Broadly, my research uses computational methods to study how meanings and concepts change over time, and how those shifts reflect wider cultural and social dynamics. At the intersection of NLP, computational social science, and psychology, I develop theory-informed approaches for tracing semantic and conceptual change in historical text corpora using contextual embeddings, large language models, lexical resources, and statistical modeling. My current work focuses especially on mental health language, media discourse, and socially contested language.

Research themes

Semantic and conceptual change
Methods for studying how meanings shift across time, including multidimensional approaches to lexical semantic change and synthetic benchmarks for method evaluation.
Key work: SIBling, LSC-Eval, Schizophrenia WSD.

Mental health language and media discourse
Applications to mental-health concepts and terminology, examining how meanings and representations shift across psychology, news media, books, and everyday language.
Key work: trauma, schizophrenia, ADHD, anxiety/depression, and generic mental-health terminology papers.

Meaning representations in humans and language models
Benchmarking how humans and language models represent word meanings, senses, connotation, denotation, valence, and arousal.
Key work: SenseRel, LSC-Eval.

Socially contested language
Computational social science work on evaluative, dehumanizing, identity-related, and contested language in online and public discourse.
Key work: ICWSM dehumanization paper, mental-health stigma benchmark, identity/person-first language paper, and common-good work.