Welcome to the July 12, 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."
|
|
|
Tool Automatically Finds Buffer Overflow Vulnerabilities
Carnegie Mellon University CyLab Security and Privacy Institute
Daniel Tkacik
July 9, 2021
A new tool designed to automatically test for memory flaws in the Rust programming language libraries could detect and mitigate the threat of buffer overflow attacks. Crafted by researchers at Carnegie Mellon University's Security and Privacy Institute (CyLab), the SyRust tool can automatically generate unit tests for library application programming interfaces, and check these library deployments for memory bugs. CyLab's Limin Jia said the team used SyRust on 30 popular libraries, unearthing four previously undiscovered vulnerabilities. Jia said the team is attempting to enhance what it calls the "improved courage" of testing to ensure a wider net has been cast and to improve users' confidence that most, if not all, bugs have been identified.
|
AI's Role in Debugging Code Expected to Grow
The Wall Street Journal
Sara Castellanos
July 6, 2021
Technology companies are developing artificial intelligence (AI)-based tools to debug code as software maintenance becomes ever more challenging. Intel Labs' Justin Gottschlich said developers find it increasingly difficult to identify bugs in code without machine assistance; debugging consumes about half of developers' time, and correcting a single bug can take weeks. Gottschlich said Intel Labs expects to issue two free AI-based software debugging tools for outside developers by year's end. The ControlFlag tool can automatically detect coding errors via statistical analysis and machine learning, and the Machine Inferred Code Similarity tool can automatically recognize code snippets that execute similar functions.
*May Require Paid Registration
|
MIT Robot Could Help People with Limited Mobility Dress Themselves
Engadget
Steve Dent
July 12, 2021
Researchers at the Massachusetts Institute of Technology have designed an algorithm to help a robot efficiently dress a human, theoretically ensuring human safety by reasoning about the human model's uncertainty. The team declined to use a single default model in which the machine only understands one potential reaction in favor of many possible models, to more closely emulate how a human understands other humans. The robot reduces uncertainty and refines those models by collecting more data. The MIT team also reclassified safety for human-aware motion planners as either collision avoidance or safe impact in case of a collision, so the robot could safely complete the dressing task faster. Carnegie Mellon University's Zackory Erickson said, "This research could potentially be applied to a wide variety of assistive robotics scenarios, towards the ultimate goal of enabling robots to provide safer physical assistance to people with disabilities."
|
UOC Team Develops Neural Network to Identify Tiger Mosquitoes
Universitat Oberta de Catalunya (Spain)
Santiago Campillo
July 8, 2021
A new technique can learn to identify tiger mosquitoes using a large set of images captured on mobile phones and uploaded to the Mosquito Alert platform by volunteers. Scientists at Spain's Universitat Oberta de Catalunya (UOC) engineered a deep neural network to differentiate between the morphological similarities of diverse mosquito species, and UOC's Gereziher Adhane said the network performs as well or nearly as well as human experts, and can process vast volumes of images. Adhane also said the algorithm, with modest tweaking, could work with mobile applications. Adhane said the neural network can “make predictions about images of mosquitoes taken using smartphones efficiently and in real time, as has happened with the Mosquito Alert project."
|
AR System Alters Sight, Sound, Touch
IEEE Spectrum
Michelle Hampson
July 2, 2021
Researchers at the University of California, Santa Barbara have developed an augmented reality (AR) system that can translate the vibrations produced by fingers touching a surface into programmed sounds. The system, Tactile Echoes, could be used for gaming, entertainment, and research purposes. A sensor placed on a user's fingernail can detect the vibrations produced within the finger as it touches a surface, creating different sounds depending on the surface touched or the pressure applied. Researcher Anzu Kawazoe said, "We were motivated by the idea of being able to almost magically augment any ordinary object or surface, such as a simple wooden table, with lively haptics and effects that playfully respond to how or where we touch." The device can be integrated with smart projectors or virtual reality or AR headsets to allow users to touch virtual objects in their real environment and receive graphic, sound, and haptic feedback.
|
Imaging Technique May Boost Biology, Neuroscience Research
The Harvard Gazette
Juan Siliezar
July 7, 2021
Researchers at Harvard University and the Massachusetts Institute of Technology have developed a computational imaging process that could improve biology and neuroscience research. The new system, De-scattering with Excitation Patterning (DEEP), uses computational imaging to generate high-resolution images 100 to 1,000 times faster than point-scanning multiphoton microscopy or temporal focusing microscopy. DEEP uses near-infrared laser light to penetrate deep into biological tissue, which scatters the light and excites the fluorescent molecules to be imaged, which emit signals to be captured by the microscope. Harvard's Dushan N. Wadduwage said, "This is very important for neuroscientists and other biologists to actually get better statistics, as well as to see what's happening around the area being imaged."
|
FAA: Tool Limits Disruptions Caused by Space Operations
Associated Press
July 8, 2021
The U.S. Federal Aviation Administration (FAA) last week announced it was using new technology that can automatically deliver data about a space vehicle's trajectory to the U.S. air traffic control system almost instantly. The Space Data Integrator tool will largely replace the manual task of sending such information, which could shorten the amount of time required to route airplanes around space operations. The regulator said the technology was first used for last month's launch of SpaceX's Transporter 2 satellite deployment vehicle; it also will be employed for the return of a SpaceX cargo ship from the International Space Station. The FAA's Stephen Dickson said, "With this capability, we will be able to safely reopen the airspace more quickly and reduce the number of aircraft and other airspace users affected by a launch or reentry."
|
Facing Skilled Worker Shortage, U.S. Technology Companies Try to Train Their Own Labor Pools
PBS NewsHour
Levi Pulkkinen
July 1, 2021
Technology companies are working to address a shortage of qualified talent. IBM, for instance, is offering a two-year program for entry-level workers without college degrees, providing classroom instruction and on-the-job training for in-demand positions it has difficulty filling. CompTIA's Amy Kardel said, "Earn-and-learn strategies can open a door for someone into a career quickly." The IT trade association is working to promote on-the-job learning with the U.S. Department of Labor. Kardel said employers who depend on technology, including manufacturers, retailers, and the government, are competing with Silicon Valley for the same workers. IBM, which is looking to hire over 400 trainees annually, offers 25 different training tracks, including software development, data science, and human resources. The company retains about 90% of participants. IBM's Kelli Jordan said the company is offering "new-collar jobs," white-collar work that does not require a four-year degree but offers necessary learning on the job.
|
Google Releases Open Source Security Software Program: Scorecards
ZDNet
Steven J. Vaughan-Nichols
July 1, 2021
Google and the Open Source Security Foundation have developed the OpenSSF Security Scorecards, an automated security tool that generates a "risk score" for open source programs. This is important because 95% of all commercial programs contain open source software, according to the Synopsys Cybersecurity Research Center, and many organizations lack the systems and processes to evaluate new open source dependencies for security issues. Scorecards v2 includes new security checks, including the Branch-Protection check, which ensures code reviews to prevent malicious contributors from introducing potential backdoors into code. The Scorecards project already has performed security evaluations for more than 50,000 open source projects.
|
'We Don't Need Another Michelangelo': In Italy, It's Robots' Turn to Sculpt
The New York Times
Emma Bubola
July 11, 2021
Scientists at the Robotor laboratory in Carrara, Italy, are developing sculpting robots to keep the country on the artistic forefront. Robotor's Giacomo Massari said the continued prosperity of the Italian marble sculpture segment depends on discarding traditional manual techniques, especially since marble has fallen out of favor in artistic circles. Robotor's Michele Basaldella said many outstanding sculptors lack distinction because manual dexterity is frowned upon, but robots can create groundbreaking works if they are built "with an artistic sensitivity." The company's founders initially used robots from local technology companies, but started designing their own from homemade software and German parts when artist clients started ordering increasingly challenging commissions.
*May Require Paid Registration
|
NASA Preps 'More Complex and Riskier' Hubble Space Telescope Fix
CNet
Amanda Kooser
July 1, 2021
The U.S. National Aeronautics and Space Administration (NASA) hopes to correct a persistent issue with the Hubble Space Telescope's payload computer, in which commands to write into or read from memory are not going through. The agency is preparing to activate backup hardware that is part of the Science Instrument Command and Data Handling unit where the payload computer resides. The NASA team is considering a power regulator element, and hardware that transmits and formats commands and data. An agency update said if one of these components is the likely culprit, it will require a more complicated and riskier backup unit-switching procedure than previously attempted. The switchover will be conducted in a simulation prior to the actual attempt, and the process highlights the reality of working with aging systems that have long exceeded operational expectations.
|
ML Helps Predict When Immunotherapy Will Be Effective
Eindhoven University of Technology (Netherlands)
June 30, 2021
Researchers at the Eindhoven University of Technology (TU/e) in the Netherlands have developed a machine learning model that can predict whether immunotherapy will work for a patient. One type of immunotherapy that involves immune checkpoint blockers (ICB) is effective in only a third of patients. The researchers used computational algorithms and datasets from previous clinical patient care to search the tumor microenvironment for biomarkers to predict patient response to ICB. TU/e's Federica Eduati said, "RNA-sequencing datasets are publicly available, but the information about which patients responded to ICB therapy is only available for a small subset of patients and cancer types." To solve the data problem, the researchers searched for substitute immune responses from the same datasets, which could be an indicator of ICB's effectiveness. Eduati said, "Our machine learning model outperforms biomarkers currently used in clinical settings to assess ICB treatments."
|
Meet the Scientist Teaching AI to Police Human Speech
The Washington Post
Drew Harwell
July 1, 2021
Facebook and Google have engineered artificial intelligence (AI) systems capable of understanding dozens of languages with remarkable accuracy through the efforts of scientists like Alexis Conneau. At Facebook, Conneau and others advanced machine learning algorithms' ability to abstract language numerically, eventually training an AI model to piece through different languages concurrently; the 100-language XLM-R model was almost as accurate as its specialized single-language peers. Conneau's final work for Facebook was on wav2vec-U, an unsupervised speech-recognition system that reads words from audio. Conneau has helped lead research on natural language processing, and spearheaded work in AI that Facebook and others have applied to the online policing of bullying, bigotry, and hate speech. He believes this problem can be addressed only through automation, while critics claim such innovations will just give companies more information on Web users to exploit.
*May Require Paid Registration
|
|
