研究者詳細 - 松原　茂樹

The Power of URLs in Scholarly Papers: Their Role as Metadata Sources for Data Repositories

Wada K., Matsubara S.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 15494 LNCS 巻頁： 85 - 92 2025年

記述言語：英語掲載種別：研究論文（学術雑誌）出版者・発行元：Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

Open science is increasingly important in academia. This paper investigates the citations using URLs, used to reference research resources in scholarly papers, and their potential as metadata sources for data repositories. A dataset of URL citations extracted from scholarly papers in natural language processing is analyzed. We first examine trends in locations and frequency of URL citations over time, then classify the types of research resources and their citation purposes. Furthermore, we assess the usability of URL citations for automatic metadata generation in data repositories. Specifically, we explore how many metadata records can be added and how much the metadata fields can be enriched. Our findings suggest that URL citations offer a promising approach for enhancing metadata quality and availability in data repositories.

DOI： 10.1007/978-981-96-0868-3_6

Capabilities and Challenges of LLMs in Metadata Extraction from Scholarly Papers

その他リンク： https://link.springer.com/chapter/10.1007/978-981-96-0868-3_6

Watanabe Y., Ito K., Matsubara S.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 15493 LNCS 巻頁： 280 - 287 2025年

Research data are cited in scholarly papers, and the construction and use of datasets are mentioned. The descriptions of research data in papers may be used as information for its metadata. In this paper, we focus on large language models (LLM), which have achieved high performance in various natural language processing tasks, and we investigate the ability of LLMs to extract metadata from papers. In the experiment, we analyzed LLMs’ metadata extraction capabilities quantitatively and qualitatively. The results demonstrate that while LLMs can extract metadata from papers extensively, the extraction accuracy is not necessarily high. We confirm that there are challenges in identifying the names of research data and linking information related to the research data.

DOI： 10.1007/978-981-96-0865-2_23

Estimation of Relevance Between Datasets for Enhancing Accessibility of Research Artifacts

その他リンク： https://dl.acm.org/doi/10.1007/978-981-96-0865-2_23

Ito K., Matsubara S.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 15494 LNCS 巻頁： 97 - 103 2025年

Open science is being promoted globally, and researchers are encouraged to publish and share research artifacts, e.g., datasets, tools, software, code, and scholarly papers. To accelerate open science, the accessibility of research artifacts is also important. Previous studies focused on the relevance between scholarly papers to improve their accessibility; however, the relevance between research artifacts other than scholarly papers has not been investigated extensively. In this study, we focus on datasets, and verify the feasibility of estimating the relevance between datasets. Access using the relevance between datasets may be useful when it is difficult to verbalize the characteristics of the desired dataset. We implemented a method to estimate the relevance between datasets that utilizes dataset metadata expressed as text, and the relevance is estimated by using a BERT-based model. Experimental results demonstrated the feasibility of estimating the relevance between datasets. Finally, we discuss the effect of utilizing metadata for each metadata field.

DOI： 10.1007/978-981-96-0868-3_8

その他リンク： https://link.springer.com/chapter/10.1007/978-981-96-0868-3_8

言語処理学会30周年記念事業実施報告 Open Access

井之上直也, 大内啓樹, 河原大輔, 黒橋禎夫, 小町守, 須藤克仁, 永田亮, 松原茂樹, 宮尾祐介

自然言語処理 32 巻 ( 1 ) 頁： 330 - 341 2025年

Negation Scope Conversion for Unifying Negation-Annotated Datasets Open Access

Yoshida Asahi, Kato Yoshihide, Matsubara Shigeki

Journal of Natural Language Processing 32 巻 ( 1 ) 頁： 300 - 329 2025年

記述言語：英語掲載種別：研究論文（学術雑誌）出版者・発行元：The Association for Natural Language Processing

Negation scope resolution is a technique that identifies the part of a sentence affected by the negation cue. The three major corpora used for it, the BioScope corpus, the SFU review corpus, and the Sherlock dataset, have different annotation schemes for negation scope. Due to the different annotations, it is difficult to use the three corpora together in the study of negation scope resolution. To address this issue by merging the corpora into a unified dataset based on a common annotation scheme, we propose a method for automatically converting the scopes of BioScope and SFU to those of Sherlock. We conducted an experiment to evaluate the accuracy of our method using a dataset obtained by manually annotating the negation scopes to a tiny portion of BioScope and SFU, verifying that our method can convert the scopes with high accuracy. In addition, we conducted another experiment to verify the effectiveness of our method from a pragmatic perspective, where we fine-tuned PLM-based negation scope resolution models using the unified dataset obtained by our method. The results demonstrated that the performances of the models increase when fine-tuned on the unified dataset, unlike the simply combined one, which supports the effectiveness of our method.

