Tatsuki Kuribayashi (栗林樹生)
My Japanese name, 栗林樹生, contains several kanji related to trees (木). My parents run a flower shop and love gardening, and, true to my name, I have also become fascinated by trees—linguistic ones!
My research lies at the intersection of natural language processing and cognitive science. I study incremental sentence processing, discourse processing, and the reasoning processes underlying language use in humans and machines.
My research lies at the intersection of natural language processing and cognitive science. I study incremental sentence processing, discourse processing, and the reasoning processes underlying language use in humans and machines.
|
|
News
I lead the Processing and Acquisition of Language in Machines and Mind (PALM) 🌴 group at MBZUAI.
Publications
Latest Publications
- Ryo Yoshida, Shinnosuke Isono, Taiga Someya, Yohei Oseki, Tatsuki Kuribayashi.
"An Existence Proof for Neural Language Models That Can Explain Garden-Path Effects via Surprisal." ACL 2026, pp. 36556-36569, 2026/07.
[paper] - Ruoxi Ning, Yongpeng Zhu, Qingcheng Zeng, Tatsuki Kuribayashi, Freda Shi.
"On the Effect of Hyperparameters in Language Modeling for Computational Linguistics." ACL 2026, pp. 41863-41880, 2026/07.
[paper] - Tatsuki Kuribayashi, Alex Warstadt, Yohei Oseki, Ethan Gotlieb Wilcox.
"Dual Alignment Between Language Model Layers and Human Sentence Processing." ACL 2026, pp. 46207-46223, 2026/07.
[paper] - Saeed Almheiri, Bilal Elbouardi, et al., Tatsuki Kuribayashi, et al.
"Multilingual Idioms in Sentences and Conversations Across High-, Medium-, and Low-Resource Languages." ACL 2026, pp. 12363-12389, 2026/07.
[paper]
Selected Publications
- Tatsuki Kuribayashi, Alex Warstadt, Yohei Oseki, Ethan Gotlieb Wilcox.
"Dual Alignment Between Language Model Layers and Human Sentence Processing." ACL 2026, pp. 46207-46223, 2026/07. [paper] - Tianyang Xu, Tatsuki Kuribayashi, Yohei Oseki, Ryan Cotterell, Alex Warstadt.
"Can Language Models Learn Typologically Implausible Languages?"
Transactions of the Association for Computational Linguistics (TACL), 14: pp. 588-611, 2026. [paper] - Tatsuki Kuribayashi, Yohei Oseki, Souhaib Ben Taieb, Kentaro Inui, Timothy Baldwin.
"Large Language Models Are Human-Like Internally." TACL, 13: pp. 1743-1766, 2025/12. [paper] - *Nadine El-Naggar, *Tatsuki Kuribayashi, Ted Briscoe. (* equal contribution)
"Which Word Orders Facilitate Length Generalization in LMs? An Investigation with GCG-Based Artificial Languages." EMNLP 2025, pp. 35599-35613, 2025/11. [paper] - Tatsuki Kuribayashi, Ryo Ueda, Ryo Yoshida, Yohei Oseki, Ted Briscoe, Timothy Baldwin.
"Emergent Word Order Universals from Cognitively-Motivated Language Models." ACL 2024, pp. 14522-14543, 2024/08. [paper] - Tatsuki Kuribayashi, Yohei Oseki, Ana Brassard, Kentaro Inui.
"Context Limitations Make Neural Language Models More Human-Like." EMNLP 2022, pp. 10421-10436, 2022/12. [paper] - Tatsuki Kuribayashi, Yohei Oseki, Takumi Ito, Ryo Yoshida, Masayuki Asahara, Kentaro Inui.
"Lower Perplexity is Not Always Human-Like." ACL 2021, pp. 5203-5217, 2021/08. [paper] - *Takumi Ito, *Tatsuki Kuribayashi, *Masatoshi Hidaka, Jun Suzuki, Kentaro Inui. (* equal contribution)
"Langsmith: An Interactive Academic Text Revision System." EMNLP 2020 System Demonstrations, pp. 216-226, 2020/11. [paper]
Peer-Reviewed Publications
- Ryo Yoshida, Shinnosuke Isono, Taiga Someya, Yohei Oseki, Tatsuki Kuribayashi.
"An Existence Proof for Neural Language Models That Can Explain Garden-Path Effects via Surprisal."
In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026 main long), pp. 36556-36569, 2026/07.
[paper] -
Ruoxi Ning, Yongpeng Zhu, Qingcheng Zeng, Tatsuki Kuribayashi, Freda Shi.
"On the Effect of Hyperparameters in Language Modeling for Computational Linguistics."
In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026 main long), pp. 41863-41880, 2026/07.
[paper] - Tatsuki Kuribayashi, Alex Warstadt, Yohei Oseki, Ethan Gotlieb Wilcox.
"Dual Alignment Between Language Model Layers and Human Sentence Processing."
In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026 main long), pp. 46207-46223, 2026/07.
