This week’s list of data news highlights covers March 13, 2021 – March 19, 2021 and includes articles about building a new supercomputer in less than a year and saving coral reefs.
1. Using Data to Determine Which COVID-19 Vaccines Need to be Updated
The U.S. Food and Drug Administration (FDA) is collaborating with the Centers for Disease Control and Prevention and the Centers for Medicare and Medicaid Services to collect data on vaccine safety, clinical trials that have resulted in vaccine failures, and the infection rates of people who have already been vaccinated against COVID-19 to determine whether vaccines will be effective against new coronavirus variants. Currently, the FDA is only using data received from public health surveillance systems and reports on vaccine side effects from patients and physicians to assess when updates to vaccinations will be needed.
2. Building a New Supercomputer in Five Months
Nvidia, a company that develops graphic chips, has finished building the UK’s fastest supercomputer in five months, which is the quickest building time among the world’s 500 fastest supercomputers where the average build time was more than two years. This supercomputer is specialized for AI computations in healthcare applications, such as designing new drugs.
3. Creating an AI-Powered Water Meter to Prevent Leaks
WINT Water Intelligence, an Israeli startup focused on using AI to prevent water hazards, has developed an AI-powered water meter that analyzes water flows in pipe systems to detect leaks. The tool is integrated into existing pipe systems and alerts staff about the location of abnormal water flows so that they can promptly shut off malfunctioning parts. Mace, a UK company that built the London Eye, has incorporated the tool into its construction sites, saving it 25 percent in excess water consumption that leaking pipes cause.
4. Using Local Data to Understand Changes to Visitor Experiences due to COVID-19
Thanksgiving Point Institute, a farm, garden, and museum complex in Utah, is evaluating visitor data to understand how the reasons for visits have changed because of the pandemic. It has collected eight years worth of data on visitors’ motivations and reasons for visiting through surveys and gallery observations, and it is now comparing the pre-pandemic data to parallel data from current visitors. From this data, the institute discovered that there was a 400 percent increase in visitors whose goal was to recharge physically, emotionally, or spiritually at the site.
Researchers from Accenture, Intel, and the Sulubaaï Environmental Foundation in the Philippines have developed AI tools to help monitor the health of coral reefs, which are dying at increasing rates due to warmer ocean temperatures. Typically, divers collect data on reefs by capturing video and pictures, but this disrupts the surrounding marine life. To address this, the team has implanted artificial reefs equipped with an AI-powered video camera that counts and classifies marine life supporting the reefs, allowing researchers to make data-driven decisions about reef restoration projects.
6. Launching a Data Dashboard for Adverse Reactions to COVID-19 Treatments
The FDA has launched a public data dashboard that shows reports of adverse human reactions to COVID-19 drugs and therapeutic products that were granted emergency use authorization. The dashboard displays reports submitted by pharmaceutical companies, healthcare providers, and consumers and can be further broken down by report types, such as reports by age, sex, or seriousness of injury.
7. Responding to Consumer Concerns Using AI
Shanghai 12345, a 24-hour public service hotline run by the government in Shanghai, China, is using AI to collect and analyze data from consumers calling about concerns, such as a store refund not processing. The AI analyzes the density of calls and determines whether they are related to a common problem, helping officials more quickly address pressing issues. The hotline is available in nine languages and last year received more than 200,00 calls.
8. Supporting Bank Customers Impacted by Natural Disasters
The Commonwealth Bank of Australia is using data from national emergency and weather alert services to identify vulnerable customers who may be impacted by natural disasters and be in need of additional support. On average, the tool runs 400 machine learning models, analyzes 157 billion data points, and delivers 35 million decisions per day based on the data sources it is given. In February, the Bank used the tool to contact and support 80,000 customers impacted by bushfires in Perth.
9. Spotting Defects in Chip Manufacturing
Applied Materials, a U.S. semiconductor manufacturing company, is using AI to detect defects in semiconductors. Normally, an optical scanner is applied to chips at different points in the manufacturing process to identify up to one million problem areas. Then, an electron microscope determines whether the areas will cause a chip to malfunction. With AI technology embedded into the optical scanner, this days-long process is now completed in 15 minutes. Rather than looking at all the problem areas, the AI system only needs to check 1,000 specific trouble spots before the electron microscope examines the chip.
10. Making Instagram Safer for Minors
Instagram has unveiled two new tools designed to protect users under the age of 18. Instagram requires a user to be over the age of 13, but one of the new tools is preventing adults from messaging users under 18 who do not already follow them by using machine learning to predict their ages. The other new tool detects suspicious behavior, such as an adult sending a large number of message requests to users under 18, and then inserts a safety notice to the underage user, where they can either end the conversation by blocking, reporting, or restricting the adult.
Image credit: Li Fei