DOI： 10.5715/jnlp.32.300

Classification of URL Citations on Scholarly Papers Using Intermediate Task Training Open Access

WADA Kazuhiro, TSUNOKAKE Masaya, MATSUBARA Shigeki

IEICE Transactions on Information and Systems advpub 巻 ( 0 ) 2025年

記述言語：英語掲載種別：研究論文（学術雑誌）出版者・発行元：The Institute of Electronics, Information and Communication Engineers

Citations using URL (URL citations) that appear in scholarly papers can be used as an information source for the research resource search engines. In particular, the information about the types of cited resources and reasons for their citation is crucial to describe the resources and their relations in the search services. To obtain this information, previous studies proposed some methods for classifying URL citations. However, their methods trained the model using a simple fine-tuning strategy and exhibited insufficient performance. We propose a classification method using a novel intermediate task. Our method trains the model on our intermediate task of identifying whether sample pairs belong to the same class before being fine-tuned on the target task. In the experiment, our method outperformed previous methods using the simple fine-tuning strategy with higher macro F-scores for different model sizes and architectures. Our analysis results indicate that the model learns the class boundaries of the target task by training our intermediate task. Our intermediate task also demonstrated higher performance and computational efficiency than an alternative intermediate task using triplet loss. Finally, we applied our method to other text classification tasks and confirmed the effectiveness when a simple fine-tuning strategy does not stably work.

DOI： 10.1587/transinf.2024edp7149

Human Performance in Incremental Dependency Parsing: Dependency Structure Annotations and their Analyses 査読有り

Hiroki Unno, Tomohiro Ohno, Koichiro Ito, Shigeki Matsubara

Proceedings of the 37th Pacific Asia Conference on Language, Information and Computation (PACLIC-2024) 2024年12月

研究データをより見つけやすくするためのメタデータ変換と学術機関リポジトリへの登録：太陽地球系物理学分野における実践招待有り

能勢正仁, 新堀淳樹, 三好由純, 堀智昭, 大平司, 岡本麻衣子, 直江千寿子, 我喜屋累, 田中幸恵, 相良毅, 青木学聡, 松原茂樹, 高橋一郎, 林秀和, 山田一成, 南山泰之, 田中良昌, 阿部修司, 上野悟, 今城峻, 齊藤泰雄, 芦北卓也, 堀優子, 清水敏之, 岡村奈々子, 平野かおる

情報の科学と技術 74 巻 ( 11 ) 頁： 487 - 493 2024年11月

記述言語：日本語掲載種別：研究論文（学術雑誌）出版者・発行元：一般社団法人情報科学技術協会

「超高層大気長期変動の全球地上ネットワーク観測・研究（IUGONET）」プロジェクトでは，太陽地球系物理学における地上観測データのメタデータデータベースを構築している。このメタデータデータベースを，幅広い研究コミュニティや一般市民が利用できるように，メタデータスキーマの変換を行い，学術機関リポジトリに登録した。その結果，データベース間の定期的なハーベストにより，学術機関リポジトリデータベース（IRDB），データカタログ横断検索システム，Google Dataset Searchなどにも登録され，研究データの可視性や到達可能性が大きく高まることとなった。また，政策決定機関からの研究メタデータ整備要求に対して，効果的なメタデータマネジメントを実施できた。

DOI： 10.18919/jkg.74.11_487

Utilization of Text Data for Response Timing Detection in Attentive Listening 査読有り

Yu Watanabe, Koichiro Ito, Shigeki Matsubara

Proceedings of Interspeech-2024 頁： 3565 - 3569 2024年9月

On an Intermediate Task for Classifying URL Citations on Scholarly Papers 査読有り

Kazuhiro Wada, Masaya Tsunokake, Shigeki Matsubara

Proceeding of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING2024) 2024年5月

Negation Scope Conversion: Towards a Unified Negation-Annotated Dataset 査読有り

Asahi Yoshida, Yoshihide Kato, Shigeki Matsubara

Proceeding of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING2024) 2024年5月

語りの傾聴において不同意を示す応答の生成査読有り

伊藤滉一朗, 村田匡輝, 大野誠寛, 松原茂樹

自然言語処理 31 巻 ( 1 ) 頁： 212 - 249 2024年3月

担当区分：最終著者記述言語：日本語掲載種別：研究論文（学術雑誌）出版者・発行元：一般社団法人　言語処理学会

コミュニケーションロボットなどの会話エージェントが語りを聴く役割を担うことが期待されている．これらが聴き手として認められるには，語りを傾聴していることを語り手に伝達する機能を備える必要がある．このための明示的な手段は語りに応答することであり，傾聴を示す目的で語りに応答する発話，すなわち傾聴応答の表出が有力である．語りの傾聴では，語り手の発話を受容することが聴き手の基本的な応答方略となる．しかし，語りには，自虐や謙遜などの発話が含まれることがある．この場合，語り手の発話に同意しないことを示す応答，すなわち，不同意応答を確実に表出できることが求められる．本論文では，語りの傾聴において不同意応答を適切に生成することの実現性を示す．そのために，本研究ではまず，時間制約のない環境で語りデータに不同意応答のタイミングと表現をタグ付けする方式を定めた．作成したコーパスを用いて，不同意応答タイミングを網羅的に，不同意応答表現を安定的にタグ付けできることを検証する．続いて，事前学習済みの Transformer ベースのモデルに基づく，不同意応答タイミングの検出手法，及び，不同意応答表現への分類手法を実装し，実験により応答コーパスを用いた不同意応答生成の実現性を検証した．

