Meet the Fellow: Byung-Doh Oh

This entry is part of our Meet the Fellow blog series, which introduces and highlights Faculty Fellows who have recently joined CDS.

Meet CDS Faculty Fellow Byung-Doh Oh, who will be joining us in September 2024. Oh has recently received his PhD in Linguistics at The Ohio State University, where he worked with Professor William Schuler. His research primarily focuses on cognitive modeling and neural network interpretability.

Oh’s academic journey began in Seoul, South Korea, where he earned his BA in English Language Education (summa cum laude) and MA in English Language Education from Seoul National University. His transition to computational linguistics marked the beginning of his interdisciplinary approach to understanding language processing.

At the intersection of linguistics and data science, Oh’s work explores how both humans and machines process language. His recent research investigates the differences in how people and large language models process syntactically challenging constructions, particularly focusing on “garden path” sentences — a type of syntactically ambiguous sentence that initially misleads the reader.

“I look forward to joining CDS and continuing to study how people and machines process language,” said Oh. “In addition to developing new collaborations, I am particularly excited to gain methodological insights from the many disciplines at CDS.”

Oh has already begun collaborating with CDS Associate Professor of Linguistics and Data Science Tal Linzen, examining the characteristics of by-item variation in both human subjects and language models when processing these challenging sentences. This work aims to provide insights into how syntax and semantics interact in both human and artificial language processing systems.

During his time at CDS, Oh also hopes to develop new interpretability techniques for studying how language models build representations while processing such complex sentences. This approach complements his ongoing work and could provide valuable insights into the inner workings of these powerful AI systems.

Oh’s research has been published in prestigious venues such as the Proceedings of the Association for Computational Linguistics (ACL), Transactions of the Association for Computational Linguistics (TACL), and Frontiers in Artificial Intelligence. His recent paper, “Frequency explains the inverse correlation of large language models’ size, training data amount, and surprisal’s fit to reading times,” was presented at the 18th Conference of the European Chapter of the Association for Computational Linguistics (EACL).

As Oh joins the vibrant community at CDS, his work promises to bridge the gap between human language processing and machine learning, contributing to our understanding of both fields and potentially improving the design of future AI systems.

To view all our current faculty fellows, please visit the CDS Faculty Fellow page on our website.

By Stephen Thomas