Welcome to the January 5, 2022 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."
|
|
|
Drone Innovation Moves from Hardware to Self-Flying Software
Financial Times
Patrick McGee
December 30, 2021
Researchers at Shield AI, a defense startup based in San Diego, CA, have developed software that enables autonomous drones. Shield AI's Brandon Tseng said the Hivemind software allows drones to "go inside a building without GPS, without communications—these are jammed by adversaries—so it can scope out a building ahead of soldiers." Other companies also are working to transition drones from human-piloted point-and-shoot flying cameras to self-flying computers able to work in teams and upload data to the cloud. Qualcomm's ModalAI unit, which supplies technology to drone makers, said around 65 U.S. startups are working to outfit drones with the latest software and artificial intelligence.
*May Require Paid Registration
|
Singapore Engineers Build Robotic Fingers with Delicate Grip
ZDNet
Eileen Yu
December 28, 2021
National University of Singapore (NUS) engineers have constructed a robotic hand that can grasp various objects using three-dimensionally-printed fingers. The air-driven fingers feature a locking mechanism for adjusting the level of stiffness required for grasping, while computer vision and deep learning technologies help identify object types and determine their orientation. The system chooses the best way to pick and place objects to reduce the need for human assistance. The NUS researchers said the gripper system may be reconfigured on demand, and can be equipped with three different grip options that permit its use in a variety of applications.
|
Detecting Evasive Malware on IoT Devices Using Electromagnetic Emanations
The Hacker News
Ravie Lakshmanan
January 3, 2022
Researchers at France’s Research Institute of Computer Science and Random Systems (IRISA) have proposed harnessing electromagnetic field emanations from Internet of Things (IoT) devices to collect side-channel data about malware, even when evasion is implemented. The idea is to exploit this information to detect anomalies in emanations that deviate from previously observed patterns, and flag suspicious behavior mimicking the malware, versus the system's normal state. The framework requires no hardware modification, and can detect and classify malware such as kernel-level rootkits, ransomware, and distributed denial-of-service botnets. The IRISA researchers said the method is robust against code transformation/obfuscation schemes including random junk insertion, packing, and virtualization, even when the transformation is previously unknown.
|
‘Sci-Fi’ Medical Implants Soon to Become Reality, Researchers Say
The Washington Post
Kenneth R. Rosen
January 1, 2022
Researchers are working to merge embedded medical devices like pacemakers, cochlear implants, and cardiac defibrillators with wearable devices that track heart rates and other vital signs. These technologies could allow patients to be screened before a doctor appointment without having to submit to lab testing, or allow them to carry their medical records with them at all times. The University of Minnesota's Hubert H. Lim said fewer volunteers are willing to participate in clinical studies of these technologies, because “People don't want things in their body.” However, Lim added, “What I've learned with deep brain stimulation, even hip replacement, is that it's something that is scary at first but, generally speaking, people do in my opinion adapt to it as they see the benefits of it."
|
Long-Reads, Powerful Algorithms Identify 'Invisible' Microbes
UC San Diego News Center
Josh Baxt
January 3, 2022
An international team of scientists has demonstrated that highly accurate, long-read genomic sequencing technology (HiFi), applied with advanced algorithms, can distinguish between nearly identical, previously "invisible" microbes. The researchers used HiFi long-reads to sequence the microbial metagenome in sheep guts; the process and associated algorithms identified genomes from 428 species with more than 90% completeness, many of which were previously overlooked by short-read technologies. The National Cancer Institute's Mikhail Kolmogorov said, "We can now sequence complete genomes of nearly all abundant bacteria in a microbiome. Short-read studies rarely provided complete sequences of even a single microbe."
|
Are Apple AirTags Being Used to Track People, Steal Cars?
The New York Times
Ryan Mac; Kashmir Hill
January 1, 2022
Concerns are mounting as Apple's location-tracking AirTag devices are being found on people's cars and in their belongings, giving rise to fear of stalking. Seven women interviewed by The New York Times think they were tracked with AirTags, including a 17-year-old whose mother put one on her car. Canadian police also reportedly have investigated cases of thieves placing AirTags on "high-end vehicles so they can later locate and steal them." Bluetooth-outfitted AirTags emit a signal that can be detected by devices using Apple's mobile operating system. While the devices incorporate abuse-preventing features include tracking alerts and automatic beeping, the Electronic Frontier Foundation's Eva Galperin warns they constitute a "uniquely harmful" threat, given the ubiquity of Apple products.
