Welcome to the October 11, 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."

States Are Calling for More Computer Science Classes States Are Calling for More Computer Science Classes. Now They Need Teachers
Education Week
Sarah D. Sparks
October 5, 2023

Code.org reported that every state had a law or policy in place promoting K-12 computer science education by 2022. During that year, it said 53% of high schools offered basic computer science courses. However, concerns are on the rise about who will teach these courses, given school staffing issues nationwide and differing state certification requirements for computer science education. In 2020, 42 states allowed teachers to add computer science to an existing license, 19 required completion of an independent certification program, and 24 had alternative certification pathways. A study led by University of Illinois Urbana-Champaign's Paul Bruno found most computer science teachers in North Carolina, for instance, are business or math teachers, with only 1% having specific computer science certification. Said Bruno, "We need to think more strategically about the teacher supply."

Full Article
*May Require Paid Registration
OmniMotion Allows for Better Video Motion Estimation
Cornell Chronicle
Tom Fleischman
October 10, 2023

The OmniMotion optimization tool developed by scientists at Cornell University and Google Research can estimate motion in videos. Cornell's Noah Snavely said OmniMotion tracks dense and long-range motion across time, capturing the ordering relationships between objects through what the researchers call "a quasi-three-dimensional representation." The tool produces a full-motion video representation from a sample of frames and motion estimates, which can be queried after optimization with any pixel in any frame to generate a smooth, accurate motion pathway across the entire video. Uses for OmniMotion could include incorporating computer-generated imagery into video editing and improving motion coherence of generated videos via movement-estimation algorithms, according to Snavely.

Full Article

Geoffrey Hinton and Scott Pelley Geoffrey Hinton on the Promise, Risks of Advanced AI
CBS News
Scott Pelley
October 8, 2023

U.K. computer scientist and 2019 ACM A.M. Turing Award recipient Geoffrey Hinton said advanced artificial intelligence (AI), for all its promise, could conceivably take over. In a "60 Minutes" interview, Hinton said AI systems are intelligent, capable of comprehension, and can make experiential decisions in the same sense that humans do; achieving self-awareness is only a matter of time, effectively making AI more intelligent than humans. Hinton and collaborators Yann LeCun and Yoshua Bengio created a neural network to learn by trial and error, strengthening connections that lead to correct outcomes. Hinton suggested modern AI systems can learn better than the human mind despite having fewer connections, even though their exact inner workings are unknown. Hinton urges experiments to improve our understanding of how the technology works, as well as government regulation, and a worldwide ban on the use of military robots.

Full Article

An artist's rendition of a biochemical model inspired by an interpretable neural network Neural Network for Genomics Explains How It Achieves Accurate Predictions
NYU News
October 6, 2023

A neural network developed by computer scientists at New York University (NYU) has the ability to explain how its predictions are made. Said NYU's Oded Regev, "Many neural networks are black boxes—these algorithms cannot explain how they work, raising concerns about their trustworthiness and stifling progress into understanding the underlying biological processes of genome encoding. By harnessing a new approach that improves both the quantity and the quality of the data for machine-learning training, we designed an interpretable neural network that can accurately predict complex outcomes and explain how it arrives at its predictions." The researchers used facts about RNA splicing to develop a neural network that enables the RNA splicing process to be traced and quantified. Regev noted, "Our model revealed that a small, hairpin-like structure in RNA can decrease splicing." This was confirmed through experiments.

Full Article

An app shows how ancient Greek sites looked thousands of years ago App Shows How Greek Sites Looked Thousands of Years Ago
Associated Press
Derek Gatopoulos; Theodora Tongas
October 8, 2023

Greece's Culture Ministry is supporting a free augmented reality application that renders ancient Greek sites and sculptures as they appeared 2,500 years ago, based on archaeological evidence. Visitors to the Parthenon in Athens and marble sculptures in London's British Museum can summon virtual overlays on their smartphone screens using Greek telecom provider Cosmote's "Chronos" app, which matches the real-world perspective as they walk around. Chronos' designers said they hope to advance features such as the artificial intelligence-powered Clio virtual guide. Said Cosmote's Panayiotis Gabrielides, "As technologies and networks advance, with better bandwidth and lower latencies, mobile devices will be able to download even higher-quality content."

Full Article
Improved Strategy for Social Media During Wildfires
University of Notre Dame Mendoza College of Business
Shannon Roddel
October 10, 2023

Researchers in the U.S. and Canada collaborated with the Canadian Red Cross to propose an alternative approach to social media communications for disaster relief organizations during wildfires. The researchers compiled Twitter data on Alberta, Canada's 2016 Fort McMurray wildfire, analyzing 934 tweets from the headquarters account that elicited 33,861 retweets and 34,722 likes, as well as 629 tweets from the Alberta account that garnered 4,802 retweets and 2,862 likes. The University of Notre Dame's Alfonso Pedraza-Martinez said, "We find that user engagement increases when the national headquarters lead the production of content and the local accounts follow either by tweeting to a matching or mismatching audience, depending on timing in the operation."

Full Article

Using drones to survey endangered species in Western Australia’s Kimberley region Indigenous Land Management Meets the Latest Drone Technology
Australian Financial Review
October 4, 2023

Indigenous Australian wildlife rangers are surveying endangered species like the wiliji (black-footed rock-wallaby) in Western Australia's Kimberley region using thermal drones. The rangers track wiliji movements with the drones and through methods like scat searches, live-trapping, and camera surveys via a partnership between Australia's Walalakoo Aboriginal Corporation, Charles Darwin University, and World Wildlife Fund-Australia. The project's goal is to find healthy wiliji populations and create a permanent monitoring program, as well as augmenting existing knowledge of the animals' responses to fire, feral predator control, forage plant protection, and cattle fences.

