研究者詳細 - 駒水　孝裕

論文 - 駒水　孝裕

分割表示 100 件中 61 - 80 件目／全件表示 >>

Exploring Relevant Parts between Legal Documents using Substructure Matching 査読有り

Takahiro Komamizu, Kazuya Fujioka, Yasuhiro Ogawa, Katsuhiko Toyama

the Thirteenth International Workshop on Juris-informatics (JURISIN 2019) 頁： 16-28 2019年11月

言い換えによる自然言語－SPARQL対訳コーパスの拡張

李偉嘉, 小川泰弘, 駒水孝裕, 外山勝彦

第17回情報学ワークショップ論文集頁： - 2019年11月

Analyzing Japanese Law History through Modeling Multi-versioned Entity 査読有り

Takahiro Komamizu, Yushi Uchida, Yasuhiro Ogawa, Katsuhiko Toyama

the 2nd International Workshop on Contextualized Knowledge Graphs (CKG 2019) 頁： - 2019年10月

利用規約中の不公平文の自動検出

青山恵子, 小川泰弘, 駒水孝裕, 外山勝彦

第15回テキストアナリティクス・シンポジウム NLC2019-8(2019-9) 頁： 1-6 2019年9月

Thai Legal Term Correction using Random Forests with Outside-the-sentence Features 査読有り

Takahiro Yamakoshi, Vee Satayamas, Hutchatai Chanlekha, Yasuhiro Ogawa, Takahiro Komamizu, Asanee Kawtrakul, Katsuhiko Toyama

the 33rd Pacific Asia Conference on Language, Information and Computation (PACLIC 33) 頁： 161-170 2019年9月

弱分類器の調整に基づく不均衡データ向けアンサンブル・フレームワーク査読有り

植原リサ, 駒水孝裕, 小川泰弘, 外山勝彦

第12回Webとデータベースに関するフォーラム頁： 81-84 2019年9月

共通BERT分類器による紛らわしい法令用語の校正

山腰貴大, 駒水孝裕, 小川泰弘, 外山勝彦

言語処理学会NLP若手の会第14回シンポジウム頁： - 2019年8月

nagoy Team's Summarization System at the NTCIR-14 QA-Lab PoliInfo 査読有り

Yasuhiro Ogawa, Michiaki Satou, Takahiro Komamizu, Katsuhiko Toyama

the Fourteenth NTCIR conference (NTCIR-14) , Revised Selected Papers 頁： 110-121 2019年6月

法律の要約のためのランダムフォレストを用いた重要文抽出

小川泰弘, 佐藤充晃, 駒水孝裕, 外山勝彦

人工知能学会全国大会(第33回)論文集 JSAI2019 巻 ( 0 ) 頁： 4E2OS7a02 - 4E2OS7a02 2019年6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：一般社団法人人工知能学会

本研究の目標は，日本法令の要約を提供することである．そのためにランダムフォレストによる重要文抽出に基づく自動要約を提案する．従来の自動要約に関する研究においては，原文書の情報のみが用いられてきた．近年では機械学習に基づく手法なども提案されている．しかし，そうした機械学習において利用される学習データの量は，特に日本語においては，充分でなかった．それに対し，本研究の法令の要約においては，政府が作成する「法令のあらまし」を利用することにより，この問題を解決する．さらに，従来利用されてきた決定木やSVMを使った手法に代えて，ランダムフォレストを用いた重要文抽出を提案し，その性能が従来手法を上回ることを示す．本論文の貢献は，従来よりもサイズの大きな要約用コーパスを作成した点と，重要文抽出におけるランダムフォレストの有効性を確認した点にある．

DOI： 10.11517/pjsai.jsai2019.0_4e2os7a02

CiNii Research

Detecting Communities and Correlated Attribute Clusters on Multi-Attributed Graphs 査読有り

Ito, H; Komamizu, T; Amagasa, T; Kitagawa, H

IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS E102D 巻 ( 4 ) 頁： 810 - 820 2019年4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：IEICE Transactions on Information and Systems

