Welcome to the June 19, 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."
|
|
Tech Majors Are Booming, but Rural Students Stuck in the Digital Divide
The Hill Lexi Lonas June 18, 2023
The digital divide leaves rural students fewer opportunities to enter expanding and well-paying technology industries, despite a boom in tech majors in college. The National Student Clearinghouse Research Center estimated a 23% surge in "Computer and Information Sciences and Support Services" majors from 423,315 to 518,844 between 2018 and 2022. However, Rural Technology Fund director Chris Sanders said many rural schools still lack computer classes, while those that exist likely focus on typing and Microsoft Office. Last week, the White House announced the Online for All initiative, which will collaborate with more than 300 national and community organizations "to close the digital divide."
|
Food Producers Band Together in Face of Cyber Threats
WSJ Pro Cybersecurity James Rundle June 15, 2023
Executives of U.S. food and agriculture companies say they are collaboratively formalizing information-sharing to fortify themselves against escalating cybersecurity threats. Last month, the Information Technology-Information Sharing and Analysis Center (IT-ISAC), which tracks threats across multiple industries, launched its own dedicated information-sharing platform. Technology is an integral component of food production, with farmers using distributed networks, remote sensors, edge-computing devices, and heavy equipment that are often insecure and vulnerable to hackers. IT-ISAC's Scott Algeier, executive director of the new Food and Ag-ISAC, said, "We're tracking threat actors. We have playbooks that we've developed and our members provide input on that, help track the adversaries, their tactics, techniques, and procedures, how they move around, how you can stop them."
*May Require Paid Registration
|
Researchers Collaborate on Dynamic 3D Structure Representation
Carnegie Mellon University School of Computer Science Aaron Aupperle June 14, 2023
The Dynamic Light Field Network (DyLiN) model developed by researchers from Carnegie Mellon University (CMU) and Japanese information and communications technology corporation Fujitsu can render two-dimensional images as a three-dimensional structure. DyLiN uses artificial intelligence to manage non-rigid deformations and topological alterations, and "has already shown significant improvements over existing methods in terms of speed and visual fidelity," explained CMU's Laszlo A. Jeni. The accompanying CoDyLiN model augments DyLiN with controllable attribute inputs; the researchers used knowledge distillation from pretrained dynamic radiance fields to train both models. DyLiN's performance on synthetic and real-world datasets containing non-rigid deformations was up to 71 times faster than state-of-the-art models without losing visual fidelity.
|
Conservationists in Peru Turn to Robots to Help Reforest the Amazon
EuroNews June 15, 2023
The Junglekeepers International nonprofit and Swiss-Swedish robotics company ABB are testing a robotic seed-planter to help reforest the Peruvian Amazon. ABB's solar-powered, dual-arm YuMi robot was engineered for automation in various industries. Solar satellite Wi-Fi keeps YuMi linked to a base in Sweden 12,000 kilometers (7,456 miles) away. Junglekeepers' Mohsin Kazmi said, "The robot is able to remove soil and plant seeds all in one box, 16 at a time. And then we take that box and put it into our nursery for it to germinate." Kazmi said the robot could plant up to 600 trees in a morning, allowing conservationists to devote more time to achieving broader reforestation goals.
|
92% of Programmers Use AI Tools: Survey
ZDNet Steven Vaughan-Nichols June 14, 2023
A recent survey by GitHub found that 92% of U.S.-based developers use artificial intelligence (AI) coding tools, with only 6% using them solely outside of work. Of the 500 U.S.-based developers polled, 70% said their code has benefited significantly from AI. The respondents said AI coding tools are useful in achieving performance standards with better code quality, faster outputs, and fewer production-level issues. However, AI code appears to be a means to an end for developers, as the survey found that they “want to upskill, design solutions, get feedback from end users, and be evaluated on their communication skills." Said GitHub's Inbal Shani, "Engineering leaders will need to ask whether measuring code volume is still the best way to measure productivity and output."
