Welcome to the November 20, 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."
|
|
|
Using Next Generation Exascale Supercomputers to Understand the Climate Crisis
Association for Computing Machinery
November 16, 2023
The first ACM Gordon Bell Prize for Climate Modelling was awarded to a nineteen-member team of researchers for its algorithmic Simple Cloud Resolving E3SM Atmosphere Model (SCREAM). SCREAM features portable C++ innovations, including using the Kokkos programming model, a C++ library for on-node parallelism; algorithmic innovations, including formulations which help the core of the model conserve energy with exact timestepping; subgrid physical parameterizations, a class of models used in climate simulations; and use of the Software for Caching Output and Reads for Parallel I/O (SCORPIO) library and the Adaptable Input Output System I/O (ADIOS) library. The team noted that its successful cloud simulation on the Frontier exascale supercomputer is a transformative accomplishment for climate science and will allow trustworthy predictions of future climate impacts which have traditionally been localized.
|
Was Argentina the First AI Election?
The New York Times
Jack Nicas; Lucía Cholakian Herrera
November 16, 2023
Sergio Massa and Javier Milei widely used artificial intelligence (AI) to create images and videos to promote themselves and attack each other prior to Sunday's presidential election in Argentina, won by Milei. AI made candidates say things they did not, put them in famous movies, and created campaign posters. Much of the content was clearly fake, but a few creations strayed into the territory of disinformation. Researchers have long worried about the impact of AI on elections, but those fears were largely speculative because the technology to produce deepfakes was too expensive and unsophisticated. “Now we’ve seen this absolute explosion of incredibly accessible and increasingly powerful democratized tool sets, and that calculation has radically changed,” said Henry Ajder, an expert who has advised governments on AI-generated content.
*May Require Paid Registration
|
Introducing EUGENe: An Easy-to-Use Deep Learning Genomics Software
UC San Diego Today
Miles Martin
November 16, 2023
Researchers at the University of California San Diego (UCSD) created EUGENe (elucidating the utility of genomic elements with neural nets), a deep-learning platform that can adapt to a wide variety of different genomics projects. With EUGENe, explains UCSD's Adam Klie, "You give an algorithm a sequence of DNA and ask it to make predictions about anything you’d expect that DNA could predict, such as whether a particular DNA sequence is functional or whether it regulates a gene in a certain biological context. This lets you explore properties of the DNA sequence and ask what would happen if I modified this piece here or moved this piece there." The researchers tested EUGENe by having it reproduce the results of three existing genomics studies that utilized different types of sequencing data.
|
Kotlin Keeps Climbing the Tiobe Popularity Index
InfoWorld
Paul Krill
November 13, 2023
Kotlin, a Java rival computer language, continues to move up Tiobe’s index of language popularity after first cracking the top 20 in September with a rating of .9%. Kotlin climbed to No. 15 with a rating of 1.15% in November, after ranking 18th the previous month. Tiobe cited Kotlin's interoperability with Java and Android accommodations as reasons for the language’s rise. The Tiobe index top 10 for November was topped by Python, with a rating of 14.16%, followed by C (11.77%), C++ (10.36%), Java (8.35%), C# (7.65%), JavaScript (3.21%), PHP (2.3%), Visual Basic (2.1%), SQL (1.88%), and Assembly (1.35%).
|
WhaleVis Turns More Than a Century of Whaling Data into an Interactive Map
University of Washington News
Stefan Milne
November 15, 2023
University of Washington (UW) researchers created an online interactive map called WhaleVis from data kept by the International Whaling Commission on commercial whale catches dating back to 1880. Rather than rendering the more than 2 million data points on a global map all at once, the researchers aggregated whale catches in clusters to minimize the drain on computing power. They also built the WhaleVis tool for Web browsers, instead of as a standalone app, to allow it to function on different computing platforms. “It was important to make this data accessible so it can easily be used to generate actionable insights,” said UW's Ameya Patil.
|
Apple to Make It Easier to Text Between iPhones, Androids
Bloomberg
Mark Gurman
November 16, 2023
Apple will adopt a technological standard next year that will allow text messaging to operate more smoothly between its iOS devices and Android devices. For more than a year, the company has pushed back on the standard known as RCS (rich communication services), an upgrade over standard SMS and MMS texting that has backing from the GSM Association. The technology allows more texting features to be shared over phones on different platforms. “Later next year, we will be adding support for RCS Universal Profile, the standard as currently published by the GSM Association,” Apple said. “We believe RCS Universal Profile will offer a better interoperability experience when compared to SMS or MMS."
|
Scientists 3D-Print Hair Follicles in Lab-Grown Skin
Rensselaer Polytechnic Institute News
Samantha Murray
November 15, 2023
A team led by scientists at Rensselaer Polytechnic Institute has three-dimensionally (3D) printed hair follicles in lab-grown human skin tissue. The researchers first allowed samples of skin and follicle cells to divide and multiply in the lab until there were enough printable cells. They then mixed each type of cell with proteins and other materials to create the “bio-ink” used by the printer. The printer builds a skin layer by layer, while creating channels for depositing hair cells. The skin cells eventually migrate to these channels, duplicating the follicle structures present in real skin. Said Rensselaer’s Pankaj Karande, "This kind of automated process is needed to make future biomanufacturing of skin possible."
|
Communication Satellite 3D-Printed, Deployed in 90 Minutes
Interesting Engineering
Ameya Paleja
November 17, 2023
Researchers at Spain's Universitat Oberta de Catalunya three-dimensionally (3D) printed a nanosatellite for deployment in just 90 minutes, demonstrating a process that could aid in the restoration of wireless connectivity in areas affected by a disaster. After being printed, the nanosatellite, also known as a CubeSat, may be launched over the impacted area using a hot air balloon. The CubeSat connects two ground stations and works as a repeater, allowing ground users to communicate. The launch route of the CubeSat can be mapped using computer simulations that consider weather conditions in the target area.
|
Technique Enables AI on Edge Devices to Keep Learning over Time
MIT News
Adam Zewe
November 16, 2023
Researchers from the Massachusetts Institute of Technology (MIT), the MIT-IBM Watson AI Lab, and the University of California San Diego developed a technique that enables deep-learning (DL) models to efficiently adapt to new sensor data directly on edge devices. The on-device PockEngine training method determines the parts of a machine-learning model that need updating to improve accuracy, and only stores and computes with those specific pieces. DL models are comprised of many interconnected layers of nodes that process data to make a prediction. PockEngine fine-tunes each layer individually on a certain task and measures the accuracy improvement after each such tuning. It removes unnecessary layers or pieces of layers, creating a pared-down graph of the model to be used during runtime. Said MIT's Song Han, “On-device fine-tuning can enable better privacy, lower costs, customization ability, and also lifelong learning."
|
Shape-Shifting Robot Could One Day Perform Surgery from Inside the Body
LiveScience
Keumars Afifi-Sabet
November 14, 2023
Scientists led by Kaushik Jayaram at the University of Colorado have designed a tiny, shape-shifting "spider-bot" that can manipulate its body as it navigates its course. Dubbed "mCLARI," the robot is 0.8 inch (2 cm) long, weighs less than 0.03 ounces (0.97 grams), and can move at maximum speeds of 2.4 inches per second (6 cm/s). It has four leg modules that can move in two different dimensions, and it can maneuver in tight spaces by adopting varied gaits and speeds. Its legs hinge to its body through passive joints, which enable mCLARI to alter its body shape based on the environment it encounters. Said Jayaram, "In the long term, we envision such robots to be effective at navigating through human bodies and performing automated surgery, such as clearing an artery clog or removing a tumor."
|
Tool Models Viability of Closed-Loop Geothermal Systems
Sandia Labs News Releases
November 14, 2023
A team of experts from several U.S. national laboratories computationally modeled a closed-loop geothermal system. Closed-loop geothermal, which involves circulating subterranean fluids to generate electricity or to directly heat buildings, was dismissed in the 1980s for not being cost-effective. Sandia National Laboratories led the computational modeling of the system, while the National Renewable Energy Laboratory (NREL) used the numerical results of that modeling to estimate the economic viability of the system. Using a simplified numerical model instead of a full three-dimensional representation and running the computations on Sandia’s high-performance computing clusters allowed the researchers to model several million sets of parameters. Said Sandia's Yaro Vasyliv, "We varied seven parameters and computed corresponding outlet temperatures and pressures. You can feed that into an aboveground model that computes the levelized cost of heat and the levelized cost of electricity, which is what NREL worked on."
|
Better Machine Learning Models with Quantum Computers
IEEE Spectrum
Tammy Xu
November 15, 2023
Researchers at European quantum computing company Terra Quantum demonstrated improved training of machine learning models using a method that combines the best features of classical and quantum computers. The researchers hypothesized that by giving classical and quantum computers the same dataset and allowing them to train models in parallel, the final model combining the two could achieve better results. Said Terra's Alexey Melnikov, “Quantum is not good for everything, classical is not good for everything, but together they improve each other.” The researchers used the technique to model gas emissions at a waste-burning thermal power plant. When they added a quantum neural network layer to an existing classical model, they found the error rate of the model dropped to one-third of what it would have been without quantum.
|
|
