Philhoon Oh

I am currently a master’s student at the Kim Jaechul Graduate School of AI, where I am fortunate to be advised by Prof. James Thorne and Prof. Jinwoo, Shin . Before joining KAIST, I was a NLP engineer in SK Inc. C&C, and I received my bachelor’s degree at UC Berkeley, majoring in Statistics.

I believe language plays a pivotal role in transferring knowledge. While humans appeared on Earth millions of years ago, civilization has only been built for thousands of years. It has only been in the last few hundred years that humans have made significant strides by accumulating knowledge through language. In other words, language enables people to think collectively beyond time and space. If we can expedite the sharing of knowledge by incorporating language into machines, it will be possible for humans to make further progress.

My primary interests lie in Information Retrieval (IR) and Retrieval Augmented Generation(RAG). I believe that any form of communication (conversations, audio, text, images) can be reformulated as query + (external knowledge) = response format. With this perspective, I am particularly eager to delve deeper into areas described below:

Retrieval Augmented Generation(RAG)
- RAG has been effective tool to deal with long-tail knowledge, hallucinations, and up-to-date information.
- My interests include how to improve the RAG pipeline and overcome the limitation of RAG for better user experience.
Information Retrieval
- Information Retrieval encompasses a range of intriguing topics, such as Open-Domain Question Answering (ODQA), Multi-hop reasoning, and more.
- Let’s search for information that can help align with the user’s intention.
LLMs (Large Langauge Models)
- Interested in efficient utilization of LLMs in various domains (Multi-Modal, Tabular, etc)
- Some potential methods include prompting, ICL (In-Context Learning), and the utilization of external knowledge sources.

Email: philhoonoh@kaist.ac.kr / vlfgns5@gmail.com

News

Feb 20, 2024	1 paper accepted to LREC-COLING 2024
Jan 2, 2024	Started as Research Intern at NAVER Search
Oct 7, 2023	2 papers accepted to EMNLP 2023
Aug 22, 2023	2nd Place in the 2023 AI Graduate School Challenge (KT CTO Award)
Feb 27, 2023	Joined XFACT Lab at KAIST as a Master student.

Publications

Under Review

Parallel Key-Value Cache Fusion for Position Invariant RAG

Philhoon Oh, Jinwoo Shin, and James Thorne

2025

arXiv Bib

@misc{oh2025parallelkeyvaluecachefusion,
  title = {Parallel Key-Value Cache Fusion for Position Invariant RAG},
  author = {Oh, Philhoon and Shin, Jinwoo and Thorne, James},
  year = {2025},
  eprint = {2501.07523},
  archiveprefix = {arXiv},
  primaryclass = {cs.AI},
  url = {https://arxiv.org/abs/2501.07523},
  bibtex_show = true,
}

LREC-COLING

CLIcK: A Benchmark Dataset of Cultural and Linguistic Intelligence in Korean

Eunsu Kim, Juyoung Suk, Philhoon Oh, and 3 more authors

In Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024), May 2024

arXiv Bib

@inproceedings{kim-etal-2024-click,
  title = {{CLI}c{K}: A Benchmark Dataset of Cultural and Linguistic Intelligence in {K}orean},
  author = {Kim, Eunsu and Suk, Juyoung and Oh, Philhoon and Yoo, Haneul and Thorne, James and Oh, Alice},
  editor = {Calzolari, Nicoletta and Kan, Min-Yen and Hoste, Veronique and Lenci, Alessandro and Sakti, Sakriani and Xue, Nianwen},
  booktitle = {Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)},
  month = may,
  year = {2024},
  address = {Torino, Italia},
  publisher = {ELRA and ICCL},
  url = {https://aclanthology.org/2024.lrec-main.296/},
  pages = {3335--3346},
  bibtex_show = true,
}

EMNLP

Detrimental Contexts in Open-Domain Question Answering

Philhoon Oh, and James Thorne

In Findings of the Association for Computational Linguistics: EMNLP 2023, Dec 2023

arXiv Bib

@inproceedings{oh-thorne-2023-detrimental,
  title = {Detrimental Contexts in Open-Domain Question Answering},
  author = {Oh, Philhoon and Thorne, James},
  editor = {Bouamor, Houda and Pino, Juan and Bali, Kalika},
  booktitle = {Findings of the Association for Computational Linguistics: EMNLP 2023},
  month = dec,
  year = {2023},
  address = {Singapore},
  publisher = {Association for Computational Linguistics},
  url = {https://aclanthology.org/2023.findings-emnlp.776},
  doi = {10.18653/v1/2023.findings-emnlp.776},
  pages = {11589--11605},
  bibtex_show = true
}

EMNLP

Knowledge Corpus Error in Question Answering

Yejoon Lee, Philhoon Oh, and James Thorne

In Findings of the Association for Computational Linguistics: EMNLP 2023, Dec 2023

arXiv Bib

@inproceedings{lee-etal-2023-knowledge,
  title = {Knowledge Corpus Error in Question Answering},
  author = {Lee, Yejoon and Oh, Philhoon and Thorne, James},
  editor = {Bouamor, Houda and Pino, Juan and Bali, Kalika},
  booktitle = {Findings of the Association for Computational Linguistics: EMNLP 2023},
  month = dec,
  year = {2023},
  address = {Singapore},
  publisher = {Association for Computational Linguistics},
  url = {https://aclanthology.org/2023.findings-emnlp.616},
  doi = {10.18653/v1/2023.findings-emnlp.616},
  pages = {9183--9197},
  bibtex_show = true
}