DOI： 10.5715/jnlp.31.212

Re-Evaluating Syntax-Based Negation Scope Resolution* Open Access

Yoshida, A; Kato, Y; Matsubara, S

IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS E107.D 巻 ( 1 ) 頁： 165 - 168 2024年1月

記述言語：英語掲載種別：研究論文（学術雑誌）出版者・発行元：一般社団法人電子情報通信学会

Negation scope resolution is the process of detecting the negated part of a sentence. Unlike the syntax-based approach employed in previous researches, state-of-the-art methods performed better without the explicit use of syntactic structure. This work revisits the syntax-based approach and re-evaluates the effectiveness of syntactic structure in negation scope resolution. We replace the parser utilized in the prior works with state-of-the-art parsers and modify the syntax-based heuristic rules. The experimental results demonstrate that the simple modifications enhance the performance of the prior syntax-based method to the same level as state-of-the-art end-to-end neural-based methods.

DOI： 10.1587/transinf.2023edl8050

Web of Science

Enhancing Findability and Searchability of Research Data: Metadata Conversion and Registration in Institutional Repositories Open Access

Nosé M., Shinbori A., Miyoshi Y., Hori T., Ohira T., Hashiba J., Naoe C., Gakiya R., Okamoto M., Sagara T., Aoki T., Matsubara S., Takahashi I., Hayashi H., Yamada K., Minamiyama Y., Tanaka Y., Abe S., Ueno S., Imajo S., Saito Y., Ashikita T., Hori Y., Shimizu T., Okamura N., Hirano K., Bargatze L.

Data Science Journal 23 巻 2024年

記述言語：英語掲載種別：研究論文（学術雑誌）出版者・発行元：Data Science Journal

This paper outlines our practice to enhance the findability and searchability of research data through metadata conversion from the Space Physics Archive Search and Extract (SPASE) schema to a more generic schema, the Japan Consortium for Open Access Repository (JPCOAR) schema, and registration of converted metadata in institutional repositories. Traditionally, earth and space science research data have been organized using the SPASE schema. Although the SPASE schema is comprehensive, its usage has been restricted to highly professional databases, limiting broader accessibility and impeding cross-disciplinary research. We discuss a case study at Nagoya University where 284 metadata records were converted from SPASE to JPCOAR, and illustrate the process and benefits of this conversion. This approach significantly improved the visibility and usability of metadata across various platforms like the Institutional Repositories DataBase (IRDB), the Data Catalog Cross-Search System, and Google’s Dataset Search, extending access to a wider range of users beyond the highly professional scientific community. This approach also aligns with national policies in Japan on research data management and simplifies metadata handling for researchers. Future direction includes expanding this conversion and registration model to other universities and institutions by leveraging the ubiquity of the SPASE schema in the earth and space science fields. Our practice may be useful in other research fields. This initiative aims to improve the overall findability of research data, to foster cross-disciplinary collaboration, and to enhance the value of research data itself and of creators and managers of the data.

DOI： 10.5334/dsj-2024-040

Estimating Metadata of Research Artifacts to Enhance their Findability

Ito, K; Matsubara, S

2024 IEEE 20TH INTERNATIONAL CONFERENCE ON E-SCIENCE, E-SCIENCE 2024 頁： 1 - 2 2024年

記述言語：英語掲載種別：研究論文（学術雑誌）出版者・発行元：Proceedings - 2024 IEEE 20th International Conference on e-Science, e-Science 2024

To accelerate open science, it is encouraged to publish and share research artifacts, e.g., scholarly papers, datasets, and code, with their metadata. Although the metadata facilitates the discovery of research artifacts, the metadata is not always complete. In this study, we verify the feasibility of estimating metadata that has not yet been filled (unfilled metadata) using metadata that has been filled (filled-in metadata) for realizing automatic completion of metadata. We implement an estimation method and evaluate its performance through experiments. The method takes filled-in metadata expressed in text as input and estimates unfilled metadata using a BERT-based model. In our experiments, the filled-in metadata was set as the name of the research artifacts, and the unfilled metadata was set as their type. Experimental results showed a high estimation performance.

DOI： 10.1109/e-Science62913.2024.10678684

Web of Science