Welcome to the December 2, 2020 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."
|
|
|
'The Game Has Changed.' AI Triumphs at Solving Protein Structures
Science
Robert F. Service
November 30, 2020
Researchers at artificial intelligence company DeepMind successfully modeled the three-dimensional structure of proteins, which the European Bioinformatics Institute's Janet Thornton predicted "will change the future of structural biology and protein research." DeepMind won the 2018 Critical Assessment of Protein Structure Prediction (CASP) competition to predict protein structure from amino acid sequences with its deep learning AlphaFold algorithm; the latest iteration combines deep learning with a tension algorithm trained on approximately 170,000 known protein structures. This year's CASP saw AlphaFold achieve a median global distance test score of 92.4, up from about 60 in 2018. The algorithm also scored a median of 87 for the most challenging proteins, 25 points above the next best prediction. Andrei Lupas at Germany's Max Planck Institute for Developmental Biology said with AlphaFold, "The game has changed."
|
Researchers Develop Algorithm From Disease Simulation
Daily Evergreen (WA)
Bradley Gamble
December 1, 2020
Researchers at Washington State University (WSU) and the U.S. Department of Energy's Pacific Northwest National Laboratory have developed an algorithm that can identify regions where vaccines should be distributed to boost immunization. WSU's Anantharaman Kalyanaraman said the program employs computer-modeled populations to calculate likely disease transmission through direct contact with others, and identifies which groups of people to inoculate to halt further infection. He added that the team used high-performance supercomputers to expand the number of scenarios it could model, as higher performance speed permits more situations to be tested faster.
|
This 3D Printer Doesn't Gloss Over the Details
MIT News
Daniel Ackerman
December 2, 2020
Researchers at the Massachusetts Institute of Technology (MIT), Princeton University, Germany's Max Planck Institute, and Switzerland's University of Lugano have developed a three-dimensional (3D) printing system that uses commercially available varnishes to give printed objects realistic, spatially varying gloss patterns. The printer is equipped with large nozzles so it is able to deposit varnish droplets of varying size. A software pipeline controls output, with users entering the desired gloss pattern on the surface to be printed, then calibrating to try out various halftoning patterns with the three supplied varnishes. The printer determines the optimal pattern for the final print job based on the reflectance of those calibration patterns.
|
Bioterrorists Can Trick Scientists Into Making Dangerous Toxins with Malware
The Daily Mail
Stacy Liberatore
November 30, 2020
Cybersecurity researchers at Israel's Ben-Gurion University of the Negev (BGU) found bioterrorists can infect biologists' computers with malware that can trick them into altering sequences of synthetic DNA in order to produce dangerous toxins or viruses. The U.S. Department of Health and Human Services (HHS) has procedures for screening DNA orders from synthetic gene providers, but the BGU team circumvented this safeguard via obfuscation; 16 of 50 obfuscated DNA samples went undetected when screened using "best-match" HHS guidelines. The BGU team performed a proof of concept in which an obfuscated DNA encoding a toxic peptide went undetected by screening software. BGU's Rami Puzis recommended a revised screening algorithm that accounts for in vivo gene editing.
|
Search and Rescue Drones Use AI to Find People Lost in Woods
Interesting Engineering
Loukia Papadopoulos
November 28, 2020
Researchers from Austria's Johannes Kepler University have developed drones equipped with a deep learning application that improves the images they collect during search and rescue missions to better distinguish people from their surroundings. The team noted, "automated person detection under occlusion conditions can be notably improved by combining multi-perspective images before classification." The researchers achieved 96% precision and 93% recall rates with image integration using airborne optical sectioning, a synthetic aperture imaging technique that captures unstructured thermal light fields using camera drones, compared to 25% achieved by traditional thermal imaging. The researchers say the drones are ready for use.
|
Disney Is Teaching Robots How to Sculpt with Clay
Gizmodo
Andrew Liszewski
December 1, 2020
A team of researchers from Disney Research and the Swiss Federal Institute of Technology, Zurich have designed and built a system that allows human artists to train robots to sculpt clay with tools that allow them to emulate artistic styles. RobotSculptor employs a six-axis robot arm with a custom sculpting tool controlled by software, which lets artists first upload a three-dimensional (3D) model of the desired rendering in clay, then define the overall look of the final sculpture and how the robot should move and carve. The software processes the model and the artists' defined parameters to formulate a series of sculpting motions that will accurately replicate the model in the desired style.
|
AI Can Turn Spoken Language Into Photorealistic Sign Language Videos
New Scientist
Chris Stokel-Walker
November 27, 2020
Researchers at the U.K.'s University of Surrey employed a neural network to render spoken language as sign language, and to map the signs onto a three-dimensional model of the human skeleton. Surrey's Ben Saunders and colleagues also trained the SignGAN system on videos of sign language interpreters, educating it to generate a photorealistic video of anyone signing based on an image of them. The artificial intelligence (AI) combines video and skeletal images to convert spoken words into sign language. Because interpreters' hands in the training videos could sometimes be blurry, the Surrey team used an existing AI that could estimate hand poses from a small area around the middle knuckle. When 10-second clips of the videos were shown to 46 people, about 25% of whom were signers, all favored SignGAN over other AI models.
