About me
Hi, my name is Eric! I’m a 2nd-year CS Ph.D. candidate at Georgia Institute of Technology, advised by Prof. Srijan Kumar and part of the Computational Data Science Lab for the Web and Social Media (CLAWS) Lab. My research interests lie broadly in natural language processing, large language models, multimodal machine learning, and computational social science.
I have 3+ years research and 7+ years programming experience. In 2023, I presented my paper, “Characterizing and Predicting Social Correction on Twitter”, at the 15th ACM Web Science Conference (WebSci’23) in Austin, TX. The same year, I was awarded the Marshal D. Williamson Fellowship at Georgia Tech’s 32nd College of Computing Awards. I have contributed to several other projects, served as an author on multiple other publications, and peer-reviewed work for ICWSM ‘25 and HKS Misinformation Review.
In November 2024, I passed my Ph.D. qualifying exam - much thanks to my committee consisting of Srijan Kumar, Polo Chau, B. Aditya Prakash, and Bo Dai!
Current Work
At the moment, I am working on two projects. In the first, I am developing methods to benchmark and improve the robustness of multimodal large language models (MLLMs) to realistic and plausible semantic perturbations in the text and image modalities. Specifically:
- (1) I am working to create an extensive suite of input perturbations to enable a vigorous evaluation of MLLMs’ robustness across a myriad of reasoning tasks and state-of-the-art models.
- (2) I am looking to develop techniques to make MLLMs more robust to such perturbations.
In the second, I am investigating the dynamics of online misinformation and counter-misinformation. Questions I am trying to answer include:
- (1) What kinds of misinformation-spreading users are more likely to receive countering or debunking responses from ordinary online users?
- (2) What kinds of users are more likely to be the creators of misinformation-countering content?