Welcome to the June 5, 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."
|
|
|
World's Most Popular Online Computer Class Turns to AI for Help
Bloomberg
Saritha Rai
June 2, 2023
Harvard University's David J. Malan said the school's CS50 introductory computer science (CS) class will utilize artificial intelligence (AI) to mark assignments, teach programming, and personalize learning tips for students. Malan is credited with turning CS50 into the world's most popular online learning course; he said tailoring support to students' questions at scale has been challenging, as there are more students online than teachers. Malan's team is refining an AI system to grade students' work, and testing a virtual teaching assistant to assess and provide feedback on their coding by asking rhetorical questions and making suggestions to help them learn. Malan said the course's use of AI could underscore its educational advantages, especially for augmenting the quality of and access to online learning.
*May Require Paid Registration
|
Robot 'Chef' Learns to Recreate Recipes from Watching Videos
University of Cambridge (U.K.)
June 5, 2023
Scientists at the U.K.'s University of Cambridge programmed a robotic "chef" to make eight salad recipes, teaching it to identify recipes and to add to its repertoire by having it watch food videos. The researchers trained the robot chef using a publicly available neural network programmed to identify various objects, including salad ingredients. The robot’s computer vision analyzed video frames to recognize different objects and features, including the demonstrator's arms, hands, and face. The researchers converted the recipes and videos to vectors, which the robot rated for similarity by executing mathematical operations on them. It identified the correct recipe 93% of the time in the 16 viewed videos, despite only detecting 83% of the human chef's actions.
|
Black Men Were Likely Underdiagnosed with Lung Problems Due to Bias in Software
Associated Press
Mike Stobbe
June 1, 2023
A study published last week suggests built-in software bias may be responsible for the underdiagnosis of lung problems among Black men. The researchers analyzed the medical histories of more than 2,700 Black men and 5,700 white men evaluated by University of Pennsylvania Health System doctors between 2010 and 2020. They focused on spirometry and lung volume tests to evaluate how many were determined to have breathing difficulties using a race-based algorithm, and found nearly 400 more Black male patients might have been diagnosed with breathing problems using a new algorithm. Massachusetts General Hospital's Darshali Vyas said race-adjusting algorithms elevate the diagnostic threshold for Black patients, which may make them less likely to initiate certain medications or to receive referrals for medical procedures or lung transplants.
|
Turning a Circle into a Square Is Possible with This Kirigami-Inspired Formula
MIT News
Jennifer Chu
June 1, 2023
Researchers at the Massachusetts Institute of Technology (MIT) and Harvard University invented a computational strategy for creating transformable two-dimensional (2D) structures inspired by the Japanese paper-cutting art of kirigami. Said MIT’s Gary Choi, “People have talked of the square and circle as one of the impossible problems in mathematics: You cannot turn one into the other. But with kirigami, we can actually turn a square shape into a circle shape.” The strategy could be used to solve some design problems for engineers, such as having a robot change its shape to perform a specific task, or to fit into a specifically shaped space.
|
Pneumatic Computer Uses Pressure Instead of Electricity
New Scientist
Karmela Padavic-Callaghan
June 2, 2023
Elliot Hui and colleagues at the University of California, Irvine built a computer from glass and silicone that encodes data using pneumatic pressure instead of electricity. The researchers sandwiched a 0.25-millimeter (0.009-inch)-thick silicone sheet between two panes of glass etched with tiny channels to conduct liquids for chemical reactions, then perforated the silicone layer to link the channels. Pressure variations impel liquids through the channels, similar to how voltage changes direct electricity through wires in electronic computer chips. The researchers coded programs by using different silicone sheets as "punch cards," and inputted data by placing their fingers over designated points to change pressure.
|
'Designer' Titanium Alloys Made Using 3D Printing
The University of Sydney (Australia)
June 2, 2023
Scientists at Australia's University of Sydney (USYD), RMIT University, and smart manufacturer Hexagon Manufacturing Intelligence have combined alloy and three-dimensional (3D) printing process designs into a new class of strong titanium alloys. The new alloys integrate alpha- and beta-titanium phase crystals, using oxygen and iron as stabilizing agents. The researchers printed the alloys from metal powder via the Laser-Directed Energy Deposition 3D printing process. USYD's Simon Ringer said, “This research delivers a new titanium alloy system capable of a wide and tunable range of mechanical properties, high manufacturability, enormous potential for emissions reduction, and insights for materials design in kindred systems.”
|
Artificial Synapses for Neuromorphic Computing
Los Alamos National Laboratory
June 1, 2023
Los Alamos National Laboratory (LANL) researchers have developed an interface-type memristive device that could be used to create artificial synapses for next-generation neuromorphic computing. Memristors can replicate the structure and function of synapses, the human brain's "in-memory processing" system; they also save time and energy by co-locating information storage and processing. Comprised of a simple structure of gold and other semiconducting materials, the interface-type memristor could be scaled down to nanometer size and needs much less processing power than transistor-based neuromorphic chips. LANL's Aiping Chen said the advantages of the new device include low-energy consumption, high parallelism, and excellent error tolerance, which ” make it very good for advanced computing tasks like learning, recognition, and decision-making."
|
'Primate AI Algorithm' Predicts Genetic Health Risks
Financial Times
Clive Cookson
June 1, 2023
An international team of academic researchers analyzed a primate DNA database with artificial intelligence (AI) to more accurately predict genetic mutations that cause disease in humans. The researchers based their PrimateAI-3D algorithm on the genetic information of roughly 800 primates across 233 species of apes, monkeys, and lemurs. They collaborated with U.S. DNA sequencing equipment maker Illumina to identify 4.3 million common genetic variants in the primates' genomes, then predicted their health effects by training PrimateAI-3D on data about the mutations and the three-dimensional protein structures they generate. The researchers then used PrimateAI-DC on the DNA data of 454,000 humans participating in the U.K. Biobank project, which yielded "greatly improved genetic risk prediction," they said.
*May Require Paid Registration
|
Wireless Monitoring Patch System Detects Sleep Apnea at Home
Georgia Institute of Technology
May 24, 2023
A wearable device developed by researchers at the Georgia Institute of Technology can measure both obstructive sleep apnea and the wearer's quality of sleep. The device could allow people with suspected sleep disorders to undergo less-expensive sleep testing at home, instead of needing to be monitored overnight at a medical facility. The silicone patch fits over the forehead, with a silicone attachment that conforms to the chin. It features three embedded electronic sensors that record brain, eye, and muscle activity and transmit data via Bluetooth to an app on a smartphone or tablet. Artificial intelligence and machine learning technology are used to analyze the data and assign a sleep score, which indicates whether the wearer has sleep apnea or the quality of their sleep.
|
Multirobot Mapping Sets 'Gold Standard'
IEEE Spectrum
Evan Ackerman
June 2, 2023
Roboticists at the Massachusetts Institute of Technology have established what was judged to be a new "gold standard" for distributed multirobot simultaneous localization and mapping (SLAM) with their Kimera-Multi distributed system, which operates locally on a group of robots simultaneously. The researchers were awarded the IEEE Transactions on Robotics Best Paper Award for 2022 for the documentation of their development. IEEE Transactions on Robotics editor in chief Kyle Lynch said his publication’s editorial board, and the reviewers, were “deeply impressed by the theoretical elegance and practical relevance of this paper and the open-source code that accompanies it. Kimera-Multi is now the gold standard for distributed multirobot SLAM.”
|
Newborn Inspires Sensor Design Simulating Human Touch
Penn State News
June 1, 2023
The medical condition of a newborn baby inspired researchers from Pennsylvania State University (Penn State) and China's Hebei University of Technology to develop an extremely pressure-sensitive sensor that simulates human touch. Penn State's Huanyu Cheng needed to weigh his daughter, who lost 10% of her body weight shortly after birth, every two days to check for additional weight loss. When a commercial scale failed to detect weight-related pressure changes, Cheng and colleagues explored the development of a pressure sensor comprised of laser-fabricated gradient micro-pyramidal structures and an ultrathin ionic layer to induce capacitive responses. Said Cheng, “An analogy I like to use is it’s like detecting a fly on top of an elephant. It can measure the slightest change in pressure, just like our skin does with touch.”
|
Better Search for the Cause of Hereditary Diseases
Technical University of Munich (Germany)
June 1, 2023
An interdisciplinary team of researchers at Germany's Technical University of Munich (TUM) and federal research institution Helmholtz Munich developed a model to better predict DNA variations that result in RNA defects to more precisely identify the genetic causes of rare hereditary diseases and cancer. The resulting algorithm was found to be six times more accurate than previous models in predicting the impact of genetic mutations on RNA formation. Said TUM and Helmholtz Munich's Julien Gagneur, "We achieved such a large advance in precision by looking at the splicing process in a tissue-specific way and by using direct splicing measurements from easily accessible tissues such as blood or skin cells in order to predict splicing errors in inaccessible tissues like the heart or the brain."
|
|
