Marvin (Richie) Limpijankit | Columbia University

About Me

Hi 👋! I am a first-year Computer Science PhD student in the NLP lab at Columbia University under Prof. Kathy McKeown.

I was previously a data scientist at IBM 👁 🐝 Ⓜ️ for 2 years, working on applied AI/NLP in the Expert Labs group. Before that, I received my BS in Computer Science from Columbia SEAS’23 🦁, where I researched computational methods (CV, NLP) for studying news media with Prof. John R. Kender.

Research Interests

Interpretability for VLMs: concept decomposition, attention-based attribution, vision tokens
Explainability Evaluation: metrics for explanations, plausibility-faithfulness tradeoff

News

[Sept. 2025] Starting my PhD in Computer Science at Columbia!
[Aug. 2025] Paper accepted to INLG 2025!
[May. 2024] Workshop paper accepted to ICWSM 2024!
[Jun. 2023] Joining IBM Expert Labs as an NLP Data Scientist

Teaching

COMS 4771 Machine Learning w/ Prof. Nakul Verma in Spring’23, Fall ‘22
CSEE 3827 Fundamentals of Computer Systems w/ Prof. Dan Rubenstein (Spring ‘22)
COMS3203 Discrete Math w/ Prof. Dan Rubenstein (Fall ‘21)

Publications

INLG

Counterfactual Simulatability of LLM Explanations for Generation Tasks

Marvin Limpijankit, Yanda Chen, Melanie Subbiah, Nicholas Deas, Kathleen McKeown

International Natural Language Generation Conference (INLG), 2025.

PDF Code

ICWSM

Detecting Cultural Differences in News Video Thumbnails via Computational Aesthetics

Marvin Limpijankit, John R. Kender

Workshop @ International AAAI Conference on Web and Social Media (ICWSM), 2024.

PDF Workshop Slides

UNU

Does the United Nations need agents? Testing the role of AI agent generated personas in humanitarian action

Eduardo Albrecht, [Acknowledgements]

United Nations University.

PDF