[paper] - Saeed Almheiri, Bilal Elbouardi, Salsabila Zahirah Pranida, Irina Nikishina, Ashwath Rao B, Parameswari Krishnamurthy, Muhammad Cendekia Airlangga, Rifo Ahmad Genadi, Nguyễn Phan Gia Bảo, Amir Hossein Yari, Hawau Olamide Toyin, Nurdaulet Mukhituly, Mena Attia, Besher Hassan, Ahmad Fathan Hidayatullah, Tatsuki Kuribayashi, Haonan Li, Suma Bhat, Fajri Koto
"Multilingual Idioms in Sentences and Conversations Across High-, Medium-, and Low-Resource Languages."
In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026 main long), pp. 12363-12389, 2026/07.
[paper] - Nadine El-Naggar, Tatsuki Kuribayashi, Ted Briscoe.
"What Kind of Language is Easy to Language-Model Under Curriculum Learning?"
In Proceedings of the 15th edition of the Workshop on Cognitive Modeling and Computational Linguistics (CMCL 2026), 2026/05.
[paper] - Tianyang Xu, Tatsuki Kuribayashi, Yohei Oseki, Ryan Cotterell, Alex Warstadt.
"Can Language Models Learn Typologically Implausible Languages?"
Transactions of the Association for Computational Linguistics (TACL), 14: pp. 588-611, 2026.
[paper] -
Keito Kudo, Yoichi Aoki, Tatsuki Kuribayashi, Shusaku Sone, Masaya Taniguchi, Ana Brassard, Keisuke Sakaguchi, Kentaro Inui.
"LLMs Faithfully and Iteratively Compute Answers During CoT: A Systematic Analysis With Multi-step Arithmetics."
Findings of 19th Conference of the European Chapter of the Association for Computational Linguistics (EACL 2026, Findings long), pp. 1114-1153, 2026/03.
[paper]- Keito Kudo, Yoichi Aoki, Tatsuki Kuribayashi, Shusaku Sone, Masaya Taniguchi, Ana Brassard, Keisuke Sakaguchi, Kentaro Inui.
"Think-to-Talk or Talk-to-Think? When LLMs Come Up with an Answer in Multi-Step Reasoning."
[arXiv]
- Keito Kudo, Yoichi Aoki, Tatsuki Kuribayashi, Shusaku Sone, Masaya Taniguchi, Ana Brassard, Keisuke Sakaguchi, Kentaro Inui.
- Mengyu Ye, Jun Suzuki, Tatsuro Inaba, Tatsuki Kuribayashi.
"Transformer Key-Value Memories Are Nearly as Interpretable as Sparse Autoencoders."
In Proceedings of the Thirty-Ninth Annual Conference on Neural Information Processing Systems (NeurIPS 2025), 2025/12.
[paper] - Tatsuki Kuribayashi, Yohei Oseki, Souhaib Ben Taieb, Kentaro Inui, Timothy Baldwin.
"Large Language Models Are Human-Like Internally."
Transactions of the Association for Computational Linguistics (TACL), 13: pp. 1743-1766, 2025/12.
[paper | code] - *Nadine El-Naggar, *Tatsuki Kuribayashi, Ted Briscoe.
"Which Word Orders Facilitate Length Generalization in LMs? An Investigation with GCG-Based Artificial Languages."
In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing (EMNLP 2025, main long), pp.35599-35613, 2025/11.
[paper] - Nadine El-Naggar, Tatsuki Kuribayashi, Ted Briscoe.
"GCG-Based Artificial Languages for Evaluating Inductive Biases of Neural Language Models."
In Proceedings of the Conference on Computational Natural Language Learning (CoNLL 2025), pp.540-556, 2025/08.
[paper] - Rena Wei Gao, Xuetong Wu, Tatsuki Kuribayashi, Mingrui Ye, Siya Qi, Carsten Roever, Yuanxing Liu, Zheng Yuan, Jey Han Lau.
"Can LLMs Simulate L2-English Dialogue? An Information-Theoretic Analysis of L1-Dependent Biases."
In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025, main long), pp. 4355-4379, 2025/08.
[paper] - Mengyu Ye, Tatsuki Kuribayashi, Goro Kobayashi, Jun Suzuki.
"Can Input Attributions Explain Inductive Reasoning in In-Context Learning?"
Findings of the 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025, Findings long), pp.21199-21225, 2025/08.
[paper] - Ryo Ueda, Tatsuki Kuribayashi, Shunsuke Kando, Kentaro Inui.
"Syntactic Learnability of Echo State Neural Language Models at Scale."
The 14th Workshop on Cognitive Modeling and Computational Linguistics (CMCL 2025, Non-archival), 2025/05.
[arXiv] - Haonan Li, Xudong Han, ..., Tatsuki Kuribayashi, ..., Eduard Hovy, Iryna Gurevych, Preslav Nakov, Monojit Choudhury, Timothy Baldwin.
"Libra-Leaderboard: Towards Responsible AI through a Balanced Leaderboard of Safety and Capability."
In Proceedings of 2025 Annual Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics (NAACL 2025, system demonstrations track), pp. 268-286, 2025/04.
[paper] - Tatsuki Kuribayashi, Timothy Baldwin.
"Does Vision Accelerate Hierarchical Generalization of Neural Language Learners?"
In Proceedings of the 31st International Conference on Computational Linguistics (COLING 2025, long), pp.1865-1879, 2025/01.
[paper] -
David Romero, Chenyang Lyu, Haryo Akbarianto Wibowo,...,Tatsuki Kuribayashi,...,Thamar Solorio, Alham Fikri Aji.
