AWS Marketplace Webinar
Welcome to the October 5, 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."

Headquarters of IQVIA, the contract research organization helping manage AstraZeneca’s Covid-19 vaccine trial. Clinical Trials Hit by Ransomware Attack on Health Tech Firm
The New York Times
Nicole Perlroth
October 3, 2020

Philadelphia-based software provider eResearch Technology (ERT) was hit two weeks ago by a ransomware attack that has slowed clinical trials. The exploit started when ERT workers learned that they were locked out of their data, and clients said this forced researchers to move certain clinical trials to pen and paper. ERT's Drew Bustos on Friday verified that ransomware had hijacked company systems on Sept. 20, when the firm took its systems offline, called in outside cybersecurity experts, and alerted the U.S. Federal Bureau of Investigation. Affected customers included IQVIA, the contract research organization helping manage AstraZeneca's Covid-19 vaccine trial, and drug maker Bristol Myers Squibb, which is leading a consortium in developing a rapid test for coronavirus.

Full Article
*May Require Paid Registration

A man wearing a mask. 3D-Printed 'Invisible' Fibers Can Sense Breath, Sound, Biological Cells
University of Cambridge (U.K.)
Sarah Collins
September 30, 2020

Researchers at the U.K.'s University of Cambridge used three-dimensional (3D) printing to manufacture ultra-thin electronic fibers invisible to the naked eye, as part of the process of building contactless, wearable, portable respiratory sensors. The devices can be attached to mobile phones to simultaneously record breath pattern data, sound, and images. Cambridge's Andy Wang employed the sensors to test the amount of breath moisture leaked through his mask for conditions like normal and rapid breathing, and simulated coughing; the sensors outperformed comparable commercial devices, especially for tracking rapid breathing. Cambridge's Yan Yan Shery Huang said, "Our fiber sensors ... could potentially be turned into home-test devices to allow the general public to perform self-administered tests to get information about their environments."

Full Article
'Digital Chemistry' Breakthrough Turns Words Into Molecules
University of Glasgow (U.K.)
October 1, 2020

Researchers from the University of Glasgow in Scotland have developed a system they say will lead to the creation of a "Spotify for chemistry." Chemify, a digital chemistry system that can convert words into molecules, is a desktop-sized chemical processing unit that can perform the repetitive, time-consuming work of creating chemicals. The researchers crafted a universal approach to chemistry digitization using a coding system that applies an algorithm called SynthReader, which can scan scientific papers and identify sections that outline procedures for chemical synthesis, parsing them into simple instructions and storing them in the open source Chemical Description Language (XDL). An interface called ChemIDE integrates with any robot chemist system, and enables the instructions to be converted into chemicals. Glasgow's Lee Cronin said, "Putting that kind of knowledge directly in the hands of people with access to robot chemists could help doctors make drugs on demand in the future."

Full Article
Women Founders of AI Startups Take Aim at Gender Bias
The Wall Street Journal
Te-Ping Chen
September 29, 2020

Women launching startups are making inroads in the largely male-dominated field of artificial intelligence, with some seeking to build more diverse workforces and promote women in tech more broadly, while others look to address concerns about potential biases against marginalized communities. Said Textio's Kieran Snyder, "I think people who've been outnumbered are more likely to have a sense of why an accurate, equitable dataset is important, and to want to do that work." McKinsey & Co.'s Lareina Yee indicates the shortage of women in tech is a recent U.S. phenomenon, citing National Science Foundation data showing that the percentage of female undergraduates studying computer sciences fell from 29% in 1995 to 18% two decades later.

Full Article
*May Require Paid Registration
Technology Helping Fire-Struck Communities Predict Air Quality Better
National Geographic
Melanie Ehrenkranz
October 2, 2020

As wildfires ravage the U.S. West Coast, communities increasingly are turning to air quality measurements captured by low-cost sensors, in conjunction with maps available on applications like AirNow, IQAir, and PurpleAir. The coffee cup-sized sensors offer near-real-time measurements by using lasers to count airborne particles. Citizen scientists can buy their own sensors for about $200, connect them to Wi-Fi, and their readings will appear on the associated air quality map. The sensors tend to overreport particulate matter levels, and federal agencies like the Environmental Protection Agency are working to correct the devices' algorithms so they will deliver data in closer alignment with higher-quality monitors.

Full Article
*May Require Paid Registration
Flawed Algorithm Used to Determine U.K. Welfare Payments Is 'Pushing People Into Poverty'
The Next Web
Thomas Macaulay
September 29, 2020

Human Rights Watch warns a flawed algorithm for calculating monthly social security benefits in Britain is causing hunger, debt, and psychological distress. The model measures changes in their earnings to dole out payments, but the non-governmental organization said the algorithm only analyzes wages people receive within a calendar month, and ignores frequency of payment. This means people who get multiple monthly paychecks can have their earnings overestimated, with their welfare payments dramatically reduced as a result. Human Rights Watch’s Amos Toh said, "The government's bid to automate the benefits system—no matter the human cost—is pushing people to the brink of poverty."

Full Article

A GitHub logo seen displayed on a smartphone. GitHub Launches Code Scanning to Unearth Vulnerabilities Early
Paul Sawers
September 30, 2020

GitHub last week launched a code-scanning tool to help developers identify flaws in code prior to its public rollout. A result of GitHub’s takeover last year of code analysis platform Semmle, the new tool is a static application security testing solution that converts code into a queryable format, then searches for vulnerability patterns. It automatically identifies flaws and errors in code revisions in real time, alerting the developer before the code approaches production. GitHub said during the scanner's beta-testing phase it scanned more than 12,000 repositories more than 1 million times, discovering 20,000 vulnerabilities; developers and maintainers corrected 72% of these errors within 30 days.

Full Article
AI Helping Scientists Discover Fresh Craters on Mars
Jet Propulsion Laboratory
October 1, 2020

Planetary scientists and artificial intelligence (AI) researchers at the U.S. National Aeronautics and Space Administration Jet Propulsion Laboratory (JPL) developed a machine learning tool that is helping to discover new craters on Mars. Fresh crater detection typically involves hours of studying low-resolution images of potential new craters captured by the Mars Reconnaissance Orbiter's Context Camera; the new process scans images from the High-Resolution Imaging Science Experiment. The classifier can complete in about five seconds what would take a human 40 minutes to accomplish. JPL's Kiri Wagstaff said. "This paves the way for an exciting symbiosis of human and AI 'investigators' working together to accelerate scientific discovery."

Full Article
Tool Helps Clear Biases From Computer Vision
Princeton Engineering News
Molly Sharlach
October 1, 2020

Princeton University computer scientists have developed an open source tool that identifies potential biases in image sets used to train artificial intelligence systems. REVISE (REvealing VIsual biaSEs) can fix issues of underrepresentation or stereotypical portrayals in datasets before those collections are used to train computer vision models. The tool employs statistical methods to check a dataset for potential biases or issues of underrepresentation along object-based, gender-based, and geography-based lines. A related Princeton study investigated strategies for preventing computer-vision models from learning spurious correlations that may reflect biases. That study found the approach of domain-independent training, or "fairness through awareness," can mitigate potential biases in computer vision by considering a protected attribute separately from other visual cues.

Full Article
Australia's Report on Agtech Confirms Technology Can Lead to Fertile Future
Aimee Chanthadavong
September 29, 2020

The Australian Council of Learned Academics' Future of Agriculture Technologies report says emerging technologies like sensors, robotics, artificial intelligence (AI), biotechnology, nanotechnology, and distributed ledgers could enhance the agricultural sector's productivity, diversity, and profitability. The report emphasized how technologies could potentially generate vast datasets that could assist farmers in decision-making and environmental monitoring, so they could focus more attention on complex tasks. Moreover, the report says, properly deployed AI and IoT could underpin solutions like asset automation and rapid testing of localized crops, reducing costs and boosting investment in computational hardware, software, and algorithm development.

Full Article

Tijana Radivojevic (left) and Hector Garcia Martin working on mechanical and statistical modeling, data visualizations, and metabolic maps at the Agile BioFoundry last year. Machine Learning Takes on Synthetic Biology: Algorithms Can Bioengineer Cells for You
Lawrence Berkeley National Laboratory
September 25, 2020

A tool developed by researchers at the Lawrence Berkeley National Laboratory adapts machines learning algorithms to the field of synthetic biology to predict how changes in a cell's DNA or biochemistry will impact its behavior, and to make recommendations for the next engineering cycle. The researchers used the Automated Recommendation Tool (ART) to guide the metabolic engineering process to increase production of the amino acid tryptophan. They selected five genes representing a total of almost 8,000 potential combinations of biological pathways; ART ultimately recommended a design that boosted tryptophan production by 106% over the state-of-the-art reference strain and 17% over the best designs used for training the model. Said Berkeley Lab's Hector Garcia Martin, "This is a clear demonstration that bioengineering led by machine learning is feasible, and disruptive if scalable."

Full Article
Critical Flaws Discovered in Popular Industrial Remote Access Systems
The Hacker News
Ravie Lakshmanan
October 1, 2020

Researchers at Israel's OTORIO industrial cybersecurity firm found critical defects in two popular industrial remote access systems that attackers could exploit to block access to industrial production floors, infiltrate company networks, tamper with data, and steal business secrets. The analysts found flaws in B&R Automation's SiteManager and GateManager ranging from path traversal to improper authentication, which could enable hackers to view sensitive data about other users, their assets, and their processes. Meanwhile, the analysts said, MB Connect Line's mbCONNECT24 was found to contain flaws that could enable attackers to access arbitrary information through Structured Query Language injection, steal session details in a cross-site request forgery attack, and leverage unused third-party libraries bundled with the software to obtain remote code execution. The flaws in both systems reportedly have been corrected.

Full Article

Researchers showed that an AI algorithm could be trained to classify COVID-19 pneumonia in computed tomography scans with up to 90% accuracy. AI Can Detect Covid-19 in the Lungs Like a Virtual Physician, Study Shows
University of Central Florida
Robert Wells
September 30, 2020

Researchers at the University of Central Florida (UCF) and the U.S. National Institutes of Health have developed an artificial intelligence (AI) algorithm that is almost as accurate as a physician in diagnosing Covid-19 in the lungs and distinguishing Covid-19 cases from influenza. They trained the algorithm to recognize Covid-19 in computed tomography (CT) scans of 1,280 patients, then tested the algorithm on CT scans of 1,337 patients with various lung diseases. They found the algorithm could be trained to classify Covid-19 pneumonia in CT scans with up to 90% accuracy, and to correctly identify positive cases 84% of the time and negative cases 93% of the time. Said UCF's Ulas Bagci, "We showed that robust AI models can achieve up to 90% accuracy in independent test populations, maintain high specificity in non-Covid-19 related pneumonias, and demonstrate sufficient generalizability to unseen patient populations and centers."

Full Article
Top 10 Online Master's
ACM Transactions on Data Science

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]