Welcome to the February 24, 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."
|
|
|
Your Data is a Weapon That Can Help Change Corporate Behavior
Fortune
Jonathan Vanian; Jeremy Kahn
February 23, 2021
A study by Northwestern University researchers highlighted how the public can influence companies they think are misusing their data or engaging in unethical behavior. The study concerns data leverage, the power consumers have over firms that rely on machine learning (ML) software; people can impact businesses by changing their online behavior, like discontinuing use of such software. Northwestern's Nicholas Vincent said if people stop using a certain artificial intelligence-powered application, the app will lose the data needed for it to learn properly. Consumers also can collaboratively infect an ML system through atypical behavior, a practice known as data poisoning, and undermine the software's performance. Vincent said the research aims to show the public how their online behavior influences the AI systems of powerful technology companies.
*May Require Paid Registration
|
EU Report Warns AI Makes Autonomous Vehicles 'Highly Vulnerable' to Attack
VentureBeat
Khari Johnson
February 22, 2021
A report by the European Union Agency for Cybersecurity (ENISA) describes autonomous vehicles as "highly vulnerable to a wide range of attacks" that could jeopardize passengers, pedestrians, and people in other vehicles. The report identifies potential threats to self-driving vehicles as including sensor attacks with light beams, as well as adversarial machine learning (ML) hacks. With growing use of artificial intelligence (AI) and the sensors that power autonomous vehicles offering greater potential for attacks, the researchers advised policymakers and businesses to foster a security culture across the automotive supply chain, including third-party providers. The researchers suggested AI and ML systems for autonomous vehicles “should be designed, implemented, and deployed by teams where the automotive domain expert, the ML expert, and the cybersecurity expert collaborate."
|
Cosmologists Create 4,000 Virtual Universes to Solve Big Bang Mystery
LiveScience
Stephanie Pappas
February 22, 2021
Cosmologists used the National Astronomical Observatory of Japan’s ATERUI II supercomputer to simulate 4,000 different versions of the universe, in an effort to better understand what occurred in the immediate aftermath of the Big Bang. The observatory’s Masato Shirasaki said, "We are trying to do something like guessing a baby photo of our universe from the latest picture." The researchers created each virtual universe simulation with slightly different initial density fluctuations, then allowed them to undergo virtual inflations, before applying a reconstruction method the cosmologists had developed to remove gravitational fluctuations and “rewind” each simulation back to its starting point. Shirasaki said the process permits the cosmologists “to extract the information of initial conditions of our universe in an efficient way."
|
Toward a Disease-Sniffing Device that Rivals a Dog's Nose
MIT News
David L. Chandler
February 17, 2021
A system developed by a multi-institutional team including Massachusetts Institute of Technology (MIT) researchers can detect an air sample's chemical and microbial content with 200 times greater sensitivity than a dog's nose. When paired with machine learning, the researchers say their system can identify the properties of disease-carrying samples. The system embeds mammalian olfactory receptors stabilized to function as sensors, whose data streams can be managed in real time by a smartphone's capabilities. The system matched the success rates of disease-sniffing dogs when testing 50 urine samples from confirmed cases of prostate cancer and disease-free controls, with the system and the dogs each achieving with greater than 70% accuracy. MIT's Andreas Mershin said such odor detectors, outfitted with advanced algorithms, could identify early signs of disease faster than typical screening regimes.
|
Light Used to Detect Quantum Information Stored in 100,000 Nuclear Quantum Bits
University of Cambridge (U.K.)
February 15, 2021
Researchers at the U.K.'s University of Cambridge were able to detect a single quantum bit in a dense cloud of quantum bits, and use it to control the entire cloud. The researchers inserted quantum information into a cloud of 100,000 nuclei, then used light from a laser to communicate with an electron, which can control the spins of the nuclei. The researchers found that injecting one quantum bit with a spin wave comprised of a single nuclear spin flip enabled the detection of that spin flip among 100,000 nuclear spins. Said Cambridge's Mete Atatüre, "We don't have a way of 'talking' to the cloud and the cloud doesn't have a way of talking to us. But what we can talk to is an electron: we can communicate with it sort of like a dog that herds sheep."
|
Positive Reinforcements Help Algorithm Forecast Underground Natural Reserves
Texas A&M Today
Vandana Suresh
February 23, 2021
Texas A&M University (TAMU) and University of Oklahoma researchers have developed a reinforcement-based algorithm that automates forecasting of subterranean properties, enabling accurate prediction of oil and gas reserves. The algorithm focuses on the correct characterization of the underground environment based on rewards accumulated for making correct predictions of pressure and flow anticipated from boreholes. The TAMU team learned that within 10 iterations of reinforcement learning, the algorithm could correctly and rapidly predict the properties of simple subsurface scenarios. TAMU's Siddharth Misra said, "We have turned history matching into a sequential decision-making problem, which has the potential to reduce engineers' efforts, mitigate human bias, and remove the need of large sets of labeled training data."
|
The Appearance of Robots Affects Our Perception of the Morality of Their Decisions
University of Helsinki (Finland)
Michael Laakasuo; Tuire Korvuo; Niina Niskanen
February 19, 2021
A study recently completed under the Moralities of Intelligent Machines project found the appearance of robots influences humans' perception of their decisions' morality. The researchers determined people felt the choice made by humanoid machines iRobot and iClooney in response to the trolley dilemma was less ethically sound than the same choice when made by a human and a robot with a traditional robot-like appearance. The University of Helsinki's Michael Laakasuo said, "Humanness in artificial intelligence [AI] is perceived as eerie or creepy, and attitudes towards such robots are more negative than towards more machine-like robots. This may be due to, for example, the difficulty of reacting to a humanoid being: is it an animal, a human, or a tool?"
|
Browser-Tracking Hack Works Even When You Flush Caches or Go Incognito
Ars Technica
Dan Goodin
February 19, 2021
Some websites are using a hack that thwarts anti-tracking countermeasures by exploiting the tiny icons that sites display in users' browser tabs and bookmark lists (favicons), according to a study by University of Illinois, Chicago (UIC) researchers. The researchers said most browsers cache images in a location independent of those that store site data, browsing history, and cookies; sites can load a series of favicons on visitors' browsers that flag them over an extended period of time. The UIC team said any website can deploy the attack workflow without user interaction or consent, even when popular anti-tracking extensions are implemented. In addition, the hack utilizes resources in the favicon cache even with incognito browsing engaged, due to improper isolation practices found in all major browsers.
|
Search-and-Rescue Drone Locates Victims by Homing in on Their Phones
IEEE Spectrum
Michelle Hampson
February 23, 2021
The Search-And-Rescue DrOne (SARDO) platform developed by researchers at Germany's NEC Laboratories Europe uses off-the-shelf components, integrating aerial drones, artificial intelligence, and smartphones to find survivors using signals from their phones. SARDO utilizes a drone as a mobile cellular base station that sweeps disaster areas and conducts time-of-flight measurements, while a machine learning (ML) algorithm surveys the area and calculates the location of victims. A second ML algorithm helps locate survivors on the move by estimating each person's trajectory. In field experiments, the drone could localize missing people to within a few tens of meters in roughly three minutes per victim. NEC Laboratories Europe's Antonio Albanese said, "We built SARDO to provide first responders with an all-in-one victims localization system capable of working in the aftermath of a disaster without existing network infrastructure support."
|
AI Predicts Nonlinear Ultrafast Dynamics in Optics
Tampere University (Finland)
February 18, 2021
Researchers at Finland's Tampere University utilized artificial intelligence (AI) to predict nonlinear dynamics when ultrashort light pulses interact with matter. This can simplify the design of experiments in fundamental research and will enable the incorporation of algorithms into next-generation laser systems to ensure real-time optimization. Said Tampere's Goëry Genty, “Our team has been able to train a neural network to recognize the patterns inherent in such complex evolution.” Genty added that this approach “allows us to bypass the conventional approach of solving an underlying mathematical model, which is very time consuming and requires sometimes prohibitive memory resources."
|
A Quantum Computer Just Solved a Decades-Old Problem Three Million Times Faster Than a Classical Computer
ZDNet
Daphne Leprince-Ringuet
February 23, 2021
Researchers from quantum computing company D-Wave and Google simulated some materials as much as 3 million times faster than a classical computer could, through the process of quantum annealing. D-Wave's Andrew King said, "This work is the clearest evidence yet that quantum effects provide a computational advantage in D-Wave processors," which are based on a quantum computing technique that identifies solutions for optimization problems. The researchers used D-Wave’s 2,000-qubit system to model a programmable quantum magnetic system and simulate exotic magnetism. The researchers also were able to use a quantum annealing processor to solve a valuable materials science problem.
|
Signs of Burnout Can Be Detected in Sweat
EPFL (Switzerland)
Julie Haffner
February 15, 2021
A wearable sensing chip developed by researchers at the Swiss Federal Institute of Technology Lausanne (EPFL) and the startup Xsensio can measure the stress hormone cortisol in human sweat, which could help improve the treatment of burnout, obesity, and other stress-related conditions. Comprised of a transistor, a graphene electrode, and negatively charged aptamers (short fragments of single-stranded DNA or RNA that can bind to specific compounds) that bind to specific compounds, the chip monitors cortisol concentrations throughout the circadian cycle. EPFL's Adrian Ionescu said, "Because it can be worn, scientists can collect quantitative, objective data on certain stress-related diseases. And they can do so in a non-invasive, precise and instantaneous manner over the full range of cortisol concentrations in human sweat."
|
|