"CVQA: Culturally-diverse Multilingual Visual Question Answering Benchmark."
In Proceedings of the Neural Information Processing Systems Track on Datasets and Benchmarks (NeurIPS 2024 Datasets and Benchmarks Track), 2024/12.
[paper] -
Yoichi Aoki, Keito Kudo, Tatsuki Kuribayashi, Shusaku Sone, Masaya Taniguchi, Keisuke Sakaguchi, Kentaro Inui.
"First Heuristic Then Rational: Dynamic Use of Heuristics in Language Model Reasoning."
In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing (EMNLP 2024, main short), pp. 14255-14271, 2024/12.
[paper] - Tatsuki Kuribayashi, Ryo Ueda, Ryo Yoshida, Yohei Oseki, Ted Briscoe, Timothy Baldwin.
"Emergent Word Order Universals from Cognitively-Motivated Language Models."
In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024, main long), pp. 14522-14543, 2024/08. (acceptance rate: 940/4407=21.3%)
[paper] - Tatsuki Kuribayashi, Yohei Oseki, Timothy Baldwin.
"Psychometric Predictive Power of Large Language Models."
Findings of the 2024 Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL 2024, Findings long), pp.1983-2005, 2024/06. (acceptance rate: top 869/2434=35.7%)
[paper] -
大羽未悠, 栗林樹生, 大内啓樹, 渡辺太郎.
言語モデルの第二言語獲得.
自然言語処理 (Japan domestic journal), Volume 31, Number 2, pp. 433-455, 2024/06.
[paper] - Yukiko Ishizuki, Tatsuki Kuribayashi, Yuichiroh Matsubayashi, Ryohei Sasano and Kentaro Inui.
"To Drop or Not to Drop? Predicting Argument Ellipsis Judgments: A Case Study in Japanese."
In Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024, long), pp. 16198-16210, 2024/05. (acceptance rate: 1556/3417=52%)
[paper] - Goro Kobayashi, Tatsuki Kuribayashi, Sho Yokoi, Kentaro Inui.
"Analyzing Feed-Forward Blocks in Transformers through the Lens of Attention Maps."
In Proceedings of the 12th International Conference on Learning Representations (ICLR 2024, spotlight, top 5%), 2024/05. (acceptance rate: 2260/7262=31%)
[paper] - Mengyu Ye, Tatsuki Kuribayashi, Jun Suzuki, Hiroaki Funayama and Goro Kobayashi.
"Assessing Step-by-Step Reasoning against Lexical Negation: A Case Study on Syllogism."
In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing (EMNLP-2023, main short), pp. 14753-14773, 2023/12. (acceptance rate: 146/1041=14.0%)
[paper]- Mengyu Ye, Tatsuki Kuribayashi, Jun Suzuki, Hiroaki Funayama and Goro Kobayashi.
"Assessing Chain-of-Thought Reasoning against Lexical Negation: A Case Study on Syllogism."
In Proceedings of the Student Research Workshop (SRW) at the 61st Annual Meeting of the Association for Computational Linguistics 2023 (ACL-SRW, Non-archival), 2023/07. 👑 Best Paper Award
- Mengyu Ye, Tatsuki Kuribayashi, Jun Suzuki, Hiroaki Funayama and Goro Kobayashi.
-
Takumi Ito, Naomi Yamashita, Tatsuki Kuribayashi, Masatoshi Hidaka, Jun Suzuki, Ge Gao, Jack Jamieson and Kentaro Inui.
"Use of an AI-powered Rewriting Support Software in Context with Other Tools: A Study of Non-Native English Speakers."
The ACM Symposium on User Interface Software and Technology 2023 (UIST 2023), Article 45, pp. 1-13, 2023/10.
[paper] - Miyu Oba, Tatsuki Kuribayashi, Hiroki Ouchi, Taro Watanabe.
"Second Language Acquisition of Neural Language Models."
Findings of the 61st Annual Meeting of the Association for Computational Linguistics (ACL-2023, Findings long), pp. 13557-13572, 2023/07. (acceptance rate: top 39.1%)
[paper] - Goro Kobayashi, Tatsuki Kuribayashi, Sho Yokoi and Kentaro Inui.
"Transformer Language Models Handle Word Frequency in Prediction Head."
Findings of the 61st Annual Meeting of the Association for Computational Linguistics (ACL-2023, Findings short), pp. 4523-4535, 2023/07. (acceptance rate: top 39.1%)
[paper] - Keito Kudo, Yoichi Aoki, Tatsuki Kuribayashi, Ana Brassard, Masashi Yoshikawa, Keisuke Sakaguchi and Kentaro Inui.
"Do Deep Neural Networks Capture Compositionality in Arithmetic Reasoning?"
Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics (EACL-2023, main short), pp. 1351-1362, 2023/05. (acceptance rate: 281/1166=24.1%)
[paper] - Yoichi Aoki, Keito Kudo, Tatsuki Kuribayashi, Ana Brassard, Masashi Yoshikawa, Keisuke Sakaguchi and Kentaro Inui.
"Empirical Investigation of Neural Symbolic Reasoning Strategies."
