Tania Chakraborty

2026

Splits! A Flexible Dataset and Evaluation Framework for Sociocultural Linguistic Investigation
ACL 2026 · Main Conference, Talk ★ SAC Highlight Award
Eylon Caplan, Tania Chakraborty, and Dan Goldwasser
paper data website
A sandbox to evaluate the merit of potential socio-linguistic hypotheses. Have a hypothesis about how people speak? Test it out here!
Datasets and Methods for Improving the Cultural Capabilities of NLP Systems: A Survey
NLP+CSS Workshop 2026
Tania Chakraborty, Eylon Caplan, Zhaoqing Wu, Kevin Cushing, Bruce Qin, Shreya Havaldar, and Dan Goldwasser
paper website
What is the field of Cultural NLP up to? Check out CultureMine to find cool work (and add yours)!

2025

VIBE: Can a VLM Read the Room?
Findings of EMNLP 2025 · Poster
Tania Chakraborty, Eylon Caplan, and Dan Goldwasser
paper data
Why do VLMs struggle to make social judgements that come so intuitively to humans?
A year and half after VIBE came out, the latest VLMs still struggle on our task. Read about it here!

2021

A Large Scale Low-Resource Pronunciation Data Set Mined From Wikipedia
arXiv Preprint
Tania Chakraborty, Manasa Prasad, Theresa Breiner, Sandy Ritchie, and Daan van Esch
paper data
We mined pronunciation data for 819 languages from Wikipedia so you don’t have to.