PhD student · Computer Science and Engineering · The Ohio State University
Hi, I'm Muge.
/mu gə/ Hear my name
I work on multilingual NLP, retrieval, and multimodal medical reasoning: how language models transfer across writing systems, find the right evidence, and reason over clinical data.
About
I am a PhD student in the Department of Computer Science and Engineering at The Ohio State University
, advised by Sachin Kumar. I am broadly interested in multimodal large language models, reasoning, and retrieval-augmented systems, and in understanding how models access information, reason over it, and ultimately become more reliable for the people who use them.
My current focus is on strengthening the connection between retrieval and reasoning, supporting multilingual and cross-modal understanding, and developing methods that help LLMs handle complex, structured inputs. More generally, I want to help build AI systems that pair strong reasoning with practical usefulness for everyday users.
News
- One Form to Transfer Them All was accepted to EMNLP 2026. See you in Budapest!
- New preprint: One Form to Transfer Them All. Pretraining multilingual language models on romanized text instead of native scripts gives the strongest cross-lingual transfer, and the gap remains with scale.
- Started as a teaching assistant this semester (Autumn 2026) for CSE 3521 Introduction to Artificial Intelligence and CSE 5052 AI for Non-Majors.
- New preprint: When Cases Get Rare, a retrieval benchmark for off-guideline clinical question answering built from expert-validated case reports.
- Started my PhD in Computer Science and Engineering at The Ohio State University, advised by Sachin Kumar.
- Awarded a Graduate School Fellowship from The Ohio State University, where I will start my PhD in the fall.
Selected publications
All publicationsOne Form to Transfer Them All: Pretraining Multilingual Language Models Beyond Native Orthography
@inproceedings{zhang2026oneform,
title = {One Form to Transfer Them All: Pretraining Multilingual Language Models Beyond Native Orthography},
author = {Zhang, Muge and Jencks, Aaron and Badikela, Krishna and Tsvetkov, Yulia and Kumar, Sachin},
booktitle = {Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing (EMNLP)},
year = {2026},
url = {https://arxiv.org/abs/2608.25904}
}When Cases Get Rare: A Retrieval Benchmark for Off-Guideline Clinical Question Answering
@article{lee2026ogcarebench,
title = {When Cases Get Rare: A Retrieval Benchmark for Off-Guideline Clinical Question Answering},
author = {Lee, Doeun and Zhang, Muge and Yu, Yi and Manne, Ashish and Koesters, Stephen and Wen, Frank and Buchanan, Brady and Villagomez, Lynda and Moninuola, Oluwatoba and Lim, James and Tobin, Kathryn and Srisuwananukorn, Andrew and Zhang, Ping and Kumar, Sachin},
journal = {arXiv preprint arXiv:2605.21807},
year = {2026},
url = {https://arxiv.org/abs/2605.21807}
}