RCSC Webinar Series: Creating a map of COVID-19 research using neural embeddings: A retrospective approach

Our fellow team from the Indiana University Bloomington will present research of creating a map of COVID-19 research.

Event Details

When: Oct. 13 at 3 p.m. ET

Where: Register for this webinar here. After registering, you will receive a confirmation email containing information about joining the meeting.

Abstract: The COVID-19 pandemic is expected to continue for a prolonged period, causing various social, political, and medical issues. Solutions to the issues are in urgent need, which has led to a proliferation of research papers on the COVID-19 across disciplines. The rapid expansion of the research makes it difficult for researchers to follow related research activities. Here, we create a map of papers on the COVID-19 and other related infectious diseases such as SARS and influenza by using more than 108,000 papers retrieved from Microsoft Academic Graph and Semantic Scholar. We employ neural embedding methods, with which we map research papers onto a low dimensional space based on citations. Then, we compare the density of papers on different diseases and annotate each cluster by using the content of the papers. The map reveals an emergent research area on the COVID-19 and less-explored areas in the COVID-19 research to date. Additionally, we find a separation between studies on network models and those on other infectious diseases.


Sadamori Kojaku is a postdoctoral fellow at Luddy School of Informatics, Computing, and Engineering, Indiana University Bloomington. He focuses on inherent biases in data and methods and leverages them to explore atypical activities in science such as cartel-like behavior in citation practice, and the explosion of publications on COVID-19 research. He earned his Ph. D in Computer Science from Hokkaido University in Japan and then worked as a research associate at Bristol University in the United Kingdom and Kobe University in Japan. Currently, he works on the science of science and computational social science using word and graph embeddings.

Read more about the researchers here.