Welcome to the October 7, 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."
|
|
Tweeting About Trump, Searching for Biden: Online Activity Shows Contrast Between the Candidates
New York University October 5, 2020
Researchers at the Courant Institute of Mathematical Sciences of New York University (NYU) analyzing online activity leading up to, during, and immediately after last week's presidential debate found higher concentrations of tweets for Donald Trump, and more Google searches for Joe Biden. NYU's Anasse Bari said, "The online community was much more likely to mention President Trump on Twitter—a platform for expression—than it was to mention Joe Biden." Meanwhile, more Google searches in all U.S. states focused on Biden, suggesting the debate sparked interest in him. Trump was mentioned 43% more frequently than Biden on Twitter, while Biden appeared in more than twice as many Google searches as Trump in each state.
|
Researchers Use Eye Tracking to Discover How Mobile Apps Grab Our Attention—With Surprising Results
SciTechDaily October 6, 2020
An international study by researchers at Finland's Aalto University, India's IIT Goa, Turkey's Yildiz Technical University, and China's Huawei Technologies has yielded insights into how mobile applications capture attention. The researchers used eye tracking and mobile interfaces to empirically test how users' eyes follow commonly used mobile app elements. The results contradicted assumptions that users' eyes should not only jump to larger or brighter elements, but linger there longer; they also ran counter to the idea that attention is drawn to the center of screens and spreads horizontally with certain images. The researchers further verified that other mobile design elements align with expectations on visual salience: for example, gaze drifts to the top-left corner, indicating exploration or scanning.
|
This 'Squidbot' Jets Around, Takes Pics of Coral, Fish
UC San Diego News Center Ioana Patringenaru October 6, 2020
Engineers at the University of California, San Diego (UCSD) have constructed a self-powered squid-like robot that can swim via water-jet propulsion and record data on coral and fish with a camera or other sensor. UCSD's Michael T. Tolley said, "This is the first untethered robot that can generate jet pulses for rapid locomotion ... and can achieve these jet pulses by changing its body shape, which improves swimming efficiency." The robot is built from soft materials, along with some rigid, three-dimensionally-printed, and laser-cut components. It absorbs water into its body while storing elastic energy in its skin and flexible ribs, then releases this energy by compressing its body, propelling itself with jets of water. Testing showed the squid robot achieves a speed of roughly 18 to 32 centimeters per second (about 0.5 mph).
|
Algorithm Could Unleash the Power of Quantum Computers
Los Alamos National Laboratory News October 5, 2020
Researchers at the U.S. Department of Energy's Los Alamos National Laboratory have developed an algorithm that accelerates simulations, which could make current and near-term quantum computers more usable. The Variational Fast Forwarding algorithm integrates aspects of classical and quantum computing, and tolerates small calculation errors for intermediate times, in order to deliver useful predictions. Los Alamos' Andrew Sornborger said, "We use machine learning to create a quantum circuit that can approximate a large number of quantum simulation operations all at once. The result is a quantum simulator that replaces a sequence of calculations with a single, rapid operation that can complete before quantum coherence breaks down.”
|
Python Set to Overtake Java in Latest Programming Language Rankings
ZDNet Liam Tung October 5, 2020
Tiobe's latest monthly rankings indicate Python is poised to dethrone Java as the world's second-most-popular programming language after C, having declined 4.32% year-on-year in October. If this happens, Java would end up outside the top two spots for the first time since the language popularity index's 2001 inception. Tiobe CEO Paul Jansen in September suggested Python "is in real trouble" due to its year-on-year 3.81% decrease in popularity, which previously soared because of its use by data scientists and the advent of machine learning. The ranking dovetails with RedMonk's own index, in which JavaScript dropped below the top two rankings for the first time in July.
|
NASA Research App Helps Fight Fatigue in Flight, Space, More
NASA Abby Tabor October 6, 2020
The U.S. National Aeronautics and Space Administration (NASA) has released a research application on the Apple App Store to help scientists study what happens when someone is fatigued, and to identify signs that can be used to alert people to their fatigue. The app is based on the psychomotor vigilance task, which measures reaction time to a visual signal, modified for a portable device. The app also includes a simplified process for data collection, so those working in real-world situations can gather valuable information to be applied to designing reliable tests and solutions. NASA has used the app for spaceflight simulations.
|
Deep Learning Gives Drug Design Boost
Rice University Mike Williams October 5, 2020
Rice University researchers have developed a deep learning-based translator for predicting chemical reactions in the human body, which could potentially augment drug design. Rice's Lydia Kavraki and colleagues unveiled Metabolite Translator, a computational tool that predicts the production of metabolites, the results of interactions between small molecules like drugs and enzymes. The researchers trained Metabolite Translator to predict metabolites through any enzyme, but quantified its success against current rules-based techniques concentrating on liver enzymes that detoxify and eliminate xenobiotics. Said Kavraki, "We are training a system to understand human metabolism without the need for explicitly encoding this knowledge in the form of rules."
