Welcome to the December 4, 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."
|
|
Light-Based Quantum Computer Exceeds Fastest Classical Supercomputers
Scientific American Daniel Garisto December 3, 2020
Researchers at the University of Science and Technology of China (USTC) have for the first time coaxed a quantum computer composed of photons to outperform the fastest classical supercomputers. The Jiuzhan system executed Gaussian boson sampling, detecting 76 photons versus classical supercomputers' previous record of five. Jiuzhan combines lasers, mirrors, prisms, and photon detectors, and its achievement is only the second demonstration of quantum primacy to date. UTSC's Chao-Yang Lu said independent corroboration that quantum computing principles can enable primacy on totally different hardware "gives us confidence that in the long term ... useful quantum simulators and a fault-tolerant quantum computer will become feasible."
|
Phishing Ploy Targets Covid-19 Vaccine Distribution Effort
Associated Press Frank Bajak; Lori Hinnant December 3, 2020
A cyberespionage campaign used phishing emails to try to capture data on the World Health Organization's efforts for distributing Covid-19 vaccine to developing nations, according to IBM security researchers. IBM said the campaign’s targets are likely connected to the development of the "cold chain" for ensuring vaccines get the sterile refrigeration they need to be effective in areas where temperature-controlled storage is insufficient. IBM's Nick Rossmann said executives with groups likely associated with the Covax initiative received spoofed emails seemingly from an executive of China's Haier Biomedical, with malicious attachments designed to coax recipients to provide credentials that could have been used to harvest sensitive information about partners in the vaccine-delivery platform. IBM suggested the campaign's organizers, whose strategy researchers said bore "the potential hallmarks of nation-state tradecraft," are pursuing "insight into the purchase and movement of a vaccine that can impact life and the global economy."
|
'Time' Names Its Kid of the Year: Water-Testing Scientist Gitanjali Rao
National Public Radio Bill Chappell December 3, 2020
Time magazine has singled out Colorado teenager Gitanjali Rao as 2020's Kid of the Year for inventing a mobile device to test for lead in drinking water. Rao, 15, garnered praise in 2017 by creating Tethys, a device that detects lead in water using carbon nanotube sensors, and collaborating with water-industry scientists to put it on the market. She more recently developed a tool called Kindly, which uses artificial intelligence to identify possible early signs of cyberbullying. Rao was named Kid of the Year partly because she has followed up her technical accomplishments by inspiring other young people to tackle real-world challenges. She said she did this because "I really want to put out that message: if I can do it, you can do it, and anyone can do it."
|
Voice Assistant Recordings Could Reveal What Someone Nearby Is Typing
New Scientist Layal Liverpool December 4, 2020
A machine learning system from researchers at the U.K.'s University of Cambridge can be used with voice assistants to detect typing on touchscreens nearby, using artificial intelligence (AI) tools to determine what is being typed. Cambridge's Ilia Shumailov and colleagues had volunteers type random five-digit personal identification numbers or English words on a touchscreen device, while recording audio with a microphone on a separate device nearby. The AI's accuracy at first guess was 28% to 47% when the typist was 20 centimeters away from the recorder, but it rose to between 60% and 76% by the third guess. Accuracy declined as the person typing moved further away from the recording device. Imperial College London's Hamed Haddadi said, "The implications reconfirm that having always-on cameras and microphones in our home will eventually come with privacy and security risks."
|
Many Popular Parental Control Solutions Are Insecure
Concordia University (Canada) Patrick Lejtenyi December 2, 2020
Many commercially-sold parental control solutions are not secure, according to researchers at Canada's Concordia Institute for Information Systems Engineering (CIISE). The team developed experimental frameworks to assess security and privacy in dozens of such solutions, including routers, popular Windows applications, Chrome extensions, and Android applications. Most were insecure in terms of protecting personal information, authentication, and barring third parties and known trackers. Particularly concerning flaws included the Blocksi router's susceptibility to uploaded malicious firmware; Android's FamiSafe, KidsPlace, and Life360 apps not encrypting personal data on shared external storage; and Windows apps Qustodio and Dr. Web using proxy servers that do not properly validate certificates and accept revoked certifications. Although these vulnerabilities typically stem from design flaws, CIISE's Mohammad Mannan thinks the privacy violations are deliberate; said Mannan, “The developers are actually sending private, personally sensitive information to third-party vendors and trackers. Their only job is to collect information and monetize it.”
|
Robot Hands One Step Closer to Human, Thanks to WMG AI Algorithms
University of Warwick (U.K.) December 3, 2020
Artificial intelligence (AI) algorithms developed by researchers at the Warwick Manufacturing Group (WMG) academic department of the U.K.'s University of Warwick enable the Shadow Robot Dexterous Hand to manipulate objects like humans do. The Hand reproduces all of the degrees of freedom of a human hand. The algorithms permit the Hand to learn how to coordinate movements and execute tasks such as throwing a ball and spinning a pen; the researchers said the algorithms can learn any task, as long as it can be simulated. Said Warwick's Giovanni Montana, "The future of digitalization relies on AI algorithms that can learn autonomously, and to be able to develop algorithms that give Shadow Robot’s hand the ability to operate like a real one is without any human input is an exciting step forward.
