "fuzzy interference system" that factors in multiple variables to calculate each drone's risk level. The researchers intend to incorporate weather and other dynamic obstacles into simulations to model real-world conditions more comprehensively.
|
Glasses Transcribe Speech in Real Time
Scientific American
Timmy Broderick
October 10, 2023
Several companies have started marketing "live-captioning glasses" to help hearing-impaired users communicate with hearing persons by overlaying real-time transcribed speech on the lenses. Stanford University's Tom Pritsky said the glasses his company TranscribeGlass makes complement hearing aids by filling in blanks in conversations through captions and subtitles. Most live-captioning glasses are eyewear equipped with a microphone, a speech-processing computer, a battery, and a text display. Dan Scarfe with speech-processing application provider XRAI Glass credits these innovations to advances in speech-recognition software, adding, "I don't think we're more than six months away from a killer piece of hardware that you can absolutely use for this."
|
New Approach to Rare Disease Diagnosis
KAUST Discovery (Saudi Arabia)
October 10, 2023
Researchers at King Abdullah University of Science and Technology (KAUST) in Saudi Arabia have introduced the Symptom-based Tool for Automatic Ranking of Variants (STARVar) method for diagnosing genetic variants underlying rare diseases. The artificial intelligence-powered tool offers standardized or natural language-formatted interpretations of symptom data. STARVar outperformed other variant prioritization tools limited to rigidly represented symptoms when assessed on different genomic datasets, consistently ranking the right disease-linked variant at or near the top of the list of potential candidates. KAUST’s Senay Kafkas said STARvar “stands a unique and efficient tool that has the advantage of prioritizing genomic variants by using flexibly expressed patient symptoms in free-form text.”
|
Robots Help Understanding of How Insects Evolved Two Flight Strategies
UC San Diego Today
Ioana Patringenaru
October 4, 2023
University of California, San Diego (UCSD) roboticists worked with Georgia Institute of Technology biophysicists for six years to design robots as part of their probe of the evolution of two different modes of insect flight. The researchers investigated how moths' muscles exhibit signs of prior asynchrony and how insects can support both synchronous and asynchronous muscular characteristics and still fly. They built flapping-wing robots to mimic asynchronous and synchronous muscle combinations and test possible evolutionary transitions. One motorized robot the researchers developed demonstrated how an insect could transition between the two modes gradually and smoothly. UCSD's Nick Gravish said such research "could help usher in a new era of responsive and adaptive flapping-wing systems."
*May Require Paid Registration
|
|
