My work applies corpus linguistics and machine learning across domains where language carries evidence about people: mental health, language learning, and law.
Machine learning, topic modeling, and corpus-linguistic analysis of depression-related language in online communities, the core of my dissertation research.
Corpus-based Studies across Humanities (2025) · PTSD/anxiety study in preparation
Tracing collocation use, readability, and cohesion across proficiency levels with interpretable machine learning on learner corpora.
ELLIPSE Corpus, IJLCR (2023) · Applied Linguistics & Applied Corpus Linguistics (under review)
Advising J.D. candidates on NLP tools and corpus-linguistic approaches to the interpretation of U.S. legal documents.
Georgia State University College of Law · research advisory roles
Building GenAI-based interaction systems, chatbots with TTS and STT, for recently immigrated English learners, and studying their linguistic effects.
Research Methods in Applied Linguistics (2026) · RCALL seed grant ($48,500) · Cambridge English Funded Research Programme (FRP), £13,000
Large collections of real language use, analyzed systematically: frequency, collocation, keyness, and concordance methods that make claims about language verifiable at scale.
Computational pipelines that turn raw text into linguistic features: tokenization, POS tagging, dependency parsing, embeddings, and psycholinguistic feature engineering.
SHAP, LIME, permutation importance, and feature ablation, used to open black-box models and state findings in linguistic terms rather than model weights.
Research on the Challenges of Acquiring Language & Literacy (RCALL) · Seed grants, Georgia State University · $48,500
YouJin Kim (PI) and Daniel Dixon (Co-PI) · May 2024 – May 2025
My role: Research Assistant, Software Developer, Georgia State University, Department of Applied Linguistics.
Cambridge English Funded Research Programme (FRP) · £13,000
YouJin Kim (PI) and Daniel Dixon (Co-PI) · August 2023 – December 2023
My role: Research Assistant, Software Developer, Georgia State University, Department of Applied Linguistics.
A curated summary of the computational toolkit behind this research.
Python and R across the full research pipeline.
From t-tests to structural equation modeling.
Feature engineering from raw text: contextual embeddings to classic corpus statistics.
Classical models through deep sequence architectures, plus ensembling.
Opening black-box models from linguistic perspectives.
GPU-accelerated workflows and principled tuning.