Welcome to the October 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."
|
|
|
Report Shows It Could Take 12 Years to Reach Equal Representation of Women in Tech
CNBC
Courtney Connley
September 29, 2020
According to AnitaB.org's 10th annual "Top Companies for Women Technologists" report, women technologists account for 28.8% of the technology workforce this year, up from 25.9% in 2018 and 26.2% in 2019. If the current rate of growth were to continue, the global organization for women technologists says it would take 12 years for women to attain equal representation in the technology industry. Women accounted for an average of 30.2% of employees at small companies (less than 1,000 employees), 29.6% at medium-sized companies (1,000-10,000 employees), and 27% at large companies (more than 10,000 employees). White women accounted for 14.1% of the technology workforce, followed by Asian women (9.6%), Black women (2.2%), and Latinx women (1.7%). The survey was based on data on more than 500,000 U.S. technologists from 51 participating companies.
|
Detector Breakthrough Pushes Boundaries of Quantum Computing
Aalto University
September 30, 2020
Physicists at Finland's Aalto University and VTT Technical Research Center have developed a detector that measures energy quanta at unprecedented resolution, in a step toward real-world quantum computing. The team hopes the detector, known as a bolometer, is sufficiently fast and sensitive to measure quantum-bit (qubit) energy and overcome issues related to voltage measurements in quantum computing scalability, power consumption, and quantum noise. Said Aalto’s Mikko Möttönen, “Bolometers are now entering the field of quantum technology and perhaps their first application could be in reading out the quantum information from qubits. The bolometer speed and accuracy seems now right for it."
|
Engineers Link Brains to Computers Using 3D Printed Implants
University of Sheffield (U.K.)
September 21, 2020
Engineers at the U.K.'s University of Sheffield, Russia's St. Petersburg State University, and Germany's Technical University of Dresden have developed a prototype three-dimensionally (3D)-printed implant that can be used to create brain-computer interfaces. The method would allow a neuroscientist to order a design that engineers can convert into a computer model, which then is 3D-printed using a palette of biocompatible, mechanically soft materials. Sheffield's Ivan Minev said, "Patients have different anatomies and the implant has to be adapted to this and their particular clinical need. Maybe in the future the implant will be printed directly in the operating theatre while the patient is being prepared for surgery.”
|
Hacker Releases Information on Las Vegas-Area Students After Officials Don’t Pay Ransom
The Wall Street Journal
Tawnell D. Hobbs
September 28, 2020
A hacker who locked computer servers in Las Vegas' Clark County School District with ransomware released documents with Social Security numbers, grades, and other stolen private data after officials refused to pay the ransom. The district is the largest known to be hit by hackers during the Covid-19 crisis, and marks an escalation in tactics for hackers who exploit schools heavily dependent on online learning and technology. Some school districts have made online learning their sole educational option during the pandemic; experts said this compounds the impact of ransomware and attackers' demands. Threat analyst Brett Callow at cybersecurity company Emsisoft said, "A big difference between this school year and last school year is they didn't steal data, and this year they do."
*May Require Paid Registration
|
BGU Designs Wearable System to Predict Seizures
American Associates, Ben-Gurion University of the Negev
September 29, 2020
Researchers at Israel's Ben-Gurion University of the Negev (BGU) have developed a wearable seizure-prediction system for epileptics, particularly those for whom medication does little to control their seizures. The Epiness system employs an array of scalp-mounted electroencephalography (EEG) electrodes that monitor the electrical activity of the brain, and a connected microprocessor running machine learning-based algorithms. The team trained the algorithms on EEG data from a dataset of epilepsy patients, enabling 97% accurate predictions of seizures up to an hour prior to onset. The algorithms can distinguish between the brain's electrical signals and background noise, which allows Epiness to run with 95% accuracy even when using just a few electrodes.
|
Study Indicates Neither Algorithmic Differences Nor Diverse Data Sets Solve Facial Recognition Bias
Venture Beat
Kyle Wiggers
September 28, 2020
A study by researchers at Wichita State University found that facial recognition models fail to recognize Black, Middle Eastern, and Latino people more often than lighter-skinned individuals. The researchers concentrated on the VGG, ResNet, and InceptionNet models, which were pretrained on 1.2-million images from the open source ImageNet dataset. Images from the UTKFace and FairFace facial recognition datasets were used to tailor each model for gender classification. The researchers calculated an accuracy rate around 91% for all three models in terms of gender classification, but model performance varied depending on the person's race. The researchers said the bias they found in gender classification "is not due to a particular algorithm,” adding that the results they achieved “suggest that a skewed training dataset can further escalate the difference in the accuracy values across gender-race groups."