*May Require Paid Registration
|
Robots Collect Underwater Litter
Technical University of Munich (Germany)
December 29, 2021
Researchers at Germany's Technical University of Munich (TUM) are developing a multi-robot system that uses machine learning to find and collect garbage underwater, as part of the European SeaClear Project. An autonomous boat scans the seafloor to identify pockets of litter, then an observation robot detects undersea litter and sends close-up images and other data to computers. An aerial drone is used to identify more litter, which is used with previously amassed data to form a map that then is used by a fourth robot to visit points on the map and collect litter. The team's artificial intelligence module predicts robot movement and behavior; TUM's Sandra Hirche said the system uses high-efficiency sampling to make precise predictions with limited computing power.
|
Model Moves Computers Closer to Understanding Human Conversation
Johns Hopkins Hub
Wick Eisenberg
December 20, 2021
Piotr Zelasko at the Johns Hopkins Center for Language and Speech Processing has developed a machine learning model that can differentiate speech functions in transcripts of conversations generated by language understanding (LU) systems. The model performs dialogue act recognition by identifying words' underlying intent, and assigns them to categories like "Statement," "Question," or "Interruption" in the final transcript. Zelasko sought to ensure his system could understand ordinary conversation, which may help with such tasks as summarization, intent recognition, and detection of key phrases. Zelasko said LU systems no longer need contend with "huge, unstructured chunks of text, which they struggle with when trying to classify things such as the topic, sentiment, or intent of the text. Instead, they can work with a series of expressions, which are saying very specific things."
|
Autonomous Trucker TuSimple Logs First No-Human Road Test
Associated Press
December 29, 2021
Autonomous technology developer TuSimple said its self-driving unmanned semitruck completed an 80-mile trip in Arizona on public roads without human intervention. The nighttime trip started at a Tucson, AZ, railyard and ended at a distribution center in Phoenix, with approval and oversight from Arizona's Department of Transportation and law enforcement. A lead vehicle scouted the route for obstacles roughly five miles ahead of the truck, while another vehicle trailed the semi by about a half-mile in order to intervene if necessary. TuSimple said the semi managed highway lane changes, traffic signals, on-ramps, and off-ramps while "naturally interacting with other motorists."
|
Standards Rolling Out for Clocking Quantum Computer Performance
IEEE Spectrum
Charles Q. Choi
January 3, 2022
Researchers at Sandia National Laboratories have developed a novel method for creating benchmarks for quantum computing. The method converts any quantum program into an array of closely related benchmark programs called "mirror circuits," each of which execute, then reverse, a set of calculations. Mirror circuits' outcomes are simple and easy to predict, and offer a way to confirm a quantum processor's capabilities using benchmarks similar to programs the quantum computers may actually run. "Ultimately, we think that benchmarks built using our method could be used to design a set of comprehensive tests for quantum computers," said Sandia's Tim Proctor. "This would make it possible to accurately compare quantum computers, and to find out which hardware is best for which task."
|
Digital Twins Transforming Manufacturing, Medicine, More
Time
Andrew R. Chow
December 30, 2021
Multiple industries, including manufacturing and medicine, have adopted digital twins—virtual replications of objects or environments—generated by feeding video and other data to three-dimensional mapping software. Last year, the technology started penetrating mainstream manufacturing and research, with Nvidia and Amazon Web Services, for example, unveiling digital twin-building services for customers. Digital twins could transform worker training and the production of technical plans without wasting physical resources, but privacy concerns abound because many such replicas are facilitated by sensors that track real-world data and movement. Carnegie Mellon University's Ding Zhao said, "Legislators and companies need to work together on [digital twin regulations]. You cannot just say, 'I am a good company, I will never do evil things, just give me your data.'"
|
More Malicious Domains Are Online Than Ever Before
TechRadar
Anthony Spadafora
December 30, 2021
The Unit 42 group of cybersecurity firm Palo Alto Networks in September used a cloud-based detector to find malicious domains strategically registered years before they were actually employed. The researchers estimated 22.3% of strategically aged domains present a threat, with 3.8% being straight-out malicious, 19% suspicious, and 2% unsafe for work environments. Malefactors keep domains dormant to establish a "clean record" so their domains are less likely to be blocked when activated. Security systems frequently flag newly registered domains as malicious, yet the Palo Alto researchers said strategically aged domains are three times more likely to be malicious; a sudden spike in traffic usually signals a domain's maliciousness, while normal Websites experience more gradual traffic growth.
|
AI-Powered Robots Replace Recycling Center Workers in Finland
EuroNews
December 28, 2021
A new recycling facility near Helsinki, Finland, can process as much as 120,000 metric tons of construction waste per year with the help of robots. The facility employs 12 robotic arms from ZenRobotics to sort wood, plastic, metal, and stone waste. The robots, which can lift up to 30 kilos (66 lbs.), could better identify valuable materials in the waste stream. Said ZenRobotics' Harri Holopainen, "The key thing for these robots is that they actually identify the waste objects on the belt. They look at each and every object that comes on the belt and then figure out whether it's wood, whether it's wood that has some nails in it, tiles, or concrete, and then they put it in the correct chute for later processing." The robots are equipped with cameras, a three-dimensional sensor system, a metal detector, and an artificial intelligence-powered brain that recognizes over 350 different types of waste and determines the best gripping point for each item.
|
|