Findings of the 17th Conference of the European Chapter of the Association for Computational Linguistics (EACL-2023, Findings short), pp. 1154-1162, 2023/05. (acceptance rate: top 482/1166=41.3%)
[paper] - Yoichi Aoki, Keito Kudo, Tatsuki Kuribayashi, Ana Brassard, Masashi Yoshikawa, Keisuke Sakaguchi and Kentaro Inui.
"Empirical Investigation of Neural Symbolic Reasoning Strategies."
Non-archival submission for the 2022 AACL-IJCNLP Student Research Workshop (AACL-IJCNLP SRW Non-archival), 2022/11. 👑 Best Paper Award - Tatsuki Kuribayashi, Yohei Oseki, Ana Brassard, Kentaro Inui.
"Context Limitations Make Neural Language Models More Human-Like."
In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing (EMNLP-2022, main long), pp.10421-10436, 2022/12. (acceptance rate: 829/4190=20%)
[paper] -
Riki Fujihara, Tatsuki Kuribayashi, Kaori Abe, Ryoko Tokuhisa, Kentaro Inui.
"Topicalization in Language Models: A Case Study on Japanese."
In Proceedings of the 29th International Conference on Computational Linguistics (COLING-2022, long), pp. 851-862, 2022/10 (acceptance rate: 522/1563=33.4%).
[paper] - Riki Fujihara, Tatsuki Kuribayashi, Kaori Abe, Kentaro Inui.
"Topicalization in Language Models: A Case Study on Japanese."
In Proceedings of the Student Research Workshop at the Joint Conference of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (ACL-IJCNLP-2021 SRW Non-archival), 2021/08 (acceptance rate: 39%).
- Goro Kobayashi, Tatsuki Kuribayashi, Sho Yokoi, Kentaro Inui.
"Incorporating Residual and Normalization Layers into Analysis of Masked Language Models."
In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing (EMNLP 2021, main long), pp. 4547-4568, 2021/11 (acceptance rate: 23.3%).
[paper | code] - Hiroki Ouchi, Jun Suzuki, Sosuke Kobayashi, Sho Yokoi, Tatsuki Kuribayashi, Masashi Yoshikawa, Kentaro Inui.
"Instance-Based Neural Dependency Parsing."
Transactions of the Association for Computational Linguistics 2021 (TACL 2021), 9: pp. 1493-1507, 2021/09.
[paper] - Tatsuki Kuribayashi, Yohei Oseki, Takumi Ito, Ryo Yoshida, Masayuki Asahara,
Kentaro Inui.
"Lower Perplexity is Not Always Human-Like."
In Proceedings of the Joint Conference of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (ACL-IJCNLP 2021, main long), pp. 5203-5217, 2021/08 (acceptance rate: 21.3%).
[paper | code] - 栗林樹生, 大内啓樹, 井之上直也, 鈴木潤, Paul Reisert, 三好利昇, 乾健太郎
「論述構造解析におけるスパン分散表現」
自然言語処理 (domestic journal), Volume 27, Number 4, pp. 753-780, 2020/12.
👑 Best Paper Award 2020, Association for Natural Language Processing
[paper | slides] - Takaki Otake, Sho Yokoi, Naoya Inoue, Ryo Takahashi, Tatsuki Kuribayashi, Kentaro
Inui.
"Modeling Event Salience in a Narrative Based on Barthes' Cardinal Function."
In Proceedings of the 28th International Conference on Computational Linguistics (COLING-2020, short), pp. 1784-1794, 2020/12 (acceptance rate: 26.2%).
[paper] - Goro Kobayashi, Tatsuki Kuribayashi, Sho Yokoi, Kentaro Inui.
"Attention is Not Only a Weight: Analyzing Transformers with Vector Norms."
In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP-2020, main long), pp. 7057-7075, 2020/11 (acceptance rate: 754/3359=22.4%).
[paper | code] - Goro Kobayashi, Tatsuki Kuribayashi, Sho Yokoi, Kentaro Inui.
"Self-Attention is Not Only a Weight: Analyzing BERT with Vector Norms."
In Proceedings of the Student Research Workshop at the 58th Annual Meeting of the Association for Computational Linguistics (ACL-SRW-2020 Non-archival), 2020/07 (acceptance rate: 72/202=35.6%).
- *Takumi Ito, *Tatsuki Kuribayashi, *Masatoshi Hidaka, Jun Suzuki, Kentaro Inui. (*
equal contribution)
"Langsmith: An Interactive Academic Text Revision System."
In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP-2020, demo), pp. 216-226, 2020/11 (acceptance rate: ???%).
[paper | demo] - Takumi Ito, Tatsuki Kuribayashi, Hayato Kobayashi, Ana Brassard, Masato Hagiwara, Jun
Suzuki, Kentaro Inui.
"Assisting Authors to Convert Raw Products into Polished Prose."
Journal of Cognitive Science, Vol.21, No.1, pp. 99-135, 2020.
[paper] - Tatsuki Kuribayashi, Takumi Ito, Jun Suzuki, Kentaro Inui.
"Language Models as an Alternative Evaluator of Word Order Hypotheses: A Case Study in Japanese."
In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL-2020, main long), pp. 6452-6459, 2020/07 (acceptance rate: 779/3429=22.7%).
[paper | data] - Hiroki Ouchi, Jun Suzuki, Sosuke Kobayashi, Sho Yokoi, Tatsuki Kuribayashi, Ryuto
Konno, Kentaro Inui.