|
Jaguar, NTT Team Up with Tech Group on Remote Access Software
Reuters
Gertrude Chavez-Dreyfuss
September 30, 2020
German technology group IOTA has announced a partnership with corporations that include automaker Jaguar Land Rover, chipmaker STMicroelectronics, and Japan's NTT Data system integration firm, to launch software that would enable consumers to remotely access and control devices like cars and appliances. The IOTA Access software also can grant and revoke conditional access for use of the remote devices by individuals other than the authorized user. IOTA’s Dominik Schiener said the software will be embedded in Jaguar vehicles. Said Schiener, “With IOTA Access, we have to focus on getting more to the consumer. You want to be able to use your phone to access anything. Our phones should be our personal keys.”
|
IBM to Open-Source Space Junk Collision Avoidance
ZDNet
Greg Nichols
October 1, 2020
IBM and researchers at the University of Texas at Austin are collaborating on an open source project to determine and predict the orbit of anthropogenic space objects (ASOs)—man-made space debris—in order to avoid collisions. Current methods for orbit prediction rely on physics-based models using location data on ASOs from terrestrial-based sensors, which tends to be imperfect. The Space Situational Awareness project uses machine learning (ML)-generated models that learn when physical models incorrectly predict an ASO's future location. The project employs data from the U.S. Strategic Command through the space-track.org website, with IBM hardware running physical models to anticipate the orbits of all ASOs in low earth orbit and train ML models on the physics model error.
|
Take-Off for U.K.-Built Supercomputer Nanosatellites
Gov.UK
September 28, 2020
The U.K. on Monday launched four shoebox-sized "nanosatellites" into low Earth orbit. Two of the Spire satellites are equipped with supercomputers running machine learning algorithms that can make hyper-accurate shipping predictions to support, maritime trade; the other two will be used to enable inter-satellite connections and to relay data between each other and ground stations on Earth. Built by Spire Global UK, the nanosatellites were funded by the U.K. Space Agency at a cost of £10 million ($12.8 million), as part of the collaborative Pioneer program with the European Space Agency (ESA). ESA's Elodie Viau said the program aims "to support industry to deliver fast and affordable access to space in the field of telecommunications."
|
Pentagon Is Clinging to Aging Technologies, House Panel Warns
The New York Times
David E. Sanger
September 29, 2020
A U.S. House of Representatives panel warned that the Pentagon is still using weapons systems rendered irrelevant by technologies like artificial intelligence (AI), quantum computing, space, and biotechnology. The Future of Defense Task Force is focusing on the next three to five decades, and concluded that the U.S. Department of Defense and Congress should be "focused on the needs of the future and not on the political and military-industrial loyalties of the past." They recommended the U.S. follow an AI initiative modeled after the Manhattan Project, and overcome cultural resistance to the implementation of new technologies. The panel urged every major military acquisition program to assess at least one AI or autonomous alternative before funding, and for the nation to spearhead development and ratification of a Geneva Convention-like global AI pact.
*May Require Paid Registration
|
Scientists Train Computers to Recognize Which Early Stage Breast Cancers Will Spread
University of Michigan
Ian Demsky
September 25, 2020
A new diagnostic approach developed by researchers at the University of Michigan Rogel Cancer Center uses artificial intelligence to help predict with high confidence whether breast cancers, known as ductal carcinoma in situ (DCIS) or stage 0, are likely to recur after surgery, or if surgery likely will cure them. The researchers created a library of microscope images associated either with aggressive or non-aggressive DCIS, then showed the machine learning program about 100 different micrographs to gauge its accuracy in predicting cancer recurrence. The program, which is continually refined using additional samples, now can identify aggressive and non-aggressive disease correctly 96% of the time.
|
Looking at Evolution's Genealogy From Home
University of Münster
September 28, 2020
Researchers at Germany's University of Münster have developed a Web application for comparing human and animal genomes. Their 2-n-way app can compare any genomes from and for anyone, and systematically seek regions characterized by the presence or absence of certain DNA sequences, enabling the recognition of genealogical relationships among species or individuals. Users only have to download the genomes to be compared on the Internet, and 2-n-way aligns them with each other. Münster's Jürgen Schmitz said, "The tool is a response to the modern genomic era—and it is a piece of software which, despite the complexity behind it, can be used by anyone, whether a non-medical layperson, a student, or a professor."
|
|
