Welcome to the October 28, 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."
|
|
Scientists Create Synthetic Human Tongue Using 3D Printing
Daily Mail (U.K.) Jonathan Chadwick October 27, 2020
Researchers at the University of Leeds in the U.K. report the development of what they are calling the first biomimetic tongue created using three-dimensional (3D) printing. The researchers used digital scans of human tongues to 3D-print the device, whose synthetic silicone structure mimics the human tongue's elasticity, “wettability,” and texture. Said Leeds’s Efren Andablo-Reyes, "We aimed to replicate these mechanically relevant characteristics of the human tongue in a surface that is easy to use in the lab to replicate oral processing conditions." The synthetic tongue could be used to detect counterfeit food and orally administered pharmaceuticals, and to help develop new nutritional technologies.
|
Companies Make Their Websites More Elderly-Friendly
The Wall Street Journal Katie Deighton October 24, 2020
As the pandemic forces older adults to do more business online and the buying power of this demographic grows, some companies have started redesigning their websites to make them more user-friendly for older adults. One example is Boston-based prescription-discount company SingleCare, which has implemented gradual changes to its website to help users ages 50 and older, including streamlined navigation, simplified nomenclature, increased button size, and white fields where text is input. London-based Web designer John Corcoran said text sizes larger than 12 points and sans serif fonts are preferable for people with deteriorating eyesight. Jonathan Hassell of accessibility consulting firm Hassell Inclusion said contrast is another important consideration. Hassell said legislation could help make the Web better for seniors, but "getting the CEOs and information officers of all companies to take it seriously" is key.
*May Require Paid Registration
|
National Guard Called in to Thwart Cyberattack in Louisiana Weeks Before Election
Reuters Christopher Bing October 23, 2020
The Louisiana National Guard has been called in to investigate a series of cyberattacks aimed at small government offices across the state, highlighting risks facing local governments as they approach the U.S. presidential election. The attacks follow the pattern of a similar case in Washington State, in which hackers infected government offices with a type of malware known for deploying ransomware. Experts investigating the Louisiana incidents found a remote access trojan previously linked to a group associated with the North Korean government; however, that tool had been publicized in a computer virus repository, making attribution difficult. While staff at several government offices in Louisiana were compromised by the latest cyberattack, it was stopped in its early stages before significant harm was done, according to insiders.
|
Brazil Launches AI Center
ZDNet Angelica Mari October 14, 2020
Brazil's Artificial Intelligence Center (C4AI) officially launched earlier this month, thanks to investments from IBM and Brazil’s São Paulo Research Foundation and University of Sao Paulo. A major focus of the center will be working to address challenges related to health, the environment, the food production chain, the future of work, and the development of Natural Language Processing technologies in Portuguese (the language spoken by nearly all Brazilians). C4AI also will work on human well-being improvement projects and diversity and inclusion initiatives. The center's opening comes nearly a year after the Brazilian government announced plans to create a national AI strategy.
|
Activists Turn Facial Recognition Tools Against the Police
The New York Times Kashmir Hill October 21, 2020
Activists are wielding facial recognition tools to strike back against police. They say it has become relatively easy to design such tools, using off-the-shelf image-recognition software. Portland, OR-based facial-recognition developer Christopher Howell used Google's TensorFlow platform to train software to recognize local police who conceal their identities from the public. Earlier attempts to identify police officers have used crowdsourcing, as did the anti-surveillance group the Lucy Parsons Lab, which developed the OpenOversight searchable database of law enforcement officers as one such resource. Technologist Andrew Maximov last month uploaded a video to YouTube demonstrating how facial recognition technology can digitally remove masks by matching masked officers to full images of police from social media channels.
*May Require Paid Registration
|
Robot Trained in Simulation Performs Better in Real Life
New Scientist Chris Stokel-Walker October 21, 2020
Researchers at the Swiss Federal Institute of Technology, Zurich (ETH Zurich) trained a neural network algorithm designed to control a four-legged robot in a simulated environment resembling a video game. The ETH Zurich team told the algorithm which direction the simulated robot should be attempting to move in, and restricted how fast it could turn, in order to reflect the capabilities of the actual robot. The researchers started with a neural network preprogrammed with knowledge about the environment so the algorithm could absorb and recall inputs from virtual sensors, then transferred this knowledge to a large network controlling the real robot. As a result, the robot was able to move on uneven, mossy terrain more than twice as fast as it was able to with its default programming.
