Welcome to the July 21, 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."
|
|
|
Lego Robot Builds DNA Structures for Tiny Machines More Quickly
New Scientist
Karmela Padavic-Callaghan
July 19, 2023
Researchers at Arizona State University built a robot out of Lego components to expedite the assembly of biological machines from DNA. The device is designed as a smaller, faster version of gradient mixers, which blend the liquid content of cylindrical tubes into a single liquid whose density decreases from bottom to top. Scientists can purify structures composed of DNA molecules with this liquid for use as machines. These machines could execute tasks within cells and gain insights into the functions of naturally occurring molecular machines such as proteins. In tests, the robot produced the required density gradient in just 60 seconds.
|
MIT Makes Probability-Based Computing a Bit Brighter
IEEE Spectrum
Edd Gent; Margo Anderson
July 19, 2023
Massachusetts Institute of Technology (MIT) researchers have produced the first probabilistic bit (p-bit) using photonics. The method's core component is an optical parametric oscillator (OPO), which is basically two mirrors reflecting light back and forth between them. The researchers can influence the likelihood with which an oscillation's phase assumes a particular state by injecting the OPO with extremely weak laser pulses. MIT's Charles Roques-Carmes explained, "We can keep the random aspect that just comes from using quantum physics, but in a way that we can control the probability distribution that is generated by those quantum variables." The researchers said they were able to generate 10,000 p-bits per second, which appear to support the necessary behavior for building a probabilistic computer.
|
Individual Whale, Dolphin ID Using Facial Recognition Tech
University of Hawaii News
July 19, 2023
An international team of scientists designed a tool that can identify individual whales and dolphins across two dozen species in the wild via facial recognition. The researchers developed the multi-species photo-identification algorithm to recognize details like scarring, pigmentation, and size. The University of Hawaii at Manoa's Philip Patton said, "When we go out and do these surveys like taking pictures of them out in the field, using an algorithm like this we can really speed up the information gathering process. Once we get back to the lab, we can run our photos through the algorithm and it will tell us who is there, and then we immediately have some information to judge things like population, space use, etc. which are important for conserving Hawaiian whales and dolphins."
|
Researchers Look at How 'E-prints' Add to—or Detract from—Online Discussions
Binghamton University News
Chris Kocher
July 20, 2023
Researchers at Binghamton University and the U.K.'s University College London investigated the use of e-prints—non-peer-reviewed scientific research—on online forums Reddit and 4chan to examine their impact on online discussions. Binghamton's Jeremy Blackburn views the availability of free e-prints to people who otherwise could not access professional research as positive, but acknowledges they could spread flawed information. Reddit posts between mid-2005 and mid-March 2021, and posts from 4chan's Politically Incorrect board from mid-2016 to mid-March 2021, covered subjects ranging from computer science and physics to genetics and neuroscience, with COVID-19 emphasized when the pandemic was worst. Posters who used e-prints to make their case sometimes misread the research or overlooked scientists' techniques, missing methodical flaws or pseudoscience. Blackburn said these findings reflect the need for scientists to consider how laypeople will interpret their research.
|
ChatGPT's Accuracy Has Gotten Worse
Popular Science
Andrew Paul
July 19, 2023
Stanford University and University of Southern California, Berkeley (UC Berkeley) researchers demonstrated an apparent decline in the reliability of OpenAI's ChatGPT large language model (LLM) over time without any solid explanation. The researchers assessed the chatbot's tendency to offer answers with varying degrees of accuracy and quality, as well as how appropriately it follows instructions. In one example, the researchers observed that GPT-4's nearly 98% accuracy in identifying prime numbers fell to less than 3% between March and June 2023, while GPT-3.5's accuracy increased; both GPT-3.5 and GPT-4's code-generation abilities worsened in that same interval. UC Berkeley's Matei Zaharia suggested the decline may reflect a limit reached by reinforcement learning from human feedback, or perhaps bugs in the system.
|
Search Engines, Social Media Can Forecast Disease Outbreaks
Waterloo News (Canada)
July 10, 2023
Researchers at Canada's University of Waterloo found search engine queries and social media data can be used to predict disease outbreaks as an early-warning real-time surveillance system. The researchers used Google Trends and Twitter data in Canada posted from January to March 2020 to extract COVID-19 symptom keywords, then cross-checked this information against data from the COVID-19 Canada Open Data Working Group. They observed a strong correlation between search terms related to COVID-19 symptoms and daily COVID-19 cases within a one- to 13-day interval. The forecasting machine learning model they devised operated better using Google Trends than it did with Twitter data.
|
U.S. Program to Identify Safe Smart Devices
Voice of America News
Bryan Lynn
July 19, 2023
The Biden administration has announced a program to identify smart devices that are more resilient against cyberattacks, to help the public decide which commercial devices are safest to use. The White House said the Federal Communications Commission (FCC)-administered program will assign the most protected devices a "U.S. Cyber Trust Mark." Analysts will use quality requirements created by the National Institute of Standards and Technology to rate the devices' safety levels. The requirements call for unique and strong passwords, data safeguards, software improvements, and built-in tools for detecting cyberattack attempts. The FCC's Jessica Rosenworcel said the program would encompass a broad spectrum of Internet of Things devices that administration officials warn have the potential to elevate cybersecurity threats.
