Prediction of COVID-19 Diagnosis Using Machine Learning

Introduction

As the world tries to deal with the COVID-19 pandemic, every bit of technological innovation and imagination harnessed to combat this virus brings us a step closer to defeating it. Artificial intelligence (AI) and Machine Learning (ML) are playing a primary role to better understand and address the COVID-19 Diagnosis. ML technology allows computers to imitate human intelligence and consume huge volumes of data to quickly and efficiently detect insights and patterns.

To combat this deadly virus, organizations have efficiently leveraged their ML expertise in various areas, such as understanding how COVID-19 spreads, scaling customer communications and speeding up research and treatment.

How can ML Help in COVID-19 Diagnosis

ML technology is playing a critical role in enabling the shift by offering the tools to enable telemedicine, sustain remote communication, and protect food security.

For government and healthcare institutions, this includes using ML-enabled chatbots to screen COVID-19 symptoms without contact and to respond to questions from the public. One such example is Clevio.io, a French start-up and AWS customer. This start-up has introduced a chatbot to enable people to find official government communications about the pandemic. With almost 3 million messages sent to date, this chatbot can answer various questions ranging from exercise to an evaluation of COVID-19 threats. The strain on healthcare and government institutions is thus automatically reduced.

To prevent any interruption to the food supply chain, food processors and governments need to analyze and understand the current state of agriculture. Mantle Labs, another AWS customer, is providing its cutting-edge AI-driven crop surveillance solution to retailers. This solution is free of charge for a period of three months to offer certainty and additional resiliency to supply chains in the United Kingdom.

ML is also assisting practitioners and researchers examine huge volumes of data to predict the spread of the pandemic, to enable an early warning system for future pandemics and to recognize vulnerable populations.

ML is helping leaders make informed decisions in the face of the pandemic. In March, a group of volunteers led by former White House Chief Data Scientist DJ Patil, reached out to AWS for help to support a scenario-planning tool that demonstrated the potential impact of COVID-19 to answer questions, such as: “How many hospital beds will be needed?” or “For how long should a shelter-in-place order needs to be issued?” They needed to scale their open-source model so that governors across the US could recognize the volume of infection, exposure, and hospitalization to scale up their response plans.

Scientists and researchers at the Chan Zuckerberg Biohub in California have created a model to evaluate the number of undetected COVID-19 infections and their consequences on public health by analyzing twelve regions throughout the globe. Partnering with the AWS Diagnostic Development Initiative and by using ML, they have invented methods to measure undetected infection while examining how the virus mutates as it spreads across the public to deduce the number of transmissions that may have been missed.

Organizations are also trying to identify ways to cut the spread of the virus, specifically amongst vulnerable populations. Closedloop, an AI start-up is using its expertise in healthcare data to identify those at the highest threat of severe complications from COVID-19. 

Conclusion

As the virus progresses, sharing and continuous recording of robust data between the scientific community and public organizations is crucial. Increasing understanding of the contribution of various symptoms to diagnosing the disease needs to be combined with additional symptoms into the ML models.