As the world struggles with COVID-19, even a gram of technological innovation and skill to combat the pandemic brings us one step closer to defeating this crisis. Artificial intelligence and machine learning play an important role in understanding and combating the COVID-19 crisis. Machine learning technology enables computers to mimic human intelligence and quickly present models and insights by analyzing big data.
In the process of combating COVID-19, organizations have quickly applied their machine learning expertise to various fields. These include scaling customer communication, understanding how COVID-19 is spreading, and accelerating research and treatment.
Enabling Your Organization to Scale and Adapt
As social distance and quarantine measures continue, all organizations, whether small or large, private or public, are looking for new ways to meet the needs of their customers and employees in the most effective way. Machine learning technology makes this transition possible by providing the necessary tools for remote communication support, telemedicine and food safety protection.
Tools for healthcare and government installations include machine learning-compatible chatbots that enable contactless detection of COVID-19 symptoms and answer the public's questions. French start-up Clevy.io, which uses the AWS infrastructure, is one example of this. Clevy.io has launched a chatbot so people can easily find government statements about COVID-19.
Using real-time information from the French government and the World Health Organization, the chatbot assesses known symptoms and answers questions about government policies. The chatbot, which has received 3 million messages to date, can answer all kinds of questions from evaluating the risks related to COVID-19 to what exercises can be done without straining the resources of health institutions and the government. Cities like Strasbourg, Orléans, and Nanterre use this chatbot to deliver accurate, validated information.
Understanding How COVID-19 Spreads
Machine learning also helps researchers and doctors analyze large amounts of data to understand the spread of COVID-19. In this way, it is aimed to establish an early warning system for future pandemic risks and to identify vulnerable communities.
Chan Zuckerberg Biohub in California created a model to estimate the number of undetectable COVID-19 infections and assess its impact on public health and analyzed 19 regions globally. They developed new methods to quantify undetectable infections using machine learning and collaborating with the AWS Diagnostic Development Initiative. They analyzed how the virus mutated while spreading in the community to determine how many infections were missed.
Canadian start-up BlueDot, which used artificial intelligence to detect the onset of outbreaks on the AWS platform in the early days of the pandemic, was one of the first organizations to detect and report the spread of a respiratory disease in Wuhan, China. BlueDot uses artificial intelligence to detect these outbreaks.
Using machine learning algorithms, BlueDot scans for news in 65 languages as well as airline data and networks related to animal diseases to detect outbreaks and predict how they can spread. The epidemiologists then evaluate the results and confirm whether the conclusions are scientifically significant. BlueDot provides these insights to public health officials, airlines and hospitals so that they can anticipate and better manage risks.
Speeding Up Research and Treatment
Healthcare providers and researchers are exposed to an increasing amount of information about COVID-19, making it difficult to extract insights into treatment. AWS introduced CORD-19 Search in response.
This new search website draws its power from machine learning and allows researchers to quickly and easily access research papers and documents, "What is the period when the virus load in saliva is highest in COVID-19?" It helps to answer questions such as. Built on the CORD-19 open research dataset, which contains more than 128,000 research papers and similar materials from the Allen Artificial Intelligence Institute, this machine learning solution can extract medical information from unstructured texts and accelerate the discovery process by offering natural language query capabilities.
In the field of medical imaging, on the other hand, researchers are using machine learning to identify patterns in images, to enable radiologists to identify the possibility of disease and make early diagnosis. UC San Diego Health has developed a new method for diagnosing pneumonia associated with severe COVID-19. This early detection helps doctors quickly classify patients at the right level before a diagnosis of COVID-19 is confirmed.
The machine learning algorithm, trained with 22 thousand records created by human radiologists, combines x-rays with color-coded maps to determine the possibility of pneumonia. With donations from the AWS Diagnostic Development Initiative, these methods are now applied to all lung x-rays and tomographs taken for UC San Diego Health's clinical trials.
