Use of AI and ML in the Fight against the Pandemic

Within a few months of COVID-19, the world has changed and we have been forced to look for solutions everywhere. As we grapple with the pandemic, harnessing the potential of technological advances brings us closer to the belief that we will overcome it. In this endeavor, artificial intelligence (AI) and machine learning (ML) have emerged as two key technologies which empower us to better understand and address the COVID-specific issues. Right from prediction to detection, screening, and diagnosing to inching closer towards a vaccine, these two technologies have a critical role to play.

Using ML to screen patients effectively

With every new pandemic, diagnosing individuals is a significant challenge. As tests are expensive at the beginning conducting them on a large scale is a near-impossible task. Any individual who develops symptoms of COVID-19 would be concerned that they have contracted the disease. Instead of taking medical samples from each patient and waiting for slow reports, a more affordable and faster ML-based test would always prove more useful.

Interestingly, a hospital in Florida was one of the first to use machine learning for a strong response to the virus. As patients enter the hospital an automatic face scan (which is based on ML) is conducted to detect whether they are running a fever. This piece of data plays a critical role in triaging patients. Additionally, ML-enabled bots help patients self-identify the next course of action once they are being detected with mild COVID-like symptoms. This saves doctors from spending excessive time on answering questions of worried patients instead of focusing on the treatment.

Detecting patients at high risk

Once a group of people is tested positive it becomes critical to detect patients at a higher risk of developing serious complications or in need of advanced medical care. While many develop symptoms which are mild others may develop acute respiratory distress syndrome (ARDS) or severe lung disease which needs immediate attention. In this context, significant inroads have been made in research and a forecast model has been proposed which depends on the XGBoost calculation (a machine learning algorithm). It helps detect the seriously sick and predict mortality risk. It takes into account three clinical pointers; lymphocytes, lactic dehydrogenase, and high affectability C-reactive protein, an imbalance in which leads to cell injury, compromised cell immunity, and inflammation, also indicative of pathophysiological progress of COVID-19.

Understanding the spread of the contagion

AI-based predictive modeling does not just help in identifying the infected individuals but also in gauging the severity and spread of the pandemic. Early warning systems which are AI-based mine mainstream news, information in other channels, and online content in multiple languages to detect epidemiological patterns which also complement other healthcare networks and syndromic surveillance. A majority of the countries worldwide have opted for contract tracing systems to ascertain the infection routes by using geo-location data to identify individuals who are coming in close contact with the virus carriers. Immediately, text messages are sent as alerts directing them to self-isolate.  

Towards a vaccine

AI has played a critical role in the race to discover a vaccine. By understanding viral protein structures, the technology has suggested probable components of the vaccine. It has also helped medical researchers to scour thousands of literature at an unprecedented pace and arrive at conclusions. As the scientific literature related to COVID continues to grow scientists find it challenging to shortlist papers that are relevant to their specific research. A reputed institute which conducts high-impact research created freely available data sets that are machine readable. This enabled researchers to create and apply NLP algorithms, which would accelerate the process of discovering a vaccine. By connecting the dots between studies, AI makes it possible to identify a hypothesis and subsequently suggests experiments.

Holding on to hope

While modern technologies have helped us so far in this intense battle against the pandemic, we hope that it would continue to assist us in achieving new victories and see to the end of this unfortunate situation. It would not be wrong to assume that the possibility of combining ML, AI, and other convergence points of sectors such as healthcare and life sciences presents a brilliant opportunity for new drug discovery. And, as we leverage this opportunity, we will only stand to gain.