Tatsuya Aoyama (青山達也)

Ph.D. Candidate, Georgetown University

ta571 [AT] georgetown.edu

Bio

I am currently a 4th year Ph.D. candidate in Computational Linguistics at Georgetown University, where I belong to GUCL, NERT, and Corpling. Broadly, my research interests include topics in computational psycholinguistics. In particular, I am interested in the following areas: (1) language acquisition of LMs, (2) transferability/inductive bias, (3) methods for probing linguistic representation/knowledge in LMs. For example, I recently worked on comparing native and non-native Englishes in an embedding space using BERT, and on applying a behavioral approach to a layer-wise probing study of BERT. I find it very exciting to apply these lines of work to model human language acquisition.

I am also interested in linguistic formalisms that capture linguistic phenomena at various levels. For example, I am working on developing a novel approach to Combinatory Categorial Grammar (CCG) supertagging with Nathan Schneider, and on expanding rhetorical structure theory (RST) by creating and examining an anchored RST corpus with Amir Zeldes.

Prior to joining Georgetown, I obtained a B.A. in Liberal Arts from Soka University of America, where my main focus was Psychology and Economics (although I wrote my thesis on a topic related to Second Language Acquisition!). Subsequently, I pursued an M.A. in Teaching English to Speakers of Other Languages (TESOL) at Michigan State University, where I was co-advised by Sandra Deshors and Kristen Johnson. My M.A. thesis focused on quantifying how Chinese and Japanese speakers' use of the English articles deviates from that of native speakers, leveraging a modeling technique called Multifactorial Prediction and Deviation Analysis with Regressions (MuPDAR).

Publications

Most recent publications on Google Scholar.

  • All
  • NLP
  • SLA
2024

Identifying Fairness Issues in Automatically Generated Testing Content. PDF

Kevin Stowe, Benny Longwill, Alyssa Francis, Tatsuya Aoyama, Debanjan Ghosh, Swapna Somasundaran

The 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA), the North American Chapter of the Association for Computational Linguistics (NAACL) 2024.

eRST: A Signaled Graph Theory of Discourse Relations and Organization. PDF

Amir Zeldes, Tatsuya Aoyama, Janet Yang Liu, Siyao Peng, Debopam Das, and Luke Gessler

arxiv preprint

J-SNACS: Adposition and Case Supersenses for Japanese Joshi. Code

Tatsuya Aoyama, Chihiro Taguchi, and Nathan Schneider

The 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)

DISRPT: A Multilingual, Multi-domain, Cross-framework Benchmark for Discourse Processing.

Chloé Braud, Amir Zeldes, Laura Rivière, Yang Janet Liu, Philippe Muller, Damien Sileo, and Tatsuya Aoyama

The 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)

2023

What’s Hard in RST Parsing? Predictive Models for Error Analysis. PDF Code

Yang Janet Liu, Tatsuya Aoyama, and Amir Zeldes

Proceedings of the 24th Annual Meeting of the Special Interest Group on Discourse and Dialogue (SIGDIAL), Prague, Czechia. Association for Computational Linguistics.

GENTLE: A Genre-Diverse Multilayer Challenge Set for English NLP and Linguistic Evaluation. PDF Code

Tatsuya Aoyama, Shabnam Behzad, Luke Gessler, Lauren Levine, Jessica Lin, Yang Janet Liu, Siyao Peng, Yilun Zhu, and Amir Zeldes

The 17th Linguistic Annotation Workshop (LAW-XVII), Association for Computational Linguistics (ACL) 2023

Corpus-Based Investigation of the Markedness and Frequency of Japanese Passives in Contemporary Written Japanese. PDF

Tatsuya Aoyama

Society for Computation in Linguistics (SCiL) 2023

2022

Comparing Native and Learner Englishes Using a Large Pre-trained Language Model. PDF

Tatsuya Aoyama

11th Natural Language Processing for Computer-Assisted Language Learning (NLP4CALL) workshop 2022

Probe-Less Probing of BERT’s Layer-Wise Linguistic Knowledge with Masked Word Prediction. PDF

Tatsuya Aoyama, Nathan Schneider

North American Chapter of the Association for Computational Linguistics Student Research Workshop (NAACL-SRW 2022)

Revisiting Layer-Wise Linguistic Knowledge with Masked Word Prediction.

Tatsuya Aoyama, Nathan Schneider

Mid-Atlantic Student Colloquium on Speech, Language and Learning (MASC-SLL 2022)

2020

International Students’ Willingness to Communicate in English as a Second Language : Effects of L2 Self-Confidence, Acculturation, and Motivational Types PDF

Tatsuya Aoyama, Tomoko Takahashi

Journal of International Students 10(3)

A Corpus-based Multifactorial Analysis of Japanese and Chinese Speakers’ English Article Use : Quantifying the Deviation Using MuPDAR PDF

Tatsuya Aoyama

Unpublished M.A. Thesis

2024

Identifying Fairness Issues in Automatically Generated Testing Content. PDF

Kevin Stowe, Benny Longwill, Alyssa Francis, Tatsuya Aoyama, Debanjan Ghosh, Swapna Somasundaran

The 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA), the North American Chapter of the Association for Computational Linguistics (NAACL) 2024.

eRST: A Signaled Graph Theory of Discourse Relations and Organization. PDF

Amir Zeldes, Tatsuya Aoyama, Janet Yang Liu, Siyao Peng, Debopam Das, and Luke Gessler

arxiv preprint

J-SNACS: Adposition and Case Supersenses for Japanese Joshi. Code

Tatsuya Aoyama, Chihiro Taguchi, and Nathan Schneider

The 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)

DISRPT: A Multilingual, Multi-domain, Cross-framework Benchmark for Discourse Processing.

