Welcome to the March 10, 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."
|
|
|
Computer Scientists Paint a Picture of Six Decades of Movies
University of Waterloo (Canada)
March 9, 2023
Computer scientists at Canada's University of Waterloo analyzed the trailers of 29,000 North American movies released from 1960 to 2019 to identify the dominant colors used in individual films, genres, and decades. Using the k-means clustering technique, the researchers produced general eight-color palettes and more detailed 15-color palettes. The algorithm eliminated skin tones and asphalt to determine more accurately the most dominant colors used. According to the analysis, westerns used muted earth tones across the decades, and neon green was prevalent in science fiction films. Said Waterloo's Andreea Pocol, "While movies in the 1960s and 1970s tended to use more saturated primary colors, the team’s analysis demonstrates modern technology has actually allowed directors to use a wider variety of colors in creative ways."
|
Women Discontent with Digital Skills, Eager to Upskill
Gallup
Benedicte Clouet
March 8, 2023
An online survey of workers aged 18 and older in nine countries across the Asia-Pacific region by Amazon Web Services and Gallup found women often are less confident than men about their digital skills and their ability to keep pace with the digital skills required in their fields. Women surveyed in Singapore, South Korea, Malaysia, and New Zealand, whose jobs involved handling information and using a computer, were much less likely than men in similar jobs to be "completely confident" or "very confident" about their current level of digital skills. In most countries, the survey found, the percentage of women who anticipated increased salaries and opportunities for promotion as a result of digital training exceeded the percentage who actually received such benefits.
|
Researchers Unveil AI-Driven Method for Improving Additive Manufacturing
Argonne National Laboratory
Nikki Forrester
March 9, 2023
Researchers at the U.S. Department of Energy's Argonne National Laboratory and the University of Virginia (UVA) have developed a new technique to enhance additive manufacturing by detecting and predicting flaws in three-dimensionally (3D) printed materials. The researchers employed imaging and machine learning (ML) to anticipate the generation of pores in 3D-printed metals in real time. Argonne's Samuel Clark said researchers can image more than 1 million frames per second using high-intensity X-ray beams generated by the Advanced Photon Source (APS) facility. Correlating X-ray and thermal images expose unique thermal signatures at the material's surface that thermal cameras can detect, while an ML model predicts pore formation from thermal images. Said UVA's Tao Sun, “The APS offered the 100% accurate ground truth that allowed us to achieve perfect prediction of pore generation with our model.”
|
Mini Robot Enters Blood Vessels, Completes Surgery
IEEE Spectrum
Michelle Hampson
March 7, 2023
Researchers at South Korea's Hanyang University recently demonstrated that a miniature robot could travel autonomously to a superficial femoral artery in a pig, deliver contrast dye, and return safely to the extraction point. This could pave the way for the use of robots to treat occlusive vascular disease in humans and eliminate the need for X-ray imaging to guide surgical equipment. The researchers developed the I-RAMAN (robotically assisted magnetic navigation system for endovascular intervention) robot, which can navigate a patient's blood vessels autonomously using a three-dimensional map generated from two-dimensional X-ray images. The robot is injected into a blood vessel via catheter, then an external magnetic field untethers the robot from the catheter and guides it to the treatment area and back.
|
Researchers Take a Step Towards Turning Interactions That Normally Ruin Quantum Information into Protecting It
Aalto University (Finland)
March 8, 2023
Scientists at Finland's Aalto University and China's IAS Tsinghua University have developed a method for predicting the behavior of many-body quantum systems connected to their environment, in order to shield quantum information from decoherence. While this linkage typically disrupts the information, the researchers found having a quantum device host non-Hermitian topology can protect quantum excitations because of their environmental exposure. Said Aalto's Guangze Chen, "The method we developed allows us to solve correlated quantum problems that present dissipation and quantum many-body interactions simultaneously." Aalto's Jose Lado said the research provides "a methodology for understanding and predicting both quantum materials and devices in realistic conditions in quantum technologies."
|
Drone Maps Previously Uncharted Ocean Floors Off Alaska, California
FOX Weather
Hillary Andrews
March 7, 2023
The Saildrone Surveyor unmanned surface vehicle has completed the mapping of the previously uncharted ocean floors off the Alaska and California coasts. The sail-, solar-, and diesel-powered drone gathered data via sonar from the ocean’s surface to chart the ocean floor more than four miles down; other instruments gathered sound, ocean, and weather data. Manufacturer Saildrone said the Surveyor drone is capable of operating in 40-m.p.h. winds and 16-foot swells. The ocean-mapping project is part of a public-private partnership co-funded by the U.S. National Oceanic and Atmospheric Administration to define the country's Exclusive Economic Zone, and also to identify Essential Fish Habitat.