"Instance-Based Learning of Span Representations: A Case Study through Named Entity Recognition."
In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL-2020, Short), pp. 6452-6459, 2020/07 (acceptance rate: 208/1185=17.6%).
[paper] - *Takumi Ito, *Tatsuki Kuribayashi, Hayato Kobayashi, Ana Brassard, Masato Hagiwara,
Jun Suzuki, Kentaro Inui. (* equal contribution)
"Diamonds in the Rough: Generating Fluent Sentences from Early-stage Drafts for Academic Writing Assistance."
In Proceedings of the 12th International Conference on Natural Language Generation (INLG-2019), pp. 40-53, 2019/10 (acceptance rate: 73/143=51.0%).
[paper | data] - Masato Hagiwara, Takumi Ito, Tatsuki Kuribayashi, Jun Suzuki, Kentaro Inui.
"TEASPN: Framework and Protocol for Integrated Writing Assistance Environments."
In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing (EMNLP-IJCNLP 2019, demo), pp. 229-234, 2019/11 (acceptance rate: ???%).
[paper | code | documentation | video] - Tatsuki Kuribayashi, Hiroki Ouchi, Naoya Inoue, Paul Reisert, Toshinori Miyoshi, Jun
Suzuki, Kentaro Inui.
"An Empirical Study of Span Representations in Argumentation Structure Parsing."
In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics (ACL-2019, short), pp. 4691-4698, 2019/07 (acceptance rate: 213/1163=18.2%).
[paper | code] - Paul Reisert, Naoya Inoue, Tatsuki Kuribayashi, Kentaro Inui.
"Feasible Annotation Scheme for Capturing Policy Argument Reasoning using Argument Templates."
In Proceedings of the 5th Workshop on Argument Mining, pp. 79-89, 2018/11 (acceptance rate: 18/32=56.3%).
[paper | data]
Preprints
-
Riku Kisako, Tatsuki Kuribayashi, Ryohei Sasano.
"On Representational Dissociation of Language and Arithmetic in Large Language Models."
[arXiv]
Domestic Conference Papers and Articles
- 吉田 遼, 磯野 真之介, 染谷 大河, 大関 洋平, 栗林 樹生. サプライザルによってガーデンパス効果を説明できる言語モデルは存在する. 言語処理学会第32回年次大会, pp. 451–456, 2026/3.
- 工藤 慧音, 青木 洋一, 栗林 樹生, 曾根 周作, 谷口 雅弥, Ana Brassard, 坂口 慶祐, 乾 健太郎. 多段算術推論タスクにおける思考の連鎖の忠実性. 言語処理学会第32回年次大会, pp. 2066–2071, 2026/3.
- 木迫 璃玖, 栗林 樹生, 笹野 遼平. 大規模言語モデル内における文法回路と計算回路の重なり. 言語処理学会第32回年次大会, pp. 2095–2100, 2026/3.
- 稲葉 達郎, 乾 健太郎, 栗林 樹生. 合成データによる事前学習から見る言語と音楽の関係性. 言語処理学会第32回年次大会, pp. 3444–3448, 2026/3.
- 栗林 樹生, 大関 洋平, Souhaib Ben Taieb, 乾 健太郎, Timothy Baldwin. 大規模言語モデルの浅い層が人間の速い言語処理を再現する. 言語処理学会第31回年次大会, 4pages, 2025/3.
- 上田 亮, 栗林 樹生, 神藤 駿介, 乾 健太郎. RNNの回帰行列を凍結しても統語構造の獲得は損なわれない. 言語処理学会第31回年次大会, 4pages, 2025/3.
- 木迫 璃玖, 栗林 樹生, 笹野 遼平. 大規模言語モデル内部における言語と計算領域の区分. 言語処理学会第31回年次大会, 4pages, 2025/3.
- 栗林樹生, 大関洋平, Timothy Baldwin. 大規模言語モデルの文処理は人間らしいのか？. 言語処理学会第30回年次大会,4pages,2024/3.
- 栗林樹生, 上田亮, 吉田遼, 大関洋平, Ted Briscoe, Timothy Baldwin. どのような言語モデルが不可能な言語を学習してしまうのか？---語順普遍を例に---. 言語処理学会第30回年次大会, 4pages, 2024/3. 👑 優秀賞
- 青木洋一, 工藤慧音, 曾根周作, 栗林樹生, 谷口雅弥, 坂口慶祐, 乾健太郎. 言語モデルの思考連鎖的推論における探索戦略の動的変化. 言語処理学会第30回年次大会, 4pages, 2024/3.
- 工藤慧音, 青木洋一, 栗林樹生, 谷口雅弥, 曾根周作, 坂口慶祐, 乾健太郎. 算術推論問題における自己回帰型言語モデルの内部機序. 言語処理学会第30回年次大会, 4pages, 2024/3.
- 葉夢宇, 栗林樹生, 小林悟郎, 鈴木潤. 文脈内学習における文脈内事例の寄与度推定. 言語処理学会第30回年次大会, 4pages, 2024/3.
- 都地悠馬, 高橋惇, 横井祥, 栗林樹生, 上田亮, 宮原英之. 長距離相互作用する文脈依存言語における相転移現象 -言語モデルの創発現象を統計力学の視点で理解する-. 言語処理学会第30回年次大会,4pages,2024/3. 👑 委員特別賞
- 栗林樹生. 百聞は一見に如かず？視覚情報は言語モデルに文の階層構造を教示するか. 言語処理学会第29回年次大会, 4pages, 2023/3. 👑 委員特別賞
- 小林悟郎, 栗林樹生, 横井祥, 乾健太郎. Transformer言語モデルの予測ヘッド内バイアスによる頻度補正効果. 言語処理学会第29回年次大会, 4pages, 2023/3.
- 葉夢宇, 栗林樹生, 舟山弘晃, 鈴木潤. 思考連鎖指示における大規模言語モデルの否定表現理解. 言語処理学会第29回年次大会, 4pages, 2023/3.
- 藤原吏生, 栗林樹生, 徳久良子, 乾健太郎. 主題化における人間と言語モデルの対照. 言語処理学会第29回年次大会, 4pages, 2023/3.
- 石月由紀子, 栗林樹生, 松林優一郎, 笹野遼平, 乾健太郎. 日本語話者の項省略判断に関するアノテーションとモデリング. 言語処理学会第29回年次大会, 4pages, 2023/3. 👑 委員特別賞
- 大羽未悠, 栗林樹生, 大内啓樹, 渡辺太郎. 言語モデルの第二言語獲得. 言語処理学会第29回年次大会, 4pages, 2023/3.
- 工藤慧音, 青木洋一, 栗林樹生, Ana Brassard, 吉川将司, 坂口慶祐, 乾健太郎. 算術問題におけるニューラルモデルの構成的推論能力. 言語処理学会第29回年次大会, 4pages, 2023/3.
- 青木洋一, 工藤慧音, Ana Brassard, 栗林樹生, 吉川将司, 坂口慶祐, 乾健太郎. ニューラル記号推論における推論過程の教示方法. 言語処理学会第29回年次大会, 4pages, 2023/3.
- 大羽未悠, 栗林樹生, 大内啓樹, 渡辺太郎. 言語モデルの第二言語獲得効率. 第254回自然言語処理研究会, 6pages, 2022/11. 👑 優秀研究賞
- 青木洋一, 工藤慧音, 栗林樹生, Ana Brassard, 吉川将司, 乾健太郎. 推論過程の性質がニューラルネットの多段推論能力に与える影響. 第17回NLP若手の会 シンポジウム (YANS), 2022/08.
- 栗林樹生. 視覚情報は言語モデルに人間らしい統語的汎化を促すか. 第17回NLP若手の会 シンポジウム (YANS), 2022/08. 👑 奨励賞
- 大羽未悠, 栗林樹生, 大内啓樹, 渡辺太郎. 言語モデルの第二言語獲得効率. 第17回NLP若手の会 シンポジウム (YANS), 2022/08.
- 小林悟郎, 栗林樹生, 横井祥, 乾健太郎. Transformerにおけるフィードフォワードネットの混ぜ合わせ作用. 第17回NLP若手の会 シンポジウム (YANS), 2022/08.
- 栗林樹生, 大関洋平, Ana Brassard, 乾健太郎. ニューラル言語モデルの過剰な作業記憶. 言語処理学会第28回年次大会, pp.1530-1535, 2022/03.
- 小林悟郎, 栗林樹生, 横井祥, 乾健太郎. Transformerにおけるフィードフォワードネットの作用. 言語処理学会第28回年次大会, pp.1072-1077, 2022/03. 👑 委員特別賞
- 石月由紀子, 栗林樹生, 松林優一郎, 大関洋平. 情報量に基づく日本語項省略の分析. 言語処理学会第28回年次大会, pp.442-447, 2022/03.
- 青木洋一, 工藤慧音, Ana Brassard, 栗林樹生, 吉川将司, 乾健太郎. 多段の数量推論タスクに対する適応的なモデルの振る舞いの検証. 言語処理学会第28回年次大会, pp.168-172, 2022/03.
- 石月由紀子, 栗林樹生, 松林優一郎, 大関洋平 情報量に基づく日本語項省略の分析. NLP若手の会 (YANS) 第16回シンポジウム, 2021/08.
- 小林悟郎, 栗林樹生, 横井祥, 乾健太郎. 非線形モジュールの加法分解に基づくTransformerレイヤーの解析. NLP若手の会 (YANS) 第16回シンポジウム, 2021/08.
- 栗林樹生, 大関洋平, 伊藤拓海, 吉田遼, 浅原正幸, 乾健太郎. 予測の正確な言語モデルがヒトらしいとは限らない. 言語処理学会第27回年次大会, pp.267-272, 2021/03. 👑 委員特別賞
- 栗林樹生, 大関洋平, 伊藤拓海, 吉田遼, 浅原正幸, 乾健太郎. 日本語の読みやすさに対する情報量に基づいた統一的な解釈. 言語処理学会第27回年次大会, pp.723-728, 2021/03.
- 伊藤拓海, 栗林樹生, 日高雅俊, 鈴木潤, 乾健太郎. Langsmith: 人とシステムの協働による論文執筆. 言語処理学会第27回年次大会, pp.1834-1839, 2021/03. [システム | ドキュメント | 動画]
- 藤原吏生, 栗林樹生, 乾健太郎. 人と言語モデルが捉える文の主題. 言語処理学会第27回年次大会, pp.1307-1312, 2021/03.
