Japanese Wikipedia Research Articles

Answering Japanese Fill-in-the-Blank Questions in the Radiological Technology Field Using BERT.

The purpose of this study was to evaluate the accuracy of Japanese fill-in-the-blank questions in the radiological technology field using six BERT models. Our original model, created using 330,000 abstracts related to radiological technology, was as accurate as other models created by Wikipedia when vocabulary size was the same. Further improvements in accuracy can be expected if we can increase the vocabulary size and add Japanese Wikipedia data in training.

An Advanced Reading Course in Japanese with Wikipedia: A Case Study

This paper proposes integrating the activity of Wikipedia-article authoring into an advanced-level Japanese reading course. With its unprecedented wealth of information and unique feature of collaborative writing, Wikipedia has been recognized as a powerful pedagogical tool. Nevertheless, adapting it to a course requires a considerable amount of planning and preparation. Probably because of this high threshold, it is yet to be widely employed by L2 educators. In the fall of 2022, we experimentally taught a course, assigning short Japanese Wikipedia articles for translation into English. This paper provides information sufficient for those interested in experimenting with this potential resource to enhance their courses.

Crowdsourcing Knowledge Production of COVID-19 Information on Japanese Wikipedia in the Face of Uncertainty: Empirical Analysis.

A worldwide overabundance of information comprising misinformation, rumors, and propaganda concerning COVID-19 has been observed in addition to the pandemic. By addressing this data confusion, Wikipedia has become an important source of information. This study aimed to investigate how the editors of Wikipedia have handled COVID-19-related information. Specifically, it focused on 2 questions: What were the knowledge preferences of the editors who participated in producing COVID-19-related information? and How did editors with different knowledge preferences collaborate? This study used a large-scale data set, including >2 million edits in the histories of 1857 editors who edited 133 articles related to COVID-19 on Japanese Wikipedia. Machine learning methods, including graph neural network methods, Bayesian inference, and Granger causality analysis, were used to establish the editors' topic proclivity and collaboration patterns. Overall, 3 trends were observed. Two groups of editors were involved in the production of information on COVID-19. One group had a strong preference for sociopolitical topics (social-political group), and the other group strongly preferred scientific and medical topics (scientific-medical group). The social-political group played a central role (contributing 16,544,495/23,485,683, 70.04% of bits of content and 57,969/76,673, 75.61% of the references) in the information production part of the COVID-19 articles on Wikipedia, whereas the scientific-medical group played only a secondary role. The severity of the pandemic in Japan activated the editing behaviors of the social-political group, leading them to contribute more to COVID-19 information production on Wikipedia while simultaneously deactivating the editing behaviors of the scientific-medical group, resulting in their less contribution to COVID-19 information production on Wikipedia (Pearson correlation coefficient=0.231; P<.001). The results of this study showed that lay experts (ie, Wikipedia editors) in the fields of science and medicine tