Welcome to the October 14, 2022, 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."
|
|
|
ACM Highlights Underuse of Risk-Limiting Audits in Confirming Accuracy of Election Results
Association for Computing Machinery
October 13, 2022
Despite their efficiency in confirming the accuracy of election results, risk-limiting audits (RLAs) are underused, according to a new TechBrief from ACM's global Technology Policy Council. The authors found only five U.S. states will require then in the upcoming November elections, while just 10 additional states either have RLA pilot programs or allow their use. Meanwhile, Denmark is the only other country to have performed an RLA of an election. "RLAs give us the best of both worlds: a high degree of accuracy and transparency without the enormous undertaking that is counting every contest on every ballot by hand," said TechBrief co-lead author Matthew Bernhard.
|
UW Computer Science Professor Wins MacArthur 'Genius Grant'
The Seattle Times
Amanda Zhou
October 12, 2022
University of Washington (UW) computer science professor Yejin Choi has been selected to receive one of this year's MacArthur Foundation “genius grant” fellowships. Choi applies natural language processing to develop artificial intelligence (AI) systems capable of understanding implied meanings in human languages. She has created models that absorb text and image inputs and has used computational linguistics to help AI detect sentiment or misleading intent in writing. Her team of researchers applied the latter concept to identify false online consumer reviews and to categorize news stories as hoaxes, satirical, or trustworthy.
|
Google Algorithm Helped Direct Financial Aid to Hurricane Ian Victims
Wired
Chris Stokel-Walker
October 10, 2022
A Google algorithm deployed in partnership with nonprofit GiveDirectly notified nearly 3,500 hurricane victims in Florida of $700 in financial aid via their smartphones. The algorithm used satellite imagery to identify residents in severely damaged neighborhoods who qualified for emergency assistance. Google's Delphi mapping software powers the Florida project, highlighting communities in need after disasters such as hurricanes by overlaying live maps of damage with information on poverty from sources including the U.S. Centers for Disease Control and Prevention. Google's Skai tool supplies storm-damage data by analyzing satellite images from before and after disasters and estimating the severity of damage.
|
How to Store Data for 1,000 Years
BBC News
Jocelyn Timperley
October 9, 2022
Scientists are exploring DNA as a storage medium for preserving digital data far beyond the longevity of current hardware. According to researchers, DNA has superior efficiency and durability, and keeping it cold and dry maintains its viability for centuries. The molecule's ultra-dense storage capacity means all human-made data produced by 2025 could be contained in a ping pong ball-sized DNA library. Enveloping DNA in materials like glass beads also can boost protection, and researchers suggest storing encapsulated DNA in ice vaults could make it virtually immortal.
|
Anonymous Data Doesn't Mean Private
Illinois Institute of Technology
Casey Moffitt
October 6, 2022
Illinois Institute of Technology (Illinois Tech) researchers used machine learning and artificial intelligence algorithms to exfiltrate personal information from anonymous cellphone data. The neural-network model estimated the gender and age of individual users via their private communications by tapping data from a Latin American cellphone company. The algorithm guessed individuals' gender with 67% accuracy and age with 78% accuracy, which significantly outperformed current models. Researchers employed commonly accessible computing equipment to extract this information, and although the dataset they used was not publicly available, malefactors could compile a similar dataset by capturing data through public Wi-Fi hotspots or by targeting service providers' computing infrastructure, said Illinois Tech's Vijay K. Gurbani.
|
World's Fastest Supercomputer Can't Run a Day without Failure
Tom's Hardware
Anton Shilov
October 8, 2022
Oak Ridge National Laboratory's Frontier supercomputer experiences numerous hardware failures on a daily basis. Frontier, which has not yet been deployed officially, aims to provide up to 1.685 FP64 ExaFLOPS peak performance via AMD's 64-core EPYC Trento processors, Instinct MI250X compute graphics processing units (GPUs), and HPE's Slingshot interconnections at 21 MW of power. "We are working through issues in hardware and making sure that we understand (what they are),” said Oak Ridge Leadership Computing Facility's Justin Whitt. “You are going to have failures at this scale. Mean time between failure on a system this size is hours; it's not days."
|
System Designs Heat-Conducting Nanomaterials
MIT News
Adam Zewe
October 7, 2022
Massachusetts Institute of Technology (MIT) researchers have created an algorithm and software for designing a specific type of heat-conducting nanoscale material that could be used in computer chips that disperse heat on their own. Researchers adapted computational methods for designing large structures to generate nanomaterials with defined thermal characteristics. "Imagine that we transform a material into a picture, and then we find the best pixel distribution that gives us the prescribed property," said MIT's Giuseppe Romano. Researchers have produced materials that can conduct heat along a preferred direction and that can efficiently transform heat into electricity.
|
Intel, Google Cloud Launch Chip to Improve Data-Center Performance
Reuters
Jane Lanhee Lee
October 11, 2022
A chip co-developed by researchers at Intel and Google Cloud aims to improve the security and efficiency of data centers. Code named Mount Evans, the E2000 chip handles data packaging for networking—work that previously had been done by central processing units (CPUs)—and enhances security between different customers sharing CPUs in the cloud. Google's Amin Vahdat said the E2000 is being offered in Google Cloud's C3 VM, which will be powered by Intel's fourth-generation Xeon processors.
|
Natural Language Processing Software Evaluates Middle School Science Essays
Penn State News
Mariah Chuprinski
October 11, 2022
Computer scientists at Pennsylvania State University (Penn State) and the University of Wisconsin-Madison (UW-Madison) appraised natural language processing software for evaluating students' science essays. Researchers augmented the PyrEval tool to assess concepts in student writing based on predetermined and computable rubrics. The PyrEval-CR software "can provide middle school students immediate feedback on their science essays," while also summarizing subjects or ideas in the essays "from one or more classrooms, so teachers can quickly determine if students have genuinely understood a science lesson," said Penn State's Rebecca Passonneau. Researchers tested PyrEval-CR on hundreds of science essays from Wisconsin public schools.
|
AI Offers Insight into Who Responds to Antidepressants
IEEE Spectrum
Michelle Hampson
October 10, 2022
Machine-learning algorithms developed by scientists at the New York Institute of Technology (NYIT) and Canada's McMaster University analyze the electrical activity of the brain to predict responses to antidepressants with high accuracy. Researchers analyzed electroencephalogram data from 228 patients with depression before treatment, who were randomly assigned to receive a placebo or Sertraline. The algorithms forecast Sertraline response with 83.7% accuracy and placebo response with 83% accuracy.
|
Scientists Augment Reality to Crack the Code of Quantum Systems
SciTechDaily
October 10, 2022
Quantum physicists from the Swiss Federal Institute of Technology, Lausanne and the Flatiron Institute's Center for Computational Quantum Physics (CCQ) have developed a new quantum-entanglement simulation method. Researchers accurately modeled quantum-system behavior using neural networks and "ghost" electrons that interact with the system's actual electrons. "This is a big deal because if you have a good way of getting the wave functions of complex molecules, you can do all sorts of things, like designing drugs and materials with specific properties,” said the CCQ's Antoine Georges. Georges said the ultimate goal is to be able to computationally predict material or molecular properties without requiring laboratory-based synthesis and testing.
|
Turtle Robot Can Walk on Land, Swim through Water
New Scientist
Matthew Sparkes
October 12, 2022
The Amphibious Robotic Turtle (ART) created by researchers at Yale and West Chester universities can shape-shift its limbs and gait to crawl on land and swim in water. ART's four morphable limbs each feature two pneumatic actuators and polymer layers that can be softened by heaters, boosting their cross-sectional area by up to four times. According to researchers, the turtle bot performed three times better than a bipedal robot developed by the Massachusetts Institute of Technology in terms of the cost of transport, which quantifies the amount of energy expended when moving between two points. ART could be helpful in ocean farming and environmental monitoring, or "could transport parts or tools from land into the water to divers working on repairing a submerged structure,” said Yale University’s Robert Baines.
|
Meet the App that Helps Gig Workers Know How Much They Really Make
The New York Times
Erin Griffith
October 11, 2022
Many gig economy workers are unsure how much they earn after factoring in all their costs. In response, David Pickerell, a former operations manager at Uber, left the company and built the free Para app to help gig workers maximize their earnings. Para allows drivers to see the tip amount before accepting the job, automatically reject low-paying jobs, and flag customers and locations they wish to avoid. Para also features an earnings tracker that enables gig workers who drive for multiple apps to calculate their income in a central location. However, the app has encountered resistance from gig companies. Similar apps include Muver, which helps drivers manage multiple gig apps; Solo, which provides a guaranteed daily rate for gig workers using multiple apps; and Gridwise, which tracks earnings and mileage.
*May Require Paid Registration
|
|