|
Computer Components Inspired by Brain Cells
Empa (Switzerland)
Anna Ettlin
March 7, 2023
New memristors developed by researchers at Switzerland's Empa and ETH Zurich and Italy's Politecnico di Milano (Polytechnic University of Milan) are more powerful and easier to manufacture than previous versions. The thin-film memristors are based on halide perovskite nanocrystals, a semiconductor material that conducts both ions and electrons. Said Empa and ETH's Rohit John, "This dual conductivity enables more complex calculations that closely resemble processes in the brain." The new memristors are easier to manufacture because perovskites crystallize at low temperatures and there is no need for complex preconditioning through the application of specific voltages, but they are difficult to integrate with existing computer chips because perovskites cannot withstand the temperatures needed to process silicon.
|
Scientists Make Biosensor for Crop Blight
The Jerusalem Post (Israel)
Judy Siegel-Itzkovich
March 6, 2023
A molecular sensing system used by researchers at Israel's Hebrew University of Jerusalem can enter plant cells to identify potential crop diseases. The researchers genetically engineered new potato strains to produce proteins that function as biological sensors. Cameras scan the plants for signals emitted by the protein sensors, which are collected as spatial information. The researchers said the images captured by the cameras helped track the plant's physiological condition throughout the progression of late blight. The biological sensor was found to be capable of detecting diseased areas of leaves even during the initial hidden stages, which led to the development of an algorithm to analyze the images and differentiate healthy from infected leaves.
|
Breakthrough in Big Data Processing Helps Trace Chemicals in Complex Mixtures
Czech Academy of Sciences (Czech Republic)
March 8, 2023
An international team led by researchers at the Czech Republic's Czech Academy of Sciences developed the open source software MZmine for analyzing large mass spectrometry datasets. The latest generation of the software, MZmine 3, can process thousands of chemical samples per hour; previous generations took days to analyze hundreds of samples. MZmine 3 also can connect different types of data, particularly time-resolved and imaging data, to more easily examine and interpret complex samples. Ansgar Korf of the University of Münster said the original MZmine software “established itself as a trusted tool for mass spectrometry researchers over the past decade.” Korf added that the software’s “modular framework has fostered community participation in the development of the MZmine code, leading to significant advancements featured in the newly released MZmine 3.”
|
These Tools Help Visually Impaired Scientists Read Data, Journals
Nature
Alla Katsnelson
March 6, 2023
Several tools have been developed to help blind scientists and those with low vision read data and journals. The Allen Institute for Artificial Intelligence created SciA11y, an online tool that extracts the content and structure of a PDF using machine learning and re-renders it in HTML so it can be navigated using screen readers. Researchers at the Massachusetts Institute of Technology and the U.K.'s University College London collaborated on Olli, a screen-reader interface that permits users to navigate different levels of description. Other relevant tools include the Georgia Institute of Technology's Highcharts Sonification Studio, which allows researchers to upload data and consider different ways to represent the data aurally.
|
Researcher Releases Code for Largest-Ever Spiking Neural Network for Language Generation
UC Santa Cruz Newscenter
Emily Cerf
March 7, 2023
The University of California, Santa Cruz's Jason Eshraghian and colleagues have open-sourced a new language generation model that addresses other models' high computational costs and reliance on maintenance from just a few companies. The SpikeGPT model incorporates the largest-ever spiking neural network (SNN), which consumes “22 times less energy” than a similar model using deep learning. Eshraghian said, "We're taking an informed approach to borrowing principles from the brain, copying this idea that neurons are usually quiet and not transmitting anything. Using spikes is a much more efficient way to represent information." Eshraghian added that enabling SpikeGPT to operate on sufficiently low power to achieve brain-level scalability could reduce people's dependence on monopolized entities to maintain such models.
|
Wiring Map of an Insect's Brain Hints at Incredible Complexity
NPR
Jon Hamilton
March 9, 2023
A multi-institutional team of scientists created the first wiring diagram, or connectome, of the brain of a fruit fly larva. The researchers sliced the larva's brain, the size of a grain of salt, into thousands of very thin sections, then used an electron microscope to image each slice, tracing neuronal links with powerful computers and specialized computational tools. Johns Hopkins University's Joshua Vogelstein said the larva brain, like a human's, contains regions corresponding to decision-making, learning, and navigation. The map will help scientists research how learning transforms the brain, gender-based differences in brain wiring, and development-driven changes to wiring.
|
High-Performance Computing Offers Guidance for Development of Systems That Emit Less Pollution
San Diego Supercomputer Center
Kimberly Mann Bruch; Cynthia Dillon
March 7, 2023
Georgia Institute of Technology (Georgia Tech) researchers identified new methods for developing cleaner, more efficient gas turbine combustion systems using supercomputer simulations. The researchers used the Expanse supercomputer at the San Diego Supercomputer Center and the Bridges-2 supercomputer at the Pittsburgh Supercomputing Center to study the impact of combustion on flow structure and flow physics under certain conditions. Georgia Tech's Vedanth Nair said, "The results of our experiments showed us that combustion has a significant impact on the flow structure under the conditions we considered in our simulations.” Nair added the experiments also “demonstrated a parameter “critical to consider while studying reacting jets in crossflow – the Reynolds number of the jet, which is a measure of the inertial versus viscous tendencies of the jet."
|
|
