Welcome to the December 13, 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."
|
|
|
Software Flaw Sparks Global Race to Patch Bug
The Wall Street Journal
Robert McMillan
December 13, 2021
Companies and governments scrambled this past weekend to patch a major bug in a piece of popular Internet software that security experts warned could grant hackers access to networks. Cybersecurity researchers said the bug, hidden in Log4j server code, is one of the most significant vulnerabilities in recent years due to its use on corporate networks; hackers began exploiting it on Friday, and Check Point Software Technologies observed more than 100,000 attempts over roughly 24 hours. Apache Software Foundation’s Ralph Goers said users must upgrade to correct the bug. The U.S. Department of Homeland Security's Cybersecurity and Infrastructure Security Agency urged corporate action, while German and Australian agencies issued similar warnings.
*May Require Paid Registration
|
Research Brings Analog Computers One Step from Digital
Washington University in St. Louis
Brandie Jefferson
December 8, 2021
A new processing-in-memory (PIM) circuit developed by researchers at Washington University in St. Louis (WUSTL) adds neural network flexibility to PIM computing. With PIM computing, the memory and processing unit are merged into a single unit, and computations are performed using the machine's physical properties. The researchers are working on resistive random-access memory PIM, with resistors in both the memory and the processor. Said WUSTL's Xuan "Silvia" Zhang, "In resistive memory, you do not have to translate to digital, or binary. You can remain in the analog domain." To overcome the bottleneck that occurs when converting analog information into a digital format, the researchers added neural approximators that perform multiple calculations in the most efficient manner.
|
Technology Helps Keep Women Safe on the Streets
BBC News
Shiona McCallum
December 12, 2021
The WalkSafe anti-harassment application's popularity has soared, making it the fastest-growing safety app in Britain. WalkSafe displays a map that allows users to plan their safest route home by reviewing recent crime data related to the route. WalkSafe founder Emma Kaye said, "Our app encourages people to have precautionary measures to avoid crime, rather than reactionary measures when in an unfortunate situation." Meanwhile, Rich Larsen thinks his bSafe smartphone app could aid prosecutions, as its voice-activated alarm livestreams video and audio recordings to chosen contacts, which he said “could be used in evidence in court cases—like rape—which are often hard to prove.”
|
A System for Designing, Training Intelligent Soft Robots
MIT News
Rachel Gordon
December 7, 2021
Scientists at the Massachusetts Institute of Technology (MIT) Computer Science and Artificial Intelligence Laboratory designed the Evolution Gym, a large-scale testing platform for co-optimizing the design and control of soft robots. The researchers integrated standard design optimization and deep reinforcement learning algorithms into the platform to assess robotic aptitude; the former evolves robotic bodies while the latter optimizes a controller for each proposed design. Said MIT’s Jagdeep Bhatia, “By creating a large-scale benchmark that focuses on speed and simplicity, we not only create a common language for exchanging ideas and results within the reinforcement learning and co-design space, but also enable researchers without state-of-the-art computing resources to contribute to algorithmic development in these areas.”
|
Intel Shows Research for Packing More Computing Power Into Chips Beyond 2025
Reuters
Stephen Nellis
December 11, 2021
Semiconductor manufacturer Intel disclosed research it expects will help it continue to accelerate and miniaturize computing chips over the next decade. Possibly most significant was a paper demonstrating a method for stacking transistors, which the company thinks will enable a 30% to 50% boost in transistor density. "We're reducing interconnect lengths and really saving energy, making this not only more cost efficient, but also better performing," said Intel's Paul Fischer. Intel also highlighted three-dimensional stacking of "tiles" or "chiplets" to increase computing power, which could facilitate 10 times as many connections between stacked tiles.
|
Crucial Leap in Error Mitigation for Quantum Computers
Lawrence Berkeley National Laboratory Computing Sciences Research
Monica Hernandez; William Schulz
December 9, 2021
Researchers at the U.S. Department of Energy's Lawrence Berkeley National Laboratory (LBNL) have dramatically reduced coherent errors in quantum algorithms through the use of randomized compiling (RC), which mitigates coherent errors by randomizing the direction in which they affect quantum bits (qubits) so they act like stochastic noise. The researchers used a four-qubit superconducting quantum processor for its experiments, as well as working with the RC protocol's authors. An experimental demonstration indicated that RC can control coherent errors in quantum algorithms to grow more slowly. "Our collaborative research demonstrated that RC works to improve algorithms in the NISQ [noisy intermediate-scale quantum] era, and we expect it will continue to be a useful protocol beyond NISQ," said LBNL's Akel Hashim.
|
DeepMind Massive Language AI Approaches Human-Level Reading Comprehension
Fortune
Jeremy Kahn
December 8, 2021
Artificial intelligence (AI) research company DeepMind unveiled the Gopher language software, which the company says approaches human-level performance in reading comprehension. Gopher performs many tasks with greater accuracy than other ultra-large language models, like answering questions about specialized subjects; for tasks like logical reasoning and mathematics, it performs as well or nearly as well as other models. Gopher is smaller than some ultra-large language models, featuring about 280 billion different parameters, compared to 535 billion in Microsoft and Nvidia's Megatron model and 1.6 trillion in Google's model. DeepMind's Koray Kavukcuoglo said the company has no plans to commercialize the software at present.
|
NIST Records Fifth Straight Year of Record Software Vulnerabilities
ZDNet
Jonathan Greig
December 8, 2021
The U.S. National Institute of Standards and Technology logged record software vulnerabilities for the fifth consecutive year in 2021, recording 18,378. The number of high-severity vulnerabilities declined from 4,381 last year to 3,646 this year, while medium- and low-risk vulnerabilities topped those observed in 2020. Bugcrowd's Casey Ellis said with fundamental technology development accelerating, vulnerabilities will grow in number as more software is created. Said Ellis, "High-impact issues tend to be more complicated, remediated more quickly once found, and—in the case of systemic high-impact vulnerability classes—are often prioritized for root-cause analysis and anti-pattern avoidance in the future, and thus can often be fewer in number." K2 Cyber Security's Pravin Madhani said the pandemic contributed to the increase in reported flaws, prompting many organizations to rush products to market with less-rigorous quality assurance.
|
Your Face Is, or Will Be, Your Boarding Pass
The New York Times
Elaine Glusac
December 7, 2021
Many biometric identification innovations use facial recognition, which the U.S. National Institute of Standards and Technology estimated is at least 99.5% accurate compared to iris scanning or fingerprints. Delta Air Lines in November launched a digital identity program for Transportation Security Administration (TSA) PreCheck members at Hartsfield-Jackson Atlanta International Airport, who can opt in to using facial recognition. Swiss biometrics technology company SITA also tested a system with United Airlines at San Francisco International Airport to compare driver licenses and passports to facial scans for baggage check and domestic boarding purposes. The CLEAR subscription service, meanwhile, allows subscribers to use dedicated kiosks to access their biometric data, confirm their identities, and advance to the front of the TSA security line.
*May Require Paid Registration
|
3D Imaging Method May Help Doctors Better Determine Prostate Cancer Aggressiveness
University of Washington News
Sarah McQuate
December 9, 2021
Researchers at the University of Washington (UW), Case Western Reserve University, the Allen Institute, and the non-profit Canary Foundation developed a three-dimensional (3D) biopsy imaging method for predicting aggressive prostate cancer. UW's Jonathan Liu said, "We show for the first time that compared to traditional pathology—where a small fraction of each biopsy is examined in 2D [two dimensions] on microscope slides—the ability to examine 100% of a biopsy in 3D is more informative and accurate." The team imaged surgically extracted biopsy cores without destroying them, using an open-top light-sheet microscope. The data gleaned from this method boosted the probability of the computer correctly predicting a cancer’s aggressiveness.
|
Robots Gathering to Help Beat Britain's Supply-Chain Shortages
The Economist
December 9, 2021
Robotic logistics is helping to mitigate labor shortages and other supply-chain disruptions in Britain. Research group Interact Analysis says the U.K.'s distribution and fulfillment centers are the world's most automated; Amazon's U.K. warehouses, for example, deploy squat robots to slide under shelves and ferry items to people for sorting and packing. Online grocer Ocado operates a robot fleet at its "Hive" fulfillment center in Erith in southeast London, and is building 50 more Hive-like systems worldwide. A robotic system from Estonia’s Starship Technologies already is providing grocery deliveries from Co-op stores in the towns of Northampton and Milton Keynes using autonomous delivery pods that travel footpaths and cycleways, equipped with sensors to avoid people and other vehicles.
*May Require Paid Registration
|
Research Enables Key Step Toward Personalized Medicine: Modeling Biological Systems
Oregon State University News
Armand Resto-Spotts
December 9, 2021
Oregon State University (OSU) researchers have found that machine learning can be used to model nonlinear biological systems at the cellular level, paving the way for personalized medicine. Although biological systems are resistant to traditional upscaling techniques, machine learning can reduce the information load for complex systems at the cellular level, allowing researchers to model the impact or response of cells with high fidelity without modeling them individually. OSU's Brian D. Wood explained the process as "simplifying a computational problem that has tens of millions of data points by reducing it to thousands of data points." Said Wood, "Our work capitalizes on what are called deep neural networks to upscale the nonlinear processes found in transport and reactions within tissues," which could lead to potential treatments based on numerical model outcomes.
|
|