|
Intrinsic Plasticity Improves Learning in Deep Neural Networks
University of Waterloo Cheriton School of Computer Science (Canada) October 26, 2020
Researchers at the David R. Cheriton School of Computer Science at the University of Waterloo in Canada used what they call "intrinsic plasticity" to improve learning in deep neural networks. In deep neural networks, usable error signal declines as information is propagated through a network from one layer to the next, making it increasingly difficult to tune the parameters of earlier layers, an issue known as the vanishing gradient problem. Batch normalization is one technique that has been used to solve this problem, and Waterloo’s Nolan Shaw has found a way to use the technique locally in a neural network. Shaw says the new method “addresses why biological neurons tune themselves. The answer seems to be to improve informational efficiency. The other question is, how do biological brains perform something like batch normalization? Intrinsic plasticity provides a possible answer to that question as well.”
|
A Wearable Sensor to Help ALS Patients Communicate
MIT News Anne Trafton October 22, 2020
A stretchable, skin-like device developed by researchers at the Massachusetts Institute of Technology could help people with amyotrophic lateral sclerosis (ALS) communicate with small movements. The device, made of four piezoelectric sensors embedded in a thin silicone film, can be attached to the wearer's face. The sensors convert the mechanical deformation of the skin into measurable electric voltage. A test of the initial version of the device on two ALS patients, using a machine-learning algorithm that could distinguish a smile, an open mouth, and pursed lips with 75% accuracy, compared with 87% on healthy subjects.
|
App to Help Voters Locate Ballot Drop Boxes
George Mason University Nanci Hellmich October 22, 2020
George Mason University's Massimiliano Albanese developed the myBallotBox app to help people locate drop-off boxes for mail-in ballots. The free app provides drop-box location data for 32 states and the District of Columbia, with more added on a daily basis. The data also can be accessed through a Web interface. Users can select their state from a drop-down menu and click markers on the map to access detailed information about ballot drop-off locations. Albanese said the project “was the perfect opportunity to combine my passion for programming with my desire to use information technology for social good.”
|
Tool Predicts Risk of Hospitalization, Death From Covid-19
Independent (U.K.) Samuel Lovett October 21, 2020
Researchers in a multi-institutional U.K. study have developed an algorithm to predict the likelihood of one’s hospitalization and death from Covid-19. The research team compiled data on more than 8 million patients from 1,205 U.K. general practices, Covid-19 test results, and the Hospital Episode Statistics database. The model predicted the 5% of people predicted to be at greatest risk of Covid-19 accounted for 75% of deaths over the study period (January 24-April 30). Said the researchers, "This study presents robust risk prediction models that could be used to stratify risk in populations for public health purposes in the event of a 'second wave' of the pandemic and support shared management of risk."
|
Brain-Inspired Robot Controller Uses Memristors for Fast, Efficient Movement
IEEE Spectrum Evan Ackerman October 21, 2020
Researchers from the University of Southern California and the U.S. Air Force Research Laboratory demonstrated a brain-inspired self-balancing robot with an analog control system based on memristors (resistance-based memory devices). The researchers added a memristor to an analog circuit with inputs from a gyroscope and an accelerometer, creating a Kalman filter to strip noise from sensor signals. A second memristor converted the sensor data into a proportional-derivative controller. These two elements formed an analog system that can keep an inverted pendulum robot upright with much greater efficiency than a conventional system. The researchers said they anticipate such memristor-based hybrid computing to help “improve the robustness and the performance of mobile robotic systems with higher” degrees of freedom.
|
Optical Wiring for Large Quantum Computers
ETH Zurich (Switzerland) Oliver Morsch October 21, 2020
Researchers at the Swiss Federal Institute of Technology, Zurich (ETH Zurich) have demonstrated a method for performing sensitive quantum operations on atoms, which could enable the creation of large-scale quantum computers based on trapped atoms. The technique directs multiple laser beams precisely to the correct sites from within a processor in such a stable manner that even the most delicate quantum operations on the atoms can be executed. ETH Zurich's Maciej Malinowski said the chip used two-quantum-bit logic gates to generate entangled states “with a fidelity that up to now could only be achieved in the very best conventional experiments.” The researchers said their approach was not only extremely stable, but also scalable.
|
|