Messi (Hojun) Lee graduated from Hong Kong University of Science and Technology in 2020. He majored in Quantitative Social Analysis, and during his time in Hong Kong, he discovered his interest in computational social sciences. As an undergraduate, he studied ideologies and sentiments in censored song lyrics during the Third and Fourth Republic of South Korea and studied coordinated activities of Twitter accounts affiliated with political astroturfing campaigns.
He joined the Computational and Data Sciences Program (DCDS) at Washington University in St.Louis as a PhD student in the Fall of 2020. Since joining the program, he has been working with Professor Jacob Montgomery in the Department of Political Science and plans to pursue his interests in multi-party systems and voting behavior.
Messi is a huge fan of the Argentinian football player, Lionel Messi, and has been using the name since high school. He also enjoys playing football, watching Netflix, cooking, and listening to peoples’ philosophies.
Scholar Voices Feature
Messi Lee Presents Research at ACHM FAccT Conference
June 2024
Messi Lee attended the ACM Conference on Fairness, Accountability, and Transparency (FAccT) held in Rio de Janeiro, Brazil on June 3-6, 2024. The conference brings together researchers and practitioners interested in fairness, accountability, and transparency in socio-technical systems. It builds on the success of the previous edition of the conference, held in Chicago, US. Messi presented his paper titled “Large Language Models Portray Socially Subordinate Groups as More Homogeneous, Consistent with a Bias Observed in Humans.” The paper, presented as part of the Bias & factuality in Large Language Models session, is now accessible as an open access article at the ACM Digital Library: https://dl.acm.org/doi/10.1145/3630106.3658975.
Messi Lee is a PhD student in the Division of Computational & Data Sciences (DCDS) program, part of the McKelvey School of Engineering. He is a graduate of McDonnell Academy partner university Hong Kong University of Science and Technology.
Messi Lee Wins Outstanding Research Award
February 2024
Messi Lee awarded the Outstanding Research Award at the Society for Personality and Social Psychology (SPSP) 2024 Annual Convention for his project titled, “The Effect of Group Status on the Variability of Group Representations in LLM-generated Text.” The award recognizes rigorous and transparent research by graduate students, according to the SPSP website. Submissions were evaluated for the quality of the research design, transparency of the reported results, and legitimacy of the statistical conclusions. The goal of this award is to highlight outstanding empirical research regardless of the statistical significance of the results.
Messi Lee Publishes First Paper
January 2024
Messi Lee published his first paper, titled, “America’s Racial Framework of Superiority and Americanness Embedded in Natural Language.” The paper, coauthored with professors Calvin Lai and Jacob Montgomery, was accepted for publication at PNAS Nexus. In this work, they find that America’s racial framework, defined by the dimensions of superiority and Americanness, is embedded in American English using word embeddings, a computational model that maps words into a numerical space where the position and distance between words reflects their meanings and relationships based on how they are used in a large collection of text. Their research findings not only reinforce the idea that racial stereotypes are communicated through spoken and written language but also reveal an interesting nuance – Asian individuals are stereotyped as more ‘American’ than Hispanic individuals.
CRE2 Awards Messi Lee Small Grant
November 2022
McDonnell Scholar Hojun “Messi” Lee receives Small Grant for his project that aims to quantify racial and ethnic stereotypes in various forms of communication. The Center for the Study of Race, Ethnicity & Equity at Washington University in St. Louis awards a limited number of Small Grants of up to $1,000 to support activities related to the advancement of the CRE2 mission, three CRE2 pillars, and its programmatic themes. The proposed method for Messi’s project leverages semantic information encoded in Large Language Models to induce nuanced representations of racial and ethnic groups from small collections of text, which are then used to assess the extent to which the racial/ethnic groups are stereotyped with respect to specific stereotype dimensions (e.g. unpleasantness, inferiority, foreignness). The method is expected to help to understand the mechanisms involved in the propagation and maintenance of societal stereotypes.
Messi Lee is a third-year Scholar pursuing a PhD in Computational & Data Sciences. He is a graduate of Academy partner institution Hong Kong University of Science and Technology.
Visit https://cre2.wustl.edu/research/funding-opportunities/grants/small-grants/, if you are interested in learning more about CRE2’s Small Grants.