- 小林悟郎, 栗林樹生, 横井祥, 乾健太郎. Transformerの文脈を混ぜる作用と混ぜない作用. 言語処理学会第27回年次大会, pp.1224-1229, 2021/03.
- 大内啓樹, 鈴木潤, 小林颯介, 横井祥, 栗林樹生, 吉川将司, 乾健太郎. 事例ベース依存構造解析のための依存関係表現学習. 言語処理学会第27回年次大会, pp.497-502, 2021/03.
- 大竹孝樹, 横井祥, 井之上直也, 高橋諒, 栗林樹生, 乾健太郎. 物語におけるイベントの顕現性推定と物語類似性計算への応用. 言語処理学会第27回年次大会, pp.1324-1329, 2021/03.
- 藤原吏生, 栗林樹生, 乾健太郎. 日本語言語モデルが選択する文形式の傾向と文脈の影響 ー 主題化・有標語順について. NLP若手の会 (YANS) 第15回シンポジウム, 2020/09.
- 小林悟郎, 栗林樹生, 横井祥, 乾健太郎. ベクトル長に基づく注意機構と残差結合の包括的な分析. NLP若手の会 (YANS) 第15回シンポジウム, 2020/09.
- 栗林樹生, 伊藤拓海, 鈴木潤, 乾健太郎. 日本語語順分析に言語モデルを用いることの妥当性について. 言語処理学会第26回年次大会,
pp.493-496,
2020/03.
[slides] - 小林 悟郎, 栗林樹生, 横井 祥, 鈴木潤, 乾健太郎. ベクトル⻑に基づく自己注意機構の解析. 言語処理学会第26回年次大会, pp.965-968, 2020/03. 👑 最優秀賞
- 大内 啓樹, 鈴木 潤, 小林 颯介, 横井 祥, 栗林樹生, 乾健太郎. スパン間の類似性に基づく事例ベース構造予測. 言語処理学会第26回年次大会, pp.331-334, 2020/03.
- 大竹 孝樹, 横井 祥, 井之上 直也, 高橋 諒, 栗林樹生, 乾健太郎. 言語モデルによる物語中のイベントの顕現性推定. 言語処理学会第26回年次大会, pp.1089-1092, 2020/03.
- 伊藤拓海, 栗林樹生, 萩原正人, 鈴木潤, 乾健太郎. 英語論文執筆のための統合ライティング支援環境. 第14回NLP若手の会 シンポジウム (YANS), 2019/08.
- 小林悟郎, 栗林樹生, 横井祥, 鈴木潤, 乾健太郎. 文脈を考慮する言語モデルが捉える品詞情報とその軌跡. 第14回NLP若手の会 シンポジウム (YANS), 2019/08.
- 栗林樹生, 大内啓樹, 井之直也, Paul Reisert, 三好利昇, 鈴木潤, 乾健太郎. 複数の言語単位に対するスパン表現を用いた論述構造解析. 言語処理学会第25回年次大会, pp.990-993, 2019/03.
- *栗林樹生, *伊藤拓海, 内山香, 鈴木潤, 乾健太郎. (* 第一著者と第二著者の貢献度は等しい．) 言語モデルを用いた日本語の語順評価と基本語順の分析. 言語処理学会第25回年次大会, pp.1053-1056, 2019/03. 👑 若手奨励賞
- *伊藤拓海, *栗林樹生, 小林隼人, 鈴木潤, 乾健太郎. (* 第一著者と第二著者の貢献度は等しい．) ライティング支援を想定した情報補完型生成. 言語処理学会第25回年次大会, pp.970-973, 2019/03.
- Tatsuki Kuribayashi, Paul Reisert, Naoya Inoue and Kentaro Inui. Towards Exploiting Argumentative Context for Argumentative Relation Identification. 言語処理学会第24回年次大会, pp.284-287, 2018/03.
- Tatsuki Kuribayashi, Paul Reisert, Naoya Inoue, Kentaro Inui. Examining Macro-level Argumentative Structure Features for Argumentative Relation Identification. 第4回自然言語処理シンポジウム・第234回自然言語処理研究会, 6pages, 2017/12.
Education and Appointments
- 2025/08–Present: Tenure-Track Assistant Professor of Natural Language Processing, Division of Computing and Mathematical Sciences, MBZUAI, UAE. [website]
- 2025/10–Present: Visiting Researcher with Prof. Yusuke Miyao's team, Graduate School of Information Science and Technology, The University of Tokyo. [website]
- 2025/02–Present: Visiting Researcher, Center for Language AI Research, Tohoku University. [website]
- 2024/04–Present: Visiting Scientist, Natural Language Understanding Team, RIKEN.
- 2023/04–2025/07: Postdoctoral Associate, MBZUAI (Advisor: Prof. Timothy Baldwin).
- 2022/04–2023/03: Specially Appointed Research Fellow, Tohoku NLP Group.
- 2018/05–2023/03: Co-Founder, Langsmith Co., Ltd.