Multi-attributed graphs, in which each node is characterized by multiple types of attributes, are ubiquitous in the real world. Detection and characterization of communities of nodes could have a significant impact on various applications. Although previous studies have attempted to tackle this task, it is still challenging due to difficulties in the integration of graph structures with multiple attributes and the presence of noises in the graphs. Therefore, in this study, we have focused on clusters of attribute values and strong correlations between communities and attribute-value clusters. The graph clustering methodology adopted in the proposed study involves Community detection, Attribute-value clustering, and deriving Relationships between communities and attribute-value clusters (CAR for short). Based on these concepts, the proposed multi-attributed graph clustering is modeled as CAR-clustering. To achieve CAR-clustering, a novel algorithm named CARNMF is developed based on non-negative matrix factorization (NMF) that can detect CAR in a cooperative manner. Results obtained from experiments using real-world datasets show that the CARNMF can detect communities and attribute-value clusters more accurately than existing comparable methods. Furthermore, clustering results obtained using the CARNMF indicate that CARNMF can successfully detect informative communities with meaningful semantic descriptions through correlations between communities and attribute-value clusters.

DOI： 10.1587/transinf.2018DAP0022

Web of Science

Scopus

CiNii Research

Japanese Mistakable Legal Term Correction using Infrequency-aware BERT Classifier

Yamakoshi, T; Komamizu, T; Ogawa, Y; Toyama, K

2019 IEEE INTERNATIONAL CONFERENCE ON BIG DATA (BIG DATA) 頁： 4342 - 4351 2019年

Thai legal term correction using random forests with outside-the-sentence features

Yamakoshi T., Satayamas V., Chanlekha H., Ogawa Y., Komamizu T., Kawtrakul A., Toyama K.

Proceedings of the 33rd Pacific Asia Conference on Language, Information and Computation, PACLIC 2019 頁： 279 - 287 2019年

出版者・発行元：Proceedings of the 33rd Pacific Asia Conference on Language, Information and Computation, PACLIC 2019

We propose a method for finding and correct- ing misused Thai legal terms in Thai statu- tory sentences. Our method predicts legal terms using Random Forest classifiers, each of which is optimized for each set of similar legal terms. Each classifier utilizes outside- the-sentence features, namely, promulgation year, title keywords, and section keywords of statutes, in addition to words adjacent to the targeted legal term. Our experiment shows that our method outperformed not only a Ran- dom Forest method without the outside-the- sentence features, but also BERT (Bidirec- tional Encoder Representations from Trans- formers), a powerful language representation model, in overall accuracy.

Scopus

nagoy Team's Summarization System at the NTCIR-14 QA Lab-PoliInfo 査読有り

Yasuhiro Ogawa, Michiaki Satou, Takahiro Komamizu, Katsuhiko Toyama

Post-conference Proceedings of the 14th NTCIR Conference on Evaluation of Information Access Technologies 11966 LNCS 巻頁： to appear - 121 2019年

記述言語：英語掲載種別：研究論文（その他学術会議資料等）出版者・発行元：Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

The nagoy team participated in the NTCIR-14 QA Lab-PoliInfo’s summarization subtask. This paper describes our summarization system for assembly member speeches using random forest classifiers. Since we encountered an imbalance in the data, we were unable to achieve good results in this subtask when training on all data. To solve this problem, we developed a new summarization system that applies multiple random forest classifiers training on different-sized data sets step by step. As a result, our system achieved good performance, especially in the evaluation by ROUGE scores. In this paper, we also compare our system with a single random forest classifier using probability.

DOI： 10.1007/978-3-030-36805-0_9

Scopus

その他リンク： https://dblp.uni-trier.de/db/conf/ntcir/ntcir2019.html#OgawaSKT19

Graph Analytical Re-ranking for Entity Search

Komamizu T.

CEUR Workshop Proceedings 2482 巻 2019年

出版者・発行元：CEUR Workshop Proceedings

Entity search is a fundamental task in Linked Data (LD). The task is, given a keyword search query, to retrieve a set of entities in LD which are relevant to the query. The state-of-the-art approaches for entity search are based on information retrieval technologies such as TF-IDF vectorization and ranking models. This paper examines the approaches by applying a traditional evaluation metrics, recall@k, and shows ranking qualities still room left for improvements. In order to improve the ranking qualities, this paper explores possibilities of graph analytical methods. LD is regarded as a large graph, graph analytical approaches are therefore appropriate for this purpose. Since query-based graph analytical approaches fit to entity search tasks, this paper proposes a personalized PageRank-based re-ranking method, PPRSD (Personalized PageRank based Score Distribution), for retrieved results by the state-of-the-art. The experimental evaluation recognizes improvements but its results are not satisfactory, yet. For further improvements, this paper reports investigations about relationship between queries and entities in terms of path lengths on the graph, and discusses future directions for graph analytical approaches.

