Welcome to the March 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."

Illustration of droplets spread by coughs while wearing nonwoven and ill-fitting double masks. Double-Masking Benefits Are Limited, Japan Supercomputer Finds
Gearoid Reidy
March 5, 2021

Double-masking, as recommended by the U.S. Centers for Disease Control and Prevention, yields limited benefits in preventing the spread of droplets that could transmit Covid-19 compared to a single well-fitted disposable mask, according to an analysis conducted with a Japanese supercomputer. Researchers at Japan's Riken research institute and Kobe University used Fugaku, the world's fastest supercomputer, to model droplet dispersal. The simulation demonstrated that wearing just one tightly-fitted disposable mask prevented the spread of 85% of virus-bearing particles, while wearing two masks prevented only 89%. One well-fitted mask captured 81% of the droplets, compared to 69% by one loosely-fitted mask. The researchers observed that a tight fit and avoiding gaps in the mask were essential to blocking droplet spread.

Full Article
Using Machine Learning to Develop Personalized Cancer Vaccines
University of Waterloo Cheriton School of Computer Science (Canada)
March 11, 2021

Researchers at Canada's University of Waterloo Cheriton School of Computer Science are applying machine learning to identify tumor-specific neoantigens, which could lead to personalized cancer vaccines. Cheriton's Hieu Tran said the team used a model similar to natural language processing to ascertain neoantigens' amino acid sequences based on one-letter amino acid codes. The researchers used the DeepNovo recurrent neural network to predict amino acid sequences, which Tran said expanded the predicted immunopeptidomes of five melanoma patients by 5% to 15%, based solely on data from mass spectrometry—and personalized the neoantigens to each patient.

Full Article

Illustration of a secure thumbprint scan. CredChain: Take Control of Your Own Digital Identity ... and Keep That Valuable Bitcoin Password Safe
UNSW Sydney Newsroom (Australia)
Neil Martin
March 12, 2021

The CredChain Self-Sovereign Identity platform architecture developed by researchers at Australia's University of New South Wales (UNSW) School of Computer Science and Engineering uses blockchain to create, share, and verify cryptocurrency credentials securely. UNSW's Helen Paik and Salil Kanhere said CredChain could offer Key Sharding, the process of splitting complicated passwords into meaningless shards stored in different locations that can only be validated when recombined. Kanhere said, "If or when the key is lost, the owner can present enough pieces of the keys to the system to prove his identity and recover the original." Paik said CredChain offers decentralized identity authority via the blockchain, and “also ensures that when a credential is shared, the user can redact parts of the credential to minimize the private data being shared, while maintaining the validity of the credential."

Full Article

A worker at the Tyson Manufacturing and Automation Center in Springdale, AK. Trading Chicken Parts is Going Digital
The Wall Street Journal
Julia-Ambra Verlaine
March 8, 2021

Ashley Honey at New Zealand-based Nui hopes to use his company's electronic trading platforms to automate trading of meat and poultry products. The goal is to remove intermediate brokers or redistributors from the supply chain and lower costs with streaming platforms that centralize supply and provide access to smaller industry players at better prices. Agricultural giant Tyson Foods also aims to upgrade technology to simplify sales transactions, in order to reduce the cost of processing and distributing food across the U.S. To this end, the firm is deploying robotic arms to package poultry, and implementing digital platforms to help sales teams recognize consumption trends.

Full Article
*May Require Paid Registration

A non-rigid point set applied to human body data. Key Task in Computer Vision, Graphics Gets Boost
Kanazawa University (Japan)
March 8, 2021

Osamu Hirose at Japan's Kanazawa University has proposed a method for accelerating a basic computer vision and graphics task called non-rigid point set registration, which is computationally efficient even for large datasets. The three-step technique involves reducing the number of points in each point set via downsampling, then applying non-rigid point set registration to the downsampled point sets using a Bayesian coherent point drift algorithm; the final step is to interpolate shape deformation vectors for the points removed during downsampling, via Gaussian process regression. Hirose found the technique is efficient even for point sets with more than 10 million points, while the time it takes to compute also is significantly faster than that of a previously state-of-the-art approach.

Full Article
Bug Bounties: More Hackers Spotting Vulnerabilities Across Web, Mobile, IoT
Danny Palmer
March 9, 2021

HackerOne's 2021 Hacker Report reveals a 63% jump in the number of hackers submitting vulnerabilities to bug bounty programs during the last year. Earnings for ethical hackers disclosing vulnerabilities to the HackerOne bug bounty program more than doubled to $40 million in 2020, from $19 million in 2019. Most of the hackers focus on Web applications, but submissions of vulnerabilities associated with Android devices, Internet of Things devices, and application programming interfaces also increased last year. Said HackerOne's Jobert Abma, "We're seeing huge growth in vulnerability submissions across all categories and an increase in hackers specializing across a wider variety of technologies."

Full Article
Anti-Feminist YouTube, Reddit Content a Gateway to the Alt-Right
New Scientist
Chris Stokel-Walker
March 9, 2021

Researchers at the Swiss Federal Institute of Technology in Lausanne (EPFL) have determined that Reddit and YouTube users who engage with anti-feminist content can become alt-right converts. EPFL's Manoel Ribeiro and colleagues analyzed 300 million comments on 115 Reddit forums and 526 YouTube channels from 2006 to 2018, tracking the type of subject matter each user engaged with: general news, content from communities that expressed hate towards women (sometimes called the "manosphere"), and alt-right material. They also checked people who in 2016 commented on YouTube videos classified as anti-feminist and on general news videos, without engaging with alt-right videos, compared to 2018. Members of the male-separatist group Men Going Their Own Way were most likely to later engage with alt-right content, while overall the migration from the manosphere to the alt-right was higher on Reddit than YouTube.