|
Researchers Fingerprint Exploit Developers Who Help Several Malware Authors
The Hacker News Ravie Lakshmanan October 2, 2020
Cybersecurity researchers from Check Point Research have deployed a methodology to determine the unique characteristics of a malware author and use that ‘fingerprint’ to identify other exploits developed by the same person. Identifying characteristics can include the use of hard-coded values or string names, how the code is organized, or how certain functions are implemented. The technique enabled the researchers to link 16 Windows local privilege escalation exploits to two zero-day sellers known as "Volodya" and "PlayBit." Said the researchers, "Both of our actors were very consistent in their respective exploitation routines, each sticking to their favorite way." The researchers believe the methodology could be used to identify additional exploit writers.
|
What If Underwater Robots Could Autonomously Dock Mid-Mission to Recharge, Transfer Data?
Purdue University News Jared Pike October 6, 2020
A team of Purdue University researchers has created a mobile docking system for autonomous underwater vehicles (AUVs), extending their performance without human intervention. Marine robots typically follow a preplanned itinerary, then surface for retrieval, data uploading, and recharging; the Purdue team's portable dock can be deployed at stationary locations, and on autonomous surface vehicles or on other AUVs. Purdue's Nina Mahmoudian said, "The robots and the docks could coordinate with each other, so that they could recharge and upload their data, and then go back out to continue exploring, without the need for human intervention." She envisions the platform-agnostic system finding use on land, sea, air, and even in space, enabling robots to function autonomously indefinitely.
|
Augmented Reality Offers Promise of Incision-Free Autopsies
The Wall Street Journal Rachel Pannett October 1, 2020
Forensic and digital healthcare experts envision the incorporation of virtual/augmented reality into forensic science, with digital reconstructions and machine learning algorithms diagnosing cause of death, identifying victims, and even triaging injuries in living patients. Critical to such advances are computed tomography (CT) scans that generate cross-sectional images of the body; Australia's Victorian Institute of Forensic Medicine has compiled a database of about 80,000 CT scans covering a wide range of causes of deaths. The institute is working with Australia's Monash University and U.S. biomedical and defense company Leidos Holding on an incisionless autopsy method that will three-dimensionally (3D) reconstruct a dead person’s body digitally.
*May Require Paid Registration
|
An Open Source Bionic Leg
IEEE Spectrum Charles Q. Choi October 5, 2020
University of Michigan researchers have released details online of the design and clinical tests of an open source bionic leg, to encourage scientists to create and test new prosthetics. The Open Source Leg includes free-to-copy step-by-step guides to help researchers build or order parts for the artificial limb, as well as videos on assembling and testing the hardware; also available is code for programming the leg to walk using a preliminary control system. The Leg’s knee and ankle joints can operate independently, enabling research on patients with above-knee and below-knee amputations. Each joint also has onboard batteries and individual sensing and control systems, to enable testing outside the laboratory. The leg boasts lightweight, brushless electric motors developed for drones that add torque, allowing users to walk with less fatigue.
|
This Interactive VR 'Film' Lets You Play God to Evolving AI
Mashable Joseph Volpe September 30, 2020
Toronto-based independent studio Transitional Forms and the National Film Board of Canada have created "Agence," an interactive "dynamic film" that combines virtual reality (VR), gaming, and cinematic storytelling that allows audiences to direct evolving artificial intelligence (AI) agents. The viewer acts as a godlike omnipresent entity who can either observe or influence five agents as they respond to the introduction of an object of desire (an intoxicating flower) that either dooms them or ensures their survival. The agents are gaming and reinforcement learning AIs that evolve over millions of interactions. Transitional Forms founder Pietro Gagliano said, "These creatures that are in the film...they’re probably not even as smart as insects. But they’ll evolve over time with our work and hopefully the work of future collaborators, if people are into this idea.”
|
Smartwatches Could Predict Your Marathon Time With Increased Accuracy
New Scientist Layal Liverpool October 6, 2020
Researchers at the French National Center for Scientific Research (CNRS) have developed a mathematical model for more accurately predicting race times from smartwatches. Most smartwatch measurements of VO 2 max (the maximum rate at which runners consume oxygen during exercise) from heart rate readings are inaccurate, and can lead to errors of up to 20%. The CNRS team's model employs smartwatch data to calculate a runner's speed at maximum oxygen uptake, and the rate at which they lose power during a race. The researchers tested the model with smartwatch data from about 14,000 runners, and could predict people's marathon times to within an average 10% of actual time (and to within less than 5% of actual time for elite athletes). Lukasz Malek at Poland's National Institute of Cardiology said more accurate race-time predictions could potentially lower the risk of overtraining, as well as "the risk of potentially life-threatening incidents."
|
|