- 2020/04–2022/03: Ph.D., Graduate School of Information Sciences, Tohoku University, Japan (Supervisor: Prof. Kentaro Inui) [thesis]
- Doctoral Dissertation Recommended by the Information Processing Society of Japan. [link]
- President's Award, Graduate School of Information Sciences, Tohoku University. (東北大学総長賞)
- Excellent Student in Electrical and Information Science Award, Tohoku University. (東北大学電気・情報系優秀学生賞)
- 2018/04–2020/03: M.E., Graduate School of Information Sciences, Tohoku University, Japan (Supervisor: Prof. Kentaro Inui) [thesis]
- Excellent Student in Electrical and Information Science Award, Tohoku University. (東北大学電気・情報系優秀学生賞) [link]
- 2014/04–2018/03: B.E., Department of Information and Intelligent Systems, Tohoku University, Japan (Supervisor: Prof. Kentaro Inui)
- Excellent Academic Award, School of Engineering, Tohoku University. (東北大学工学部長賞; top 2.9%) [link]
Grants
- 2026/04–2029/03: Co-Investigator (研究分担者), JSPS Grant-in-Aid for Scientific Research (A).
「言語知能における永続的逐次適応機能の基盤理論構築と概念実証」(Principal Investigator: Jun Suzuki)
Building a Foundational Theory and Proof-of-Concept for Persistent Sequential Adaptation in Language Intelligence. - 2025/04–2028/03 (resumed following a suspension beginning in 2023/04): Principal Investigator, JSPS Grant-in-Aid for Early-Career Scientists.
「人間らしい言語獲得効率に対するマルチモーダル言語処理を通した構成論的探求」
Constructive Exploration of Human-Like Language Acquisition Through Multimodal Language Processing. - 2020/04–2022/03: JSPS Research Fellowship for Young Scientists (DC1) (interview waived; 54/277 = 19.5%).
「テクストの数理的モデリングと、数理モデルを通したテクストらしさの解明への挑戦」
Understanding Textuality Through Mathematical Modeling of Text.
Service
Teaching
- 2026: NLP701, NLP805 (Natural Language Processing) at MBZUAI
- 2025, 2026: NLP8504 (Cognitive Science through the Lens of Language Models) at MBZUAI
Reviewing
- Reviewer, Area Chair, and Senior Area Chair for ACL Rolling Review (primarily in Interpretability and Analysis of Models for NLP; Linguistic Theories; Cognitive Modeling; and Psycholinguistics)
- Reviewer for TACL
- Occasional reviewer for Computational Linguistics
- COLM: 2024
- COLING: 2020, 2022, 2024, 2025 (Language Modeling; Integrated Systems and Applications)
- LREC: 2022, 2024, 2026
- INLG: 2020, 2021
- CoNLL: 2024, 2025, 2026
- BlackboxNLP: 2024
Organizing
- Organizer, 15th Workshop on Cognitive Modeling and Computational Linguistics (CMCL 2026)
- Lead Organizer, 14th Workshop on Cognitive Modeling and Computational Linguistics (CMCL 2025)
- Lead Organizer, 13th Workshop on Cognitive Modeling and Computational Linguistics (CMCL 2024)
査読・編集委員
- 2025/09-: 会誌「自然言語処理」編集委員.
- 2020-: 言語処理学会年次大会, 学会誌自然言語処理.
委員活動等
- 2026: 言語処理学会年次大会委員
- 2020–2023: 最先端NLP勉強会 (2020–2022: 実行委員、2023: 実行委員長)
Research Visits, Talks, and Writings
- 2025/12: Research talk at LeM🍋N Lab, UCSD [slides]
- 2025/03: Research talk at MBZUAI [slides]
- 2024/05: Research talk at ETH Zürich [slides]
- 2024/05: Online research talk at The Hong Kong Polytechnic University [slides]
- 2022/12: Research talk at MBZUAI. "Cognitive Plausibility of Neual Language Models." [slides]
- 2025/09: Invited talk at the Evidence-based Linguistics Workshop 2025, NINJAL, Japan. "言語モデルは言語学に何のエビデンスも与えないのでしょうか？" [資料]
- 2024/12: 栗林樹生. 回想記事：Emergent Word Order Universals from Cognitively-Motivated Language Models. [link]
- 2024/12: 栗林樹生, 大関洋平. 開催報告：The 13th Edition of the Workshop on Cognitive Modeling and Computational Linguistics (CMCL 2024). [link]
- 2023/08: 最先端NLP勉強会. [slides] [slides]
- 2023/02: 自然言語処理による論文執筆支援. 東北大学（大学院改革推進センター主催）. [slides]
- 2022/09: 最先端NLP勉強会. [slides]
- 2021/09: 最先端NLP勉強会. [slides]
- 2021/06: 栗林樹生. Lower Perplexity is Not Always Human-Like (Talk). NLPコロキウム. [動画]
- 2021/06: 栗林樹生. 学会記事「論述構造解析におけるスパン分散表現」の研究を通して. 自然言語処理 (学会記事No.27), Volume 28, Number 2, pp.677-681.
- 2020/09: 栗林樹生. 学会記事 Language Models as an Alternative Evaluator of Word Order Hypotheses: A Case Study in Japanese. 自然言語処理 (学会記事), Volume 27, Number 3, pp.671-676.
- 2020/09: 最先端NLP勉強会. [slides]