|
Reforms to Home Appraisal Bias Target Algorithms, Tech
Bloomberg Brentin Mock June 12, 2023
Vice President Kamala Harris announced reforms designed to narrow home-appraisal gaps between white neighborhoods and neighborhoods of color through technology-based solutions. The Biden administration has proposed a regulation requiring banks and other lenders to safeguard automated valuation models (AVMs) from algorithmic biases that might generate racial appraisal inequities. Financial institutions also would have to follow non-discrimination quality control standards. Melody Taylor at the Interagency Task Force on Property Appraisal and Valuation Equity said these standards "would require financial institutions that use AVMs to adopt and maintain policies to ensure greater confidence in their valuation estimates," while also requiring some "to conduct random sampling, testing, and reviews of their systems."
|
Bot Detection Software Not as Accurate as It Seems
MIT Sloan School of Management Dylan Walsh June 12, 2023
Massachusetts Institute of Technology (MIT) researchers found the accuracy of third-party bot-detection models may be lower than reported due to limitations in their training data. The researchers applied a commercially available machine learning model to a Twitter dataset from a repository hosted by Indiana University, which sorted bots from people with 99% accuracy. Deeper analysis showed models trained to perform well on one dataset did not necessarily outperform random guessing on a different dataset, suggesting general-purpose models trained on such data may be vulnerable to error in real-world scenarios. The researchers also learned relatively simple algorithms yielded accuracy similar to that of more complex models for many datasets.
|
Computer Vision Technique Enhances Microscopy Image Analysis for Cancer Diagnosis
University of Michigan Computer Science and Engineering June 9, 2023
Researchers at University of Michigan (U-M) and its Michigan Medicine health care complex have developed a computer vision learning technique that analyzes microscopic images of tumors using artificial intelligence (AI) and machine learning to make cancer diagnoses. The HiDisc tool analyzes multiple patches of a single tumor; it also relies less on data augmentation to transform and label images. In tests against state-of-the-art self-supervised learning methods for diagnosing cancer, the researchers observed an almost 4% improvement in image classification compared to the best-performing baseline method. U-M's Cheng Jiang said, "Our goal is to use AI to help clinicians make a diagnosis based on these tumor images virtually instantaneously during surgery, while the surgeon is waiting."
|
Benchmarking Robots with Dog-Inspired Barkour
IEEE Spectrum Sarah Wells June 7, 2023
Scientists from Google DeepMind evaluated the agility of quadripedal robots with a dog-agility-inspired course called Barkour. DeepMind's Atil Iscen said Barkour features assorted obstacles requiring "a combination of different types of behaviors such as precise walking, climbing, and jumping," while a timing-based metric rewards faster behavior. Barkour's obstacles include a pause table, weave poles, climbing an A-frame, and a jump to test “multiple axes of agility, including speed, acceleration, and balance," according to Iscen. The researchers used a specialist training strategy that focused on each type of skill required for the course and a generalized framework that analyzed models run using the specialist program.
|
When Materials Discovery Glitters
Pacific Northwest National Laboratory Rebekah Orton June 13, 2023
An open source quantum materials database developed by researchers at the U.S. Department of Energy's Pacific Northwest National Laboratory (PNNL) serves as a virtual laboratory for the discovery of new materials. PNNL's Tim Pope said the database is helping scientists explore transition metal dichalcogenides (TMDs), which possess identical crystal structures but exhibit different properties based on how they are combined and grown. The researchers used the database to expose differences in electrical and magnetic characteristics between different TMD combinations, and to gain new insights into quantum-level transition metal chemistry. Said PNNL's Peter Sushko, "Projects like this provide critical data to the machine learning community and could streamline materials development."
|
Mercedes Bringing ChatGPT into its Cars
CNN Business Peter Valdes-Dapena June 15, 2023
German automaker Mercedes-Benz has partnered with Microsoft to add ChatGPT generative artificial intelligence software to Mercedes-Benz cars in the U.S. Microsoft said ChatGPT would make the vehicles' voice-command capability smoother by supporting more natural-seeming dialogue. The system will be able to recall the context of discussions and engage in back-and-forth conversation with the driver or occupants. Microsoft said the chatbot will allow the system to respond to more diverse requests, including those not related to the car or driver, and interact with other functions, like buying movie tickets. U.S.-based Mercedes owners whose vehicles include the MBUX infotainment system have been able to beta-test ChatGPT since June 16.
|
|