Full Article

A grid of six female faces; can you tell which are fake? How to Spot Deepfakes? Look at Light Reflection in the Eyes
UB News Center
Melvin Bankhead III
March 10, 2021

A tool developed by University at Buffalo computer scientists can automatically identify deepfake photos of people by analyzing light reflections in their eyes for minute deviations. The tool exploits the fact that most artificial intelligence (AI)-generated images cannot accurately or consistently reflect the image of what the pictured person is seeing, possibly because many photos are combined to create the fake image. The tool first maps out each face, then analyzes the eyes, the eyeballs, and finally the light reflected in each eyeball. The tool was 94% effective in spotting deepfakes among portrait-like photos taken from actual images in the Flickr Faces-HQ dataset, as well as fake AI-generated faces from the www.thispersondoesnotexist.com repository.

Full Article

A dashboard GIS screen. Classic Math Conundrum Solved: Danish Computer Scientist Developed Algorithm for Finding the Shortest Route
University of Copenhagen (Denmark) Faculty of Science
March 10, 2021

Computer scientists at Denmark's University of Copenhagen (UCPH) have solved the classic algorithmic problem of calculating the shortest path between two points when the route traverses a changing network. UCPH's Christian Wulff-Nilsen and colleagues represent a network as a dynamic graph, and their new algorithm accommodates changes consisting of deleted edges. Wulff-Nilsen said, "The tremendous advantage of seeing a network as an abstract graph is that it can be used to represent any type of network. It could be the Internet, where you want to send data via as short a route as possible, a human brain, or the network of friendship relations on Facebook." Wulff-Nilsen described the algorithm as "better than every other algorithm up to now—and the closest thing to optimal that will ever be, even if we look 1,000 years into the future."

Full Article

A huge wave breaking. Drones Used in Search and Rescue Trials at North Sea Wind Farm
Anmar Frangoul
March 11, 2021

The international Dredging, Environmental, and Marine Engineering (DEME) Offshore conglomerate and Belgian aerospace firm Sabca has completed tests of drones in various scenarios at a wind farm in the North Sea. DEME Offshore said the pilot program involved using drone for turbine inspections, environmental surveys, and parcel deliveries, as well as a search and rescue operation involving an automated drone using infrared detection to locate its target before dropping a life buoy into the sea. DEME Offshore's Bart De Poorter said, "We are convinced that these innovative, advanced technologies, which focus on fully autonomous operations without the need for any vessels and people offshore, have a game-changing potential to increase safety, lower the impact on the environment in the [operations and management] phase of a project, and reduce the overall costs."

Full Article

A large computer monitor packed with colorful of lines of code. Large Computer Language Models Carry Environmental, Social Risks
UW News
Jackson Holtz
March 10, 2021

University of Washington (UW) researchers warn that fast-growing computerized natural-language models can worsen environmental and social issues as the amount of training data increases. UW's Emily M. Bender and colleagues said the enormous energy consumption needed to drive the model language programs' computing muscle induces environmental degradation, with the costs borne by marginalized people. Furthermore, the massive scale of compute power can limit model access to only the most well-resourced enterprises and research groups. Critically, such models can perpetuate hegemonic language because the computers read language from the Web and other sources, and can fool people into thinking they are having an actual conversation with a human rather than a machine. Bender said, "It produces this seemingly coherent text, but it has no communicative intent. It has no idea what it's saying. There's no there there."

Full Article

Aerial robots developed by UT Dallas researchers can carry several cameras. Robots Designed to Avoid Environmental Dangers, Deliver Data Quickly
University of Texas at Dallas
Phil Roth
March 3, 2021

An autonomous team of robotic devices developed by researchers at the University of Texas at Dallas (UT Dallas) can be deployed to perform a general survey of ecosystems or at hazardous or hard-to-reach sites for real-time decision support. The devices collect thousands of data records while on the ground, in the air, or in water within minutes. The autonomous devices include robot boats to measure water composition, sonar to detect objects below the water’s surface, aerial drones with multiple onboard sensors, and a ground vehicle to collect soil samples and deploy ground-penetrating radar. UT Dallas' David Lary said, "An autonomous team like this could do a survey and rapidly sample what’s in the air and the water so that people could be kept out of harm’s way.”

Full Article

Three yellow balls from the videogame Pong, in which an “adversary” appeared to pull the ball slightly further down than it actually was. Algorithm Helps AI Systems Dodge 'Adversarial' Inputs
MIT News
Jennifer Chu
March 8, 2021

Massachusetts Institute of Technology (MIT) researchers have developed a deep learning algorithm designed to help machines navigate real-world environments by incorporating a level of skepticism of received measurements and inputs. The team mated a reinforcement-learning algorithm with a deep neural network, each used separately to train computers in playing games like Go and chess, to support the Certified Adversarial Robustness for Deep Reinforcement Learning (CARRL) approach. CARRL outperformed standard machine learning techniques in tests using simulated collision-avoidance and the videogame Pong, even when confronted with adversarial inputs. MIT's Michael Everett said, "Our approach helps to account for [imperfect sensor measurements] and make a safe decision. In any safety-critical domain, this is an important approach to be thinking about."

Full Article
Drexel University Master's in HCI/UX
ACM Author's Rights

Association for Computing Machinery

1601 Broadway, 10th Floor
New York, NY 10019-7434

ACM Media Sales

If you are interested in advertising in ACM TechNews or other ACM publications, please contact ACM Media Sales or (212) 626-0686, or visit ACM Media for more information.

To submit feedback about ACM TechNews, contact: [email protected]