Scopus

Analyzing Japanese law history through modeling multi-versioned entity

Komamizu T., Uchida Y., Ogawa Y., Toyama K.

CEUR Workshop Proceedings 2599 巻 2019年

出版者・発行元：CEUR Workshop Proceedings

As law is a blueprint of a society and is changed over time as social environments changed, analyzing histories (change provenances) of laws can reveal important facts such as legislative facts and critical events for the society. Linked Open Data (LOD) has emerged as a preferred method for publishing and sharing open data, however, there is an ontological barrier for publishing law history data as LOD. To break through the barrier, this paper proposes an ontology for law history data of the Japanese statute law. The ontology is inspired from PROV-O and SIOC ontologies. The LOD dataset based on the proposed ontology enables wide variety of analyses on the law history data by simple SPARQL queries. The analyses include simple search, visualization, temporal analysis, data mining, etc. This paper displays parts of the analyses which indicate several legislative facts behind changes of laws. The analyses demonstrate the proposed ontology and LOD dataset are useful for legal data analysis. The proposed ontology is comparable with ELI (European Legislation Identifier) which is designed for EU laws, this paper thus discusses the comparability and future directions of the proposed ontology.

Scopus

Graph Analytical Re-ranking for Entity Search

Takahiro Komamizu

Proceedings of the 1st International Workshop on EntitY REtrieval (EYRE 2018) 頁： (to appear) 2018年10月

Learning Interpretable Entity Representation in Linked Data 査読有り

Takahiro Komamizu

Proc. the 29th International Conference on Database and Expert Systems Applications 頁： 153-168 2018年9月

Community Detection and Correlated Attribute Cluster Analysis on Multi-Attributed Graphs 査読有り

Hiroyoshi Ito, Takahiro Komamizu, Toshiyuki Amagasa, Hiroyuki Kitagawa

Proc. the 2nd International workshop on Data Analytics solutions for Real-LIfe APplications (DARLI-AP 2018) co-located with the 21st International Conference on Extending Database Technology (EDBT 2018) 2018年3月

Network-Word Embedding for Dynamic Text Attributed Networks 査読有り

Hiroyoshi Ito, Takahiro Komamizu, Toshiyuki Amagasa, Hiroyuki Kitagawa

Proc. the 6th International Workshop on Semantic Computing for Social Networks and Organization Sciences (SCSN 2018) co-located with the 12th IEEE International Conference on Semantic Computing (ICSC 2018) 2018年1月

Analytical toolbox for smart city applications: Garbage collection log use case 査読有り

Takahiro Komamizu, Jin Nakazawa, Toshiyuki Amagasa, Hiroyuki Kitagawa, Hideyuki Tokuda

Proceedings - 2017 IEEE International Conference on Big Data, Big Data 2017 2018- 巻頁： 4105 - 4110 2018年1月

記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：Institute of Electrical and Electronics Engineers Inc.

Analyzing and feeding back the results on real-world services are important missions in the Big Data era to realize smart city. However, analyzing real-world data is still challenging because of dirtiness of data and large variety of analytic requirements. To cope with the challenges, this paper proposes and develops an analytical toolbox for smart city applications. The analytical toolbox consists of three phases: preparation, analysis, and visualization. The preparation phase deals with the dirtiness of the data by including fundamental data cleansing techniques and data integration techniques. The analysis phase is responsible for ETL (extract, transform and load) process and analytical query processing from the next phase. The visualization phase deals with analytical requirements from users and visualization of analytical results. This paper showcases a real-world use case of the proposed analytical toolbox. The use case is now open in public with help of Fujisawa city, Japan, and this fact indicates that the proposed analytical toolbox is feasible for real-world data analysis and feeding back to citizens.

DOI： 10.1109/BigData.2017.8258429

Scopus

　前のページ - 次のページ　

<駒水 孝裕>

論文 - 駒水 孝裕

<駒水　孝裕>

論文 - 駒水　孝裕