Welcome to the November 22, 2021 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 Gordon Bell Special Prize Awarded to Japanese Team for Novel Aerosolized Droplet Simulation
ACM
November 18, 2021
A team of Japanese scientists was awarded the 2021 ACM Gordon Bell Special Prize for High Performance Computing-Based COVID-19 Research for the novel aerosolized droplet simulation they produced on Japan's Fugaku supercomputer. The researchers ran various models of how the virus might propagate from person to person via aerosolized droplets using techniques that combined massive scalability with high resolution, while also effecting a fast time to solution. The models covered real-world settings of public areas like classrooms, concert halls, restaurants, commuter trains, and airports. The simulations were applied to risk assessments and countermeasures that were circulated to various stakeholders, which informed safety guidelines.
|
You Should See Her in a Crown. Now You Can See Her Face.
The New York Times
Jennifer Pinkowski
November 17, 2021
Joana Bruno at Spain's Autonomous University of Barcelona digitally replicated the faces of 36 people buried at the ancient European settlement of La Almoloya, located in what is now Murcia, Spain. Bruno used software to combine facial reconstruction techniques and anatomical knowledge to produce a series of gray-toned faces in profile, "to see if the resemblance between certain traits could point us towards a shared genetic relationship," she said. The methodology indicated a 3,700-year-old double burial at La Almoloya was a husband and wife, with the man's daughter buried nearby. Analysis of mitochondrial DNA determined the women and girls were mostly descended from local people, while the men and boys were not.
*May Require Paid Registration
|
Early Warning System Model Predicts Cancer Patients' Deterioration
The Source (Washington University in St. Louis)
Beth Miller
November 17, 2021
Chenyang Lu and colleagues at Washington University in St. Louis have developed a machine learning predictive model for hospitalized cancer patients that integrates heterogeneous data in electronic health records. The end-to-end CrossNet recurrent neural network model combines static data collected at time of admission with time-series data gathered repeatedly during the hospital stay. CrossNet learns to forecast deterioration while accurately entering any missing static or time-series data. Said Lu, “Humans cannot see these hidden patterns or trends in the data, so this is where machine learning is very good at picking up these patterns.” The research is part of an effort to design an early warning system for predicting cancer patients' deterioration while hospitalized, as well as enhancing patient outcomes.
|
U.K., U.S. Join Forces to Strike Back in Cyberspace
BBC News
Gordon Corera
November 19, 2021
The U.S. and U.K. announced a collaboration to strike back against mutual adversaries engaging in malicious cyber-activities, to address "evolving threats with a full range of capabilities." At an annual meeting of intelligence chiefs, U.K. General Sir Patrick Sanders and Government Communications Headquarters director Sir Jeremy Fleming and U.S. Cyber Command director General Paul Nakasone "reaffirmed" their pledge to jointly disrupt and deter new and emergent cyberthreats. They said they would accomplish this "by planning enduring combined cyberspace operations that enable a collective defense and deterrence and impose consequences on our common adversaries who conduct malicious cyber-activity." The "persistent engagement" approach used by the U.S. entails contesting adversaries daily to disrupt their cyberattack infrastructure, while Britain suggests a similar strategy with the launch of its National Cyber Force.
|
Reading the Mind of a Worm
Salk News
November 19, 2021
Scientists can learn which chemical a worm smelled seconds before by studying its brain activity, through research conducted at the Salk Institute for Biological Studies. Salk's Sreekanth Chalasani and colleagues studied the neuronal reactions of the Caenorhabditis elegans worm to smelling five different chemicals. They had developed worms in which each of their 302 neurons contains a fluorescent sensor that brightens when the neuron is active, then exposed 48 of the worms to repeated bursts of five chemicals; an average of 50 to 60 neurons activated in response to each chemical. The researchers applied graph theory to differentiate between the chemicals, then used a machine learning algorithm to make finer distinctions in neural response. "It's a start, but we're still only getting a partial answer as to how the brain discriminates these things," said Chalasani.
|
Robot Waiters Take Iraq's Mosulites Back to the Future
France24
November 22, 2021
Two robot waiters serve diners at the White Fox restaurant in Mosul, Iraq. The robots were imported from an unnamed source and programmed by researchers from the University of Mosul's department of mechatronics, who also connected the robots to a network and server in the restaurant. The robots deliver food orders to customers who order from touchscreens with built-in menus at the restaurant’s 15 tables. The robots have not entirely displaced the human staff; a team of four waiters assists the robots by moving food items from their trays to customers’ tables.
|
This Tool Protects Your Private Data While You Browse
UC San Diego News Center
Ioana Patringenaru
November 18, 2021
The SugarCoat tool developed by researchers at the University of California, San Diego (UCSD) and Brave Software can better protect users' private data as they browse the Web. SugarCoat targets scripts that attack privacy but are critical to Website operations, replacing them with safe scripts that possess the same properties. The open-source SugarCoat is configured to integrate with existing privacy focused browsers like Brave, Firefox, and Tor, and browser extensions like uBlock Origin. UCSD's Michael Smith said, "SugarCoat integrates with existing content-blocking tools like ad blockers, to empower users to browse the Web without giving up their privacy."
|
Microrobotic Trajectory Tracking Method Using Broad Learning System
Chinese Academy of Sciences (China)
Li Yuan
November 19, 2021
Researchers at the Chinese Academy of Sciences have developed a broad learning system (BLS)-based technique for tracking microrobots' trajectories flexibly and accurately. The researchers designed a learning-based microrobot servo control algorithm that derived the controller parameters' limitations when combined with the Lyapunov theory; they also created a BLS-based controller training algorithm that employs multiple tracking trajectories as demonstrations, and used it to acquire controller parameters. Models and experimental results found the BLS-based technique could train the controller in about six seconds.
|
Researchers Develop Method for Early Detection of Bacterial Infection in Crops
Singapore-MIT Alliance for Research and Technology (SMART)
November 18, 2021
A novel Raman spectroscopy-based method developed by researchers at the Singapore-Massachusetts Institute of Technology Alliance for Research and Technology (SMART) and Temasek Life Sciences Laboratory (TLL) can detect and quantify early-stage bacterial infection in crops. The diagnostic algorithm harnesses Raman spectral biomarkers and can easily be deployed in cloud-based computing and prediction platforms. TLL's Rajani Sarojam said experiments on edible plants showed the technique can rapidly identify and measure innate immunity response in plants infected with bacterial strains. Said TLL's Chua Nam Hai, "The early detection of pathogen-infected crop plants is a significant step to improve plant disease management. It will allow the fast and selective removal of pathogen load and curb the further spread of disease to other neighboring crops."
|
ML Helps Locally Restore Wetlands for Coastal Protection
Royal Netherlands Institute for Sea Research (Netherlands)
November 16, 2021
Researchers in China, the Netherlands, Britain, and Belgium collaborated on the use of machine learning (ML) to convert an experimental dataset into models that can forecast coastal wetland restoration for flood protection. "From the large datasets generated in recent field and laboratory experiments, we know that the establishment process of wetland vegetation is complex and depending on a diverse set of factors in its living environment," said Zhan Hu of China's Sun Yat-Sen University. The ML model showed marsh establishment can be managed well despite continuing global climate change. The model also revealed that "It is especially the sediment supply, the local wave height, and shape of the tidal flat in front of the marsh that we need to control to counteract the threats of changing wind climate and rising sea level," Hu said.
|
Supercomputers Flex Their AI Muscles
IEEE Spectrum
Samuel K. Moore
November 20, 2021
Japan's Fugaku supercomputer and the Summit supercomputer at the U.S. Department of Energy's Oak Ridge National Laboratory each have surpassed the exascale benchmark High Performance Linpack for Accelerator Introspection (HPL-AI). By this measure, Fugaku has maintained its 2-exaflop performance from June, while Summit's performance improved 23%, reaching 1.4 exaflops. Meanwhile, industry organization MLCommons released the results from version 1.0 of its high-performance computing benchmarks (MLPerf HPC), which demonstrated a four- to seven-fold improvement in the time needed to train machine learning models to execute scientific tasks. Eight participating supercomputing centers produced 30 benchmark results in testing three neural networks.
|
|