|
A Microscope for Everyone: Team Develops Optical Toolbox for Research, Education
Leibniz Institute for Solid State and Materials Research (Germany)
November 25, 2020
Collaborators from Germany’s Leibniz Institute of Photonic Technology (Leibniz-IPHT), Friedrich Schiller University Jena, and Jena University Hospital have created an optical toolbox that can be used to build high-resolution microscopes for biological imaging for only a few hundred euros. A simple three-dimensionally (3D) printed cube that can incorporate lenses, LEDs, cameras, or other components is the basic building block of the UC2 (You. See. Too.) system. Said Leibniz-IPHT's Barbora Maršíková, "The components can be combined to form a projector or a telescope; you can build a spectrometer or a smartphone microscope." The system's construction plans and software have been published on GitHub.
|
Researchers Design Software to Automate Traffic Surveys
University of Western Australia
November 24, 2020
An artificial intelligence system developed by researchers at the University of Western Australia (UWA) aims to automate traffic surveys in order to obtain better data and ultimately improve road designs. The Video Content Analytics software developed in partnership with Main Roads Western Australia, uses computer vision to extract traffic data automatically. The software allows more extreme angles than drones that fly almost directly above an intersection. Main Roads found the software can detect and track vehicles under challenging conditions with 97% accuracy. Said UWA's Mark Reynolds, "Better data leads to better modeling, which leads to better intersection and road design, and more efficient use of government investment."
|
Do You Have a Conflict of Interest? This Robotic Assistant May Find It First
The New York Times
Dalmeet Singh Chawla
November 23, 2020
Frontiers, a Swiss publisher of open-access journals, has rolled out the Artificial Intelligence Review Assistant (AIRA) to check for potential conflicts of interest. The software flags whether the authors of a manuscript, as well as the editors and peer reviewers handling it, have been co-authors on previous papers. Said Frontiers' Kamila Markram, "AIRA is designed to direct the attention of human experts to potential issues in manuscripts." Other publications are using similar artificial intelligence (AI) tools, but some researchers note that conflicts of interest can be subjective and difficult to unravel. Canada's McGill University computer scientist Kaleem Siddiqi noted, "There's no real solution."
|
Companies Urged to Adjust Hiring Requirements for Cyber Jobs
The Wall Street Journal
James Rundle
November 30, 2020
The International Information System Security Certification Consortium estimated that roughly 3.1 million professionals are needed to close a cybersecurity talent gap, but the ability to reach that number is hobbled by what some consider excessive job requirements. The consortium said global cyber employment must grow 89% to fulfill security requirements, but Forrester analyst Chase Cunningham cited requirements for years of experience and professional certifications as obstacles. A September report from the federal Cyberspace Solarium Commission urged federal agencies to consider candidates with nontraditional academic or professional backgrounds, while apprenticeships and firm career development pathways for new cybersecurity professionals would help amend hiring practices.
*May Require Paid Registration
|
Researchers Bring Paradigm-Shifting Technology to Endoscopic Procedures
Vanderbilt University
Marissa Shapiro
November 19, 2020
Researchers at Vanderbilt University and the U.K.'s University of Leeds have designed an intelligent and autonomous Magnetic Flexible Endoscope, with the potential of making colonoscopies more comfortable, affordable, and more widely available. Vanderbilt's Keith Obstein and Leeds' Pietro Valdastri redesigned an endoscope-pulling mechanism by attaching magnets at the front end of the device and to a robotic arm outside of the body; this enables the endoscope to be manipulated to an accuracy of five millimeters and six degrees in real time. Experiments showed that both neophyte and experienced operators could efficiently perform semi-autonomous navigation colonoscopies, with operator and robotic process collaborating. Valdastri said, "With the deployment of autonomous and intelligent medical devices, this work becomes safer for operators and more patient-centric."
|
Accurate, Efficient 3D Motion Tracking Using Deep Learning
Tohuku University (Japan)
November 27, 2020
At Japan's Tohoku University, researchers used a flexible magnetic flux sensor array to record three-dimensional (3D) motion data via deep learning and a novel structure-aware temporal bilateral filter. The team used a deep neural network to learn the regression from the simulation flux values to LC coils' 3D configuration at any location and orientation. The bilateral filter compensates the data to rebuild smooth and accurate motion, while observation time can be maximized since markers require no batteries. The integrated system can track multiple LC coils at 100-hertz speed at millimeter-scale accuracy, while the system's self-learning capability can reconstruct dead-angle-related tracking loss. Tohoku's Yoshifumi Kitamura said, "The application of our research is widespread. Hand motions can be tracked to make creating smooth animations easier, markers can be put into fluids to track its flow, and tracking can be placed on small animals."
|
|