|
A Pink Rover Tackles the Red Planet, Barriers for Women in Science
Nature
Amanda Heidt
July 19, 2023
A team of engineers from Australia's Monash University placed second in the University Rover Challenge in Hanksville, UT, with a pink-colored robotic rover that navigated desert terrain mimicking the surface of Mars. The Monash Nova Rover team, which is more than half female, aimed to address the competition's under-representation of women. Monash's Chloe Chang said the team designed the Waratah rover to spur discussions about women in science, technology, engineering, and math, and robotics is particular. Said Chang, "Robotics is fundamentally creative, and we want to be accessing a greater pool of creative thinking and ideas to ultimately find the best solutions to engineering problems. Seeing more women coming into leadership and having that visibility definitely helps."
|
Algorithm Learns Chemical Language, Accelerates Polymer Research
Georgia Tech News Center
July 18, 2023
The polyBERT machine learning model developed by the Georgia Institute of Technology's Christopher Kuenneth and Rampi Ramprasad could revolutionize polymer research. The researchers trained polyBERT on a dataset of 80 million polymer chemical structures so it could become fluent in the “chemical language” of polymers. The algorithm extracts the most meaningful information from chemical structures using the Transformer architecture employed in natural language models. PolyBERT outperforms traditional chemical fingerprinting solutions by more than two orders of magnitude, enabling the rapid screening of vast polymer spaces at an unparalleled scale.
|
Quantum-in-the-Loop Interface Connects Power Grids, Quantum Computers
Interesting Engineering
Tejasri Gururaj
July 19, 2023
Researchers from the National Renewable Energy Laboratory (NREL) connected quantum computers to power research equipment via an open source software interface created in collaboration with quantum computer developer Atom Computing and Canadian power industry protection and testing equipment provider RTDS Technologies. The interface facilitates quantum-in-the-loop experiments in which quantum computers solve power system optimization problems efficiently. Researchers can render optimization problems as quantum variables and communicate with power system simulations in real time using the quantum-in-the-loop framework. The interface connects NREL's real-time grid simulators with Atom Computing's quantum emulator and prototype system. The team demonstrated the interface near Boulder, CO, using NREL's Advanced Research on Integrated Energy Systems as a power system environment for the quantum-in-the-loop experiments.
|
Detecting Threats Beyond the Limits of Human, Sensor Sight
Sandia National Laboratories
July 20, 2023
Tian Ma and Robert Anderson at the U.S. Department of Energy's Sandia National Laboratories created a software system to detect and track pixel-sized objects moving in streaming video and images from remote sensors. Ma said he and Anderson developed the Multi-frame Moving Object Detection System (MMODS) to detect small objects in low visibility. Images from sensors stream into a computer station, which MMODS filters frame by frame in real time. An algorithm spots and matches movement in frames with target signals that can be correlated and combined across a series of frame sequences, enhancing the signal-to-noise ratio or image quality. In a test, adding MMODS to a baseline detector system boosted the odds of detecting a moving object from 30% to 90% without elevating the rate of false positives.
|
Farm Robots Poised for Growth as Labor Costs Rise
Financial Times
Sarah Murray
July 12, 2023
Labor shortages are making robots like Harvard University's RoboBees increasingly economically viable. The RoboBees can take off vertically, hover, and steer, which could minimize water and agrichemical use by enabling plant-by-plant cultivation. Agricultural technology developer FarmWise has engineered a robot that can distinguish weeds from crops using computer vision and artificial intelligence; it then can kill the weeds with a precision application of herbicide. Field-deployed drones and robots can use smart sensors and computer vision to collect and send information to farmers in real time; machine learning can extract new insights into cultivation from the resulting large datasets. FarmWise's Tjarko Leifer thinks precision farming facilitated by robotics will fuel its adoption.
*May Require Paid Registration
|
Amazon Cashless 'Pay by Palm' Technology Requires Only a Hand Wave
CBS News
Simrin Singh
July 20, 2023
Retail giant Amazon has announced a new contactless transaction service that allows shoppers to pay with their palms. Users can enable transactions by hovering their palms over an Amazon One device, which can facilitate payment, identification, loyalty program membership, and entry. Amazon said palm payment is impossible to replicate because the system creates unique "palm signatures" for each customer by examining the palm and the underlying vein arrangement. Each palm signature, the company added, corresponds to a numerical vector representation, and is securely warehoused in the Amazon Web Services cloud. The technology is already available at 200 Amazon locations in 20 U.S. states, and the company intends to deploy it at more than 500 Whole Foods and Amazon Fresh outlets by year's end.
|
|