|
Johns Hopkins Team Develops Software That Cuts Time, Cost From Gene Sequencing
Johns Hopkins University December 3, 2020
New open source software from Johns Hopkins University (JHU) researchers could transform gene sequencing by significantly reducing time and cost. Utility for Nanopore Current Alignment to Large Expanses of DNA (UNCALLED) targets, collects, and sequences specific genes without sample preparation or the need to map surrounding genetic material. By shortening the gene mutation profiling process from roughly 15 days to just three, UNCALLED enables scientists to understand and diagnose conditions almost instantly. The software also functions on standard hardware used for nanopore sequencing, without requiring special reagents or accelerators. JHU's Sam Kovaka said, "UNCALLED allows for unprecedented flexibility in targeted sequencing. The fact that it's purely software-based means researchers can target any genomic region with no added cost compared to a normal sequencing run, and they can easily change targets just by running a different command."
|
Army Computer Models Unveil Secret to Quieter Small Drones
U.S. Army Research Laboratory December 3, 2020
Researchers with the U.S. Army Combat Capabilities Development Command's Army Research Laboratory (ARL) demonstrated how aviation specialists can acquire information about airfoil boundary layers, using computational fluid dynamics to facilitate development of quieter unmanned aerial drones. The University of Maryland College Park's Miranda Costenoble said the researchers are using semi-empirical computer models developed more than three decades ago for a specific airfoil to account for broadband noise generated by the drone's rotor blades passing through air; she acknowledged those models “may need to be updated to account for the physics of different airfoil shapes.” The project is part of a research program at ARL to address unmanned aerial system platform design and control challenges.
|
Driverless Cars Are Coming, But Not Yet to Take Over
The Wall Street Journal Stephen Wilmot December 2, 2020
Driverless cars are making their way onto U.S. roads, but it will take time for them to reach their full potential. Alphabet's Waymo began offering "robotaxis" with no backup driver in Phoenix suburbs in October, and by year’s end, General Motors' Cruise will no longer require backup drivers in its autonomous test cars in California. However, after safely operating driverless-cars in one urban district in good weather, companies must adapt to new areas and other driving conditions. It will take years, and billions of dollars in capital, to enable self-driving vehicles to adapt to multiple geographies and weather conditions, prompting partnerships between tech giants and the auto industry. For instance, Amazon has taken a stake in Aurora Innovation, a startup looking to take over Uber's operation, and acquired Aurora's rival Zoox for $1.3 billion this summer. Waymo's Larry Burns predicts that freight, rather than taxis, will fuel the commercialization of driverless technology.
*May Require Paid Registration
|
Shrinking Massive Neural Networks Used to Model Language
MIT News Daniel Ackerman December 1, 2020
Researchers at the Massachusetts Institute of Technology (MIT), the University of Texas at Austin, and the MIT-IBM Watson Artificial Intelligence Laboratory identified lean subnetworks within a state-of-the-art neural network approach to natural language processing (NLP). These subnetworks, found in the Bidirectional Encoder Representations from Transformers (BERT) network, could potentially enable more users to develop NLP tools using less bulky and more efficient systems, like smartphones. BERT is trained by repeatedly attempting to fill in words omitted from a passage of writing, using a massive dataset; users can then refine its neural network to a specific task. By iteratively trimming parameters from the BERT model, then comparing the new subnetwork's performance to that of the original model, the team found effective subnetworks that were 40% to 90% leaner, and required no task-specific fine-tuning to identify "winning ticket" subnetworks that executed tasks successfully.
|
Open Source Software Security Vulnerabilities Exist for Over Four Years Before Detection
ZDNet Charlie Osborne December 2, 2020
GitHub's annual State of the Octoverse report found security vulnerabilities in open source software can take more than four years to be detected, on average. The platform surveyed more than 56 million developers on GitHub this year, where over 60 million new repositories were generated and more than 1.9 billion contributions added. Compared to 2019, 94% of projects on the platform currently rely on open source components, with nearly 700 dependencies on average; open source dependencies occur most frequently in JavaScript (94%), Ruby (90%) ,and .NET (90%). Most bugs in open source code are non-malicious, with 83% of GitHub-issued common vulnerabilities and exposures alerts rooted in mistakes and human error. GitHub recommended project developers, maintainers, and users regularly check their dependencies for flaws, and consider implementing automated alerts to fix vulnerabilities faster and more efficiently.
|
Researchers Claim AI Algorithm Uses Heart Rate, Motion Data to Predict Age, Sex, More
VentureBeat Kyle Wiggers November 11, 2020
Researchers at the U.K.'s University of Cambridge and Alan Turing Institute claim to have trained an artificial intelligence algorithm on information from wrist accelerometers and wearable electrocardiograms to capture personalized data, with more than 70% AUC. Their self-supervised Step2Heart system generates person-specific profiles, and can predict health-related outcomes with classifiers via transfer learning. The team culled the dataset from 2,100 men and women participating in a study on the interaction between environmental and genetic factors in determining obesity, type 2 diabetes, and related metabolic diseases. The researchers used Step2Heart to forecast blood oxygen consumption rate, height, weight, sex, age, body-mass index, resting heart rate, and physical activity energy expenditure. The team said this experiment proves that unlabeled wearable data can be used to learn profiles that generalize in scenarios where collecting ground truth is infeasible.
|
|