Munjung Kim

PhD Student in Data Science · University of Virginia
qns8tc@virginia.edu

Hi! I'm a PhD candidate in Data Science at University of Virginia. My research focuses on using and developing advanced AI and machine learning techniques to understand innovation in science and technology. Thesedays, I'm also interested in exploring how large language models can be enhanced and understood through the lens of collective intelligence. Before joining the PhD program, I completed my undergraduate studies in Physics at Pohang University of Science and Technology (POSTECH).

Machine Learning · Large Language Models · Science of Science


Recent News

Jan 2025 I delivered an invited talk at SNU and KISTI.

Dec 2024 I delivered an invited talk at KDI School and KAIST STP.

June - August 2024 I completed an internship at Nokia Bell Labs.

July 2023 I gave a poster presentation in ICS2S2.


Projects & Papers

The Murmuration of Scientists


Scientists shift their research interests—or “move” the space of knowledge—sparking new ideas and novel methodologies. While prior studies have explored how individual characteristics or direct collaborations shape scientists’ research topic shifts, they have largely overlooked the impact of the broader, dynamically evolving behavior of the peer community. Here, we create a continuous mapping of authors’ positions in a high-dimensional topic space to examine how scientists “move” through the knowledge space in relations with their peers. Drawing inspiration from the boids model of collective animal behaviors, we characterize in- dividual research movements using simple rules: alignment (matching the movement of peer scientists), cohesion (clustering with peers on similar topics), and separation (avoiding overlap when peers are too close). Our empirical analysis demonstrates that authors exhibit these propensities, with alignment and cohesion being modulated with the academic prominence. Furthermore, we develop a generative model grounded in these simple rules, which effectively predict future topic shifts based on observed peer dynamics. These findings reveal that the complex evolution of research interests may be driven by simple collective dynamics principles that resemble animal behaviors in physical space.

[paper]

project1

The Potential Impact of Disruptive AI Innovations on U.S. Occupations


The rapid rise of AI is poised to disrupt the labor market. , AI is not a monolith; its impact depends on both the nature of the innovation and the jobs it affects. While computational approaches are emerging, there is no consensus on how to systematically measure an innovation's disruptive potential. Here, we calculate the disruption index of 3,237 U.S. AI patents (2015-2022) and link them to job tasks to distinguish between "consolidating" AI innovations that reinforce existing structures and "disruptive" AI innovations that alter them. Our analysis reveals that consolidating AI primarily targets physical, routine, and solo tasks, common in manufacturing and construction in the Midwest and central states. By contrast, disruptive AI affects unpredictable and mental tasks, particularly in coastal science and technology sectors. Surprisingly, we also find that disruptive AI disproportionately affects areas already facing skilled labor shortages, suggesting disruptive AI technologies may accelerate change where workers are scarce rather than replacing a surplus. Ultimately, consolidating AI appears to extend current automation trends, while disruptive AI is set to transform complex mental work, with a notable exception for collaborative tasks.

[paper]

project1

Quantifying Disruptiveness using neural embedding method


Although understanding disruptive breakthroughs and their drivers hinges upon accurately quantifying disruptiveness, the core metric used in previous studies—the disruption index—remains insufficiently understood and tested. Here, after demonstratinits conflicting evaluations for simultaneous discoveries or “multiples”, we propose a new, continuous measure of disruptiveness based on a neural embedding framework that addresses these limitations. Our measure not only better distinguishes disruptive works, such as Nobel Prize-winning papers, from others, but also reveals simultaneous disruptions by allowing us to examine the “twins” that have the most similar future context. By offering a more robust and precise lens for identifying disruptive innovations and simultaneous discoveries, our study provides a foundation for deepening insights into the mechanisms driving scientific breakthroughs while establishing a more equitable basis for evaluating transformative contributions.

[paper]

project2

Quantifying topic diversity and disparity using neural embedding methods


Citation count is a popular index for assessing scientific papers. However, it depends on not only the quality of a paper but also various factors, such as conventionality, journal, team size, career age, and gender. Here, we examine the extent to which the conventionality of a paper is related to its citation count by using our measure, topic disparity. The topic disparity is the cosine distance between a paper and its discipline on a neural embedding space. Using this measure, we show that the topic disparity is negatively associated with citation count, even after controlling journal impact, team size, and the career age and gender of the first and last authors. This result indicates that less conventional research tends to receive fewer citations than conventional research. The topic disparity can be used to complement citation count and to recommend papers at the periphery of a discipline because of their less conventional topics.

[paper]

project2


Experience

Research Intern

Nokia Bell Labs

- Analyzed the differential impacts of disruptive vs. consolidating AI on job tasks using text embeddings, large language models, and network analysis.

- Accomplishment: One full paper [paper ]

Jun 2024 - Aug 2024

Education

University of Virginia

Doctor of Philosophy, Data Science
Advisor : Yong-Yeol Ahn, (transferred from Indiana University with advisor)
August 2025 - Present

Indiana University

Doctor of Philosophy, Informatics
Advisor : Yong-Yeol Ahn
August 2022 - July 2025

Pohang University of Science and Technology

Bachelor of Science, Physics
Feb 2017 - Dec 2021

Skills

Computer Language & Operating Systems

    Python, R, Matlab,C, HTML, UNIX Bash, Linux


AI, ML & DS Libraries

    Pytorch, scikit-learn, Numpy, Scipy, Pandas, NetworkX, igraph, gensim, statsmodels, Snakesmake


Awards & Certifications

  • Humane Studies Fellowship, Institute for Humane Studies, 2024, $5,000
  • National Scholarship for Excellence (Sci. & Eng.) - Korea Studnet Aid Foundation, 2017-2021, $20,000
  • Global Talent Attraction Program, Indiana University Bloomington, 2020, $4,000