Chloé Braud, Amir Zeldes, Laura Rivière, Yang Janet Liu, Philippe Muller, Damien Sileo, and Tatsuya Aoyama

The 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)

2023

What’s Hard in RST Parsing? Predictive Models for Error Analysis. PDF Code

Yang Janet Liu, Tatsuya Aoyama, and Amir Zeldes

Proceedings of the 24th Annual Meeting of the Special Interest Group on Discourse and Dialogue (SIGDIAL), Prague, Czechia. Association for Computational Linguistics.

GENTLE: A Genre-Diverse Multilayer Challenge Set for English NLP and Linguistic Evaluation. PDF Code

Tatsuya Aoyama, Shabnam Behzad, Luke Gessler, Lauren Levine, Jessica Lin, Yang Janet Liu, Siyao Peng, Yilun Zhu, and Amir Zeldes

The 17th Linguistic Annotation Workshop (LAW-XVII), Association for Computational Linguistics (ACL) 2023

Corpus-Based Investigation of the Markedness and Frequency of Japanese Passives in Contemporary Written Japanese. PDF

Tatsuya Aoyama

Society for Computation in Linguistics (SCiL) 2023

2022

Comparing Native and Learner Englishes Using a Large Pre-trained Language Model. PDF

Tatsuya Aoyama

11th Natural Language Processing for Computer-Assisted Language Learning (NLP4CALL) workshop 2022

Probe-Less Probing of BERT’s Layer-Wise Linguistic Knowledge with Masked Word Prediction. PDF

Tatsuya Aoyama, Nathan Schneider

North American Chapter of the Association for Computational Linguistics Student Research Workshop (NAACL-SRW 2022)

Revisiting Layer-Wise Linguistic Knowledge with Masked Word Prediction.

Tatsuya Aoyama, Nathan Schneider

Mid-Atlantic Student Colloquium on Speech, Language and Learning (MASC-SLL 2022)

2020

A Corpus-based Multifactorial Analysis of Japanese and Chinese Speakers’ English Article Use : Quantifying the Deviation Using MuPDAR PDF

Tatsuya Aoyama

Unpublished M.A. Thesis

2022

Comparing Native and Learner Englishes Using a Large Pre-trained Language Model. PDF

Tatsuya Aoyama

11th Natural Language Processing for Computer-Assisted Language Learning (NLP4CALL) workshop 2022

2020

International Students’ Willingness to Communicate in English as a Second Language : Effects of L2 Self-Confidence, Acculturation, and Motivational Types PDF

Tatsuya Aoyama, Tomoko Takahashi

Journal of International Students 10(3)

A Corpus-based Multifactorial Analysis of Japanese and Chinese Speakers’ English Article Use : Quantifying the Deviation Using MuPDAR PDF

Tatsuya Aoyama

Unpublished M.A. Thesis

Experience

  • JPMorgan Chase & Co. Summer 2024
    Summer Associate, Data Science & AI Associate Program
    Machine Learning Center of Excellence (MLCOE)
  • ETS, AI Lab Summer 2023
    NLP Research Intern
  • Posh Technologies Summer 2022
    NLP Research Intern
  • Money Forward, Inc. Summer 2021
    NLP Engineer Intern
  • Georgetown University 2020 - now (exp. 2025)
    Ph.D. in Computational Linguistics
    Advised by Nathan Schneider and Amir Zeldes
  • Center for Applied Linguistics Summer 2019
    Research Intern
  • Michigan State University 2018 - 2020
    M.A. in Teaching English to Speakers of Other Languages (TESOL)
    Advised by Sandra Deshors and Kristen Johnson
  • Soka University of America 2014 - 2018
    B.A. in Liberal Arts
    Advised by Tomoko Takahashi

Talks

Second Language Research Forum (SLRF 2020) (October 2020): A Corpus-based Multifactorial Analysis of Japanese and Chinese Learners’ English Article Use : Quantifying theDeviation using MuPDAR

The Asian Conference on Language (ACL 2020) (March 2020): Japanese ESL Students’ Willingness to Communicate in English : The Effects of L2 Self-Confidence, Acculturation, and Motivational Types

MSU LLT 860 Guest Lecture (February 2020): Universal Grammar and Second Language Acquisition

Teaching

JPN-102 (Spring 2019): Elementary Japanese II

JPN-101 ( Fall 2018, Fall 2019): Elementary Japanese I

Acknowledgement

This website uses the website design and template by Martin Saveski