Full Article

Mum who took part in pilot Software Detects Hidden Emotions in Parents
University of Bristol News (U.K.)
October 5, 2023

Software developed by researchers at the U.K.'s University of Bristol and Manchester Metropolitan University can identify complex hidden emotions by mapping facial features and evaluating the intensities of multiple facial expressions. The researchers used data from participants recorded by headcams worn by their infants. The participants' facial expressions in the videos were analyzed by automated facial coding software and human coders; the researchers assessed how often the software detected faces in the videos and how often the software and humans were in agreement. Machine learning then was used to predict human judgments of parent facial expressions based on the decisions made by the computer. University of Bristol's Romana Burgess said, "Deploying automated facial analysis in the parents' home environment could change how we detect early signs of mood or mental health disorders, such as postnatal depression."

Full Article
Technology Helping with Biodiversity Survey
BBC News
October 10, 2023

U.K.-based volunteer group Friends of Sherford Country Park is undertaking a technology-enhanced citizen science wildlife biodiversity survey in the county of Devon under the auspices of the Pollenize conservation organization. Participants use their cellphones to record local wildlife and plants to be identified by a phone application, while Floradex software created by the U.K.'s University of Plymouth (UoP) identifies biodiversity improvement opportunities. So far, the Bioblitz survey has captured 777 observations of 186 distinct plant and animal species. UoP's Lauren Ansell said, "Pollinators are really hard to track—it can take years of study just to know what's in one area. By getting citizen science involved, getting people out there doing the dirty and hard work for us, it gives us a really good spectrum of what's going on in local areas."

Full Article

Study co-first authors Russell Kunes and Dr. Thomas Walle, and senior author Dr. Dana Pe'er l Open Source Method Improves Single-Cell Data Decoding
Memorial Sloan Kettering Cancer Center
Ian Demsky
October 5, 2023

The open source Spectra computational method developed by Memorial Sloan Kettering Cancer Center researchers can enhance analysis of single-cell transcriptomic data. The researchers said the algorithm can tease out clinically meaningful patient traits within data from large cohorts. Spectra uses libraries of expert-produced gene programs derived from previous data to guide analysis, while also adjusting to given data to detect new and revised programs. The algorithm also factors in genetic information that defines different cell types, augmenting its ability to discover programs underpinning cellular functions. The researchers utilized Spectra to analyze two breast cancer immunotherapy datasets and a lung cancer atlas that cumulatively amassed more than 1.5 million cells from 375 individuals in 21 studies.

Full Article
Robotic Hand Has the Dexterity to Handle Tricky Objects with Care
New Scientist
Chris Stokel-Walker
October 3, 2023

University of California, Berkeley (UC Berkeley) researchers have developed a robotic hand that can deftly manipulate complicated objects by rotating them in three axes. The RotateIt system was trained in a simulation in which it received details about the shapes and sizes of objects; a neural network then was trained to rotate the objects as much as possible within the simulation. The robotic hand's precision improved in real-world tests in which it was given visual and touch inputs. UC Berkeley's Haozhi Qi acknowledged the system needs to be improved to handle thinner objects like pencils and screwdrivers. However, Jonathan Aitken at the U.K.'s University of Sheffield said, "This algorithm provides an interesting and easily accessible method for changing the grasp on an object once it has been picked up."

Full Article
*May Require Paid Registration

A 3D schematic of the new device, known a four-phase electrooptic modulator, is shown Technology Could Reduce Lag, Improve Reliability of Online Gaming, Meetings
UCF Today
Robert Wells
October 4, 2023

A new class of optical modulators developed by researchers at the University of Central Florida (UCF) and the University of California, Los Angeles aims to increase the speed and efficiency of data transfer through optical fiber communication, reducing lag in online meetings and games. The four-phase electro-optic modulators leverage phase diversity (varied timing of signals) and differential operations (comparison of light signals) to overcome signal distortion and unwanted signal interference. The modulators simultaneously control data transmission and compare the amount and timing of data moving through the system. Said UCF's Ehsan Ordouie, "This breakthrough represents a noteworthy advancement in the practical implementation of photonic systems and opens up new possibilities for faster and more efficient data communication and acquisition."

Full Article
Optimizing Continuous-Variable Functions with Quantum Annealing
Tokyo Institute of Technology News (Japan)
October 4, 2023

Researchers at Japan's Tokyo Institute of Technology (Tokyo Tech) and Tohoku University have demonstrated the ability of quantum annealing (QA) algorithms to optimize continuous-variable functions better than classical algorithms. Tokyo Tech's Hidetoshi Nishimori said, "We systematically investigated whether QA has an advantage over classical algorithms by optimizing the Rastrigin function, a one-dimensional continuous function used as a standard for benchmarking optimization algorithms." The researchers assessed QA running on the D-Wave 2000Q quantum computer to determine it could outperform classical continuous-variable function and classical discrete-variable optimization algorithms. They observed that QA could overtake many of the algorithms by suppressing thermal noise.

Full Article
2023 SIGGRAPH Asia Conference
ACM Author's Rights

Association for Computing Machinery

1601 Broadway, 10th Floor
New York, NY 10019-7434

ACM Media Sales

If you are interested in advertising in ACM TechNews or other ACM publications, please contact ACM Media Sales or (212) 626-0686, or visit ACM Media for more information.

To submit feedback about ACM TechNews, contact: [email protected]