Qin Dai

/ 代　勤

Researcher in Natural Language Processing and Large Language Models

Stockmark Inc. | Tohoku NLP Group

Google Scholar
Email

About Me

I am a researcher in natural language processing and large language models. My research focuses on knowledge acquisition and representation, particularly scientific knowledge acquisition, as well as understanding how language models represent and organize structured information such as entities, relations, and relational bindings. I am also interested in LLM interpretability, personalized language models, and recommender systems.

More broadly, I aim to understand the internal mechanisms of large language models and use these insights to build more effective and reliable NLP systems.

Research Interests

• Large Language Models and Interpretability: understanding representations, relational binding, reasoning, grounding, and internal mechanisms
• Knowledge Acquisition and Representation: scientific knowledge acquisition, knowledge discovery, and structured knowledge representation
• Knowledge-based Reasoning: reasoning over text, knowledge bases, and universal graphs
• Personalized Language Models and Recommender Systems: user modeling, personalization, news recommendation, and serendipity

International Conference Papers

• Qin Dai, Benjamin Heinzerling and Kentaro Inui. Cell-Based Representation of Relational Binding in Language Models.
  In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (ACL2026) (acceptance rate: 18.9%) [paper][arxiv].
• Qin Dai, Benjamin Heinzerling and Kentaro Inui. Representational Analysis of Binding in Language Models.
  In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing (EMNLP2024).(acceptance rate: 20.8%) [paper] [arxiv]
• Qin Dai, Benjamin Heinzerling and Kentaro Inui. Cross-stitching Text and Knowledge Graph Encoders for Distantly Supervised Relation Extraction.
  In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing (EMNLP2022).(acceptance rate: 20.0%) [paper][arxiv]
• Qin Dai, Naoya Inoue, Ryo Takahashi and Kentaro Inui. Two Training Strategies for Improving Relation Extraction over Universal Graph.
  In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics (EACL 2021), virtual. (acceptance rate: 24.7%) [paper][arxiv]
• Qin Dai, Naoya Inoue, Ryo Takahashi and Kentaro Inui. Enhancing Biomedical Relation Extraction with Indirect Evidences.
  In Proceedings of Fourth International Workshop on SCIentific DOCument Analysis (SCIDOCA), November 2020. [Link]
• Qin Dai, Naoya Inoue, Paul Reisert, Takahashi Ryo, and Kentaro Inui. Incorporating chains of reasoningover knowledge graph for distantly supervised biomedical knowledge acquisition.
  In Proceedings of the 33rd Pacific Asia Conference on Language, Information and Computation (PACLIC33), September 2019. (acceptance rate: 28.0%) [paper]
• Qin Dai, Naoya Inoue, Paul Reisert, Ryo Takahashi and Kentaro Inui. Distantly Supervised Biomedical Knowledge Acquisition via Knowledge Graph Based Attention.
  In Proceedings of First Workshop on Extracting Structured Knowledge from Scientific Publications (ESSP), June 2019. (acceptance rate: 33.3%)[paper]
• Qin Dai, Naoya Inoue, Paul Reisert and Kentaro Inui. Scientific Knowledge Acquisition via the Interaction between Relation Extraction and Knowledge Graph Completion.
  In Proceedings of Third International Workshop on SCIentific DOCument Analysis (SCIDOCA), November 2018. [Link]
• Qin Dai, Naoya Inoue, Paul Reisert and Kentaro Inui. Improving Scientific Relation Classification with Task Specific Supersense.
  In Proceedings of the 32nd Pacific Asia Conference on Language, Information and Computing (PACLIC32), December 2018. (acceptance rate: 32.2%) [paper]
• Qin Dai, Naoya Inoue, Paul Reisert, and Kentaro Inui. Leveraging Document-specific Information for Identifying Relations in Scientific Articles.
  In Proceedings of Second International Workshop on SCIentific DOCument Analysis (SCIDOCA), November 2017. [paper][Link]

Journals

• Qin Dai, Naoya Inoue, Paul Reisert, and Kentaro Inui. Leveraging Unannotated Texts for Scientific Relation Extraction.
  IEICE Transactions on Information and Systems, Vol. E101-D, No. 12, pp.3209-3217, December 2018. [paper]
• Qin Dai, Benjamin Heinzerling, Naoya Inoue and Kentaro Inui. Universal Graph based Distantly Supervised Relation Extraction.
  Journal of Natural Language Processing, Vol. 29, No. 4, pp.1138-1164, December 2022. [paper]

Domestic Conference Papers

• Qin Dai, Benjamin Heinzerling and Kentaro Inui. Cell-Based Representation of Relational Binding in Language Models.
  言語処理学会第32回年次大会, March 2026.
• 朝倉卓人, 広田 航, 乾健太郎, Benjamin Heinzerling, Qin Dai, 有馬幸介. ニーズ知識グラフ構築による異業種アナロジーの探索.
  言語処理学会第32回年次大会, March 2026.
• 高橋 洸丞, 近江 崇宏, 有馬 幸介, Benjamin Heinzerling, Qin Dai, 乾 健太郎. 継続事前学習によるLLMの知識獲得.
  言語処理学会第31回年次大会, March 2025.
• 広田 航, 高橋 洸丞, Qin Dai, Benjamin Heinzerling, 近江 崇宏, 乾 健太郎. Anchoring を行う生成的関係抽出.
  言語処理学会第31回年次大会, March 2025.
• Qin Dai, Benjamin Heinzerling and Kentaro Inui. Representational Analysis of Binding in Language Models.
  言語処理学会第31回年次大会, March 2025.
• Qin Dai, Benjamin Heinzerling and Kentaro Inui. Cross-stitching Text and Knowledge Graph Encoders for Distantly Supervised Relation Extraction.
  言語処理学会第29回年次大会, March 2023.
• Qin Dai, Benjamin Heinzerling and Kentaro Inui. Universal Graph based Relation Extraction.
  言語処理学会第28回年次大会, March 2022.
• Qin Dai, Naoya Inoue and Kentaro Inui. Improving Distantly Supervised Relation Extraction via Textual Representation of Multi-hop Inference over Text.
  言語処理学会第26回年次大会, March 2020.
• Qin Dai, Naoya Inoue, Paul Reisert and Kentaro Inui. End-to-End Scientific Knowledge Graph Completion via Word Embedding based Entity Type Classification.
  言語処理学会第25回年次大会, March 2019.

Experience

• Researcher, Stockmark Inc.
  Aug. 2026 – Present
• Researcher, Tohoku NLP Group, Tohoku University
  Apr. 2020 – Present

Education

Graduate School of Information Sciences, Tohoku University, Japan
PhD in Information Sciences (Mar. 2020)
Supervisor: Prof. Kentaro Inui

Service

Reviewer for top conferences in natural language processing.

• ACL: 2026, 2023
• EMNLP: 2024, 2023, 2021
• COLING: 2025, 2024, 2022, 2020
• COLM: 2026, 2024
• LREC: 2026, 2025, 2024
• NeurIPS: 2026

Reviewer for Journal of Natural Language Processing (JNLP): 2025, 2022

Awards

• Outstanding Paper Award at the 29th Annual Meeting of the Association of Natural Language Processing (NLP 2023)(11 out of 579 submissions)[Link]

Computer skills

• Languages: Python, C, JavaScript (basic)
• Machine Learning: PyTorch, TensorFlow, Keras, Theano

Hobbies

• Soccer, Jogging, Photography(青葉工業会 令和４年度写真コンテストで入選)