COVID-19: Machine Learning & AI

Video

This was presented at the DREAM Fellowship Symposium in 2020.

Background

Due to the onset of the unprecedented global pandemic in 2020, all labs and research projects at UCSD were suspended from physical activity on campus. This put quite a large impediment on my research as I do much of my work on campus, interacting with the university's systems. As I returned home due to the lockdown imposed statewide, I thought about how I would adapt my research in response to this turbulent shift in locale and availability of resources. After mulling over possible avenues of research and (virtual) discussions with my fellowship coordinator and research advisor, I decided that I would take a detailed dive into possibly the most relevant topic in my field at the time: how AI and machine learning could be used in the fight against COVID-19.

Being such an important topic, there were certain considerations that had to be thoughtfully examined before I could start working on this project.

1. What medium should the project ultimately be in?
2. What sources should my research in this area look at?
3. What level of depth should these sources be explained in the final project?

In deciding the medium for the project, I knew that it had to be delivered in some form at a research symposium at the end of the year. However, given the importance of the information, I also wanted to make it easily accessible to anyone who was interested in the topic. I decided that a video around five minutes in length that incorporated relevant visuals and narration would be optimal in terms of both being able to present virtually to the Symposium as well as sharing online. When pondering the next question, the sources of research, I recognized that information about COVID-19 was rapidly being disseminated and that traditional outlets may not have everything I was looking for. With this in mind, I focused on finding sources from sources that were scientifically reputable and provided timely information.

Research

To find how AI & ML were being used in the pandemic, I reviewed the literature and found preprint papers on medRxiv that detailed the use of these technologies in the context of the coronavirus. I also looked at reputed publications like the MIT Technology Review for information on companies using AI to do things like verify proper social distancing. Lastly, I reached out to my network at UCSD and interviewed a professor I had taken a course with who is renowned in virology. After I gathered a sufficient amount of material from a broad array of sources, I arrived at the final question: how deep should I delve into this information? Considering that I was going to present my research into this topic at a symposium to academics not all in my field (as well as publicly online), a clear, methodical overview without copious scientific detail seemed best.

Finished Product

I produced and edited the video with an intention to make it visually engaging and deliver interesting information, deliberately taking care to not bore the viewer. I also took great care to ensure the video was of a professional quality and something one may see from a large news organization. As I presented the video at the symposium, it received a glowing response, with many commending the timeliness of the video and its ability to communicate complex topics effectively. The link was shared around various departments at UCSD and was featured in my department's newsletter. Hundreds of people from all over the world viewed the video on YouTube and I received many messages from people I had no connection with saying they found it fascinating.