This week’s list of data news highlights covers May 30-June 5, 2020, and includes articles about using AI to protect renters and increasing data portability between social media platforms.
SK Telecom, a South Korean telecommunications operator, and Omron Electronics Korea, which manufactures electronic equipment, have developed a robot that uses AI to combat the spread of coronavirus in SK Telecom’s office. The robot can detect if an individual’s temperature is over 99.5 degrees Fahrenheit, if an individual is wearing a mask, and if individuals are gathering in groups. The robot also uses ultraviolet lamps and disinfectant sprayers to disinfect surfaces.
Researchers led by an individual at Tel-Aviv University have analyzed the DNA of the Dead Sea Scrolls, which are made of sheepskin and cow skin, to determine which fragments are likely to be part of the same documents. The scrolls consist of more than 25,000 fragments, and the researchers made comparisons of mitochondrial DNA to determine which fragments came from the same or related animals. The researchers also used the DNA to determine that the scrolls likely originated from more areas in Israel than researchers previously thought.
Researchers from McMaster University in Canada and the American University of Beirut have analyzed data from 172 observational studies about COVID-19 to determine transmission rates associated with specific distances and protections, such as masks. For example, the researchers found the rate of transmission was 13 percent when individuals stood within one meter of each other, but three percent when they stood one meter apart. The researchers also found that wearing a mask and eye protection significantly reduced the risks of transmission.
LegalSifter, a technology firm based in Pittsburgh, and CloverContracts, a U.S. firm that offers legal software, have created software that uses AI to automate the review of leases. The software uses natural language processing to analyze contracts and provide guidance to minimize risks for renters. For example, the software may suggest “force majeure” text provisions that shield renters such as college students from being responsible for paying rent if they do not use the apartment due to coronavirus.
Researchers from Pucho Technology Information Limited, an Indian technology firm, have used a neural network and a supercomputer to develop drug molecules that may inhibit coronavirus’ spread in an infected person. The researchers generated chemical structures for 200 molecules that may inhibit 3CLpro, an important enzyme in coronavirus. The researchers identified 10 of the 200 molecules as potential inhibitors.
RealNetworks, a software company based in Washington, has created software that uses facial recognition to automatically recognize the faces of celebrities in YouTube and Netflix content. The software is a free Internet browser extension and allows users to skip ahead to a scene or segment where a particular individual is talking.
The U.S. Department of Health and Human Services has announced new guidance that requires laboratories to include demographic data in their COVID-19 testing reports to the federal government. The demographic data, which includes information such as race, ethnicity, age, and sex, can help illuminate how coronavirus is affecting African American and minority communities. Officials can then use the data to create better access to testing and treatment in areas that need it.
Researchers from MIT have developed an unsupervised learning algorithm to help map the ocean into communities based on the interactions of phytoplankton species. Concentrations of phytoplankton can indicate the health of an ecosystem, and the researchers used the algorithm to cluster areas of the ocean into groups using a dataset with information on 51 species of phytoplankton. The researchers found that 100 areas of the ocean are distinct in their ecological makeup.
Facebook has made a data portability tool that allows users to directly transfer their photos to Google Photos globally available. Users no longer have to download and manually upload the files themselves, making it easier for them to port their photos to another service. The tool uses open-source code from the Data Transfer Project, a collaboration by several organizations to make transferring data between platforms seamless.
Researchers from the University of Leeds in the U.K. used an AI system to classify 2,000 stars as large protostars, which are young stars that are still forming. Cataloging stars can help researchers understand the differences in the formation of low- and high-mass stars, and the researchers used the algorithm to analyze data on 4.1 million stars. The AI system, which identified roughly 2,000 stars as having at least an 85 percent probability of being a young protostar, increased the number of known stars in the class by order of magnitude.
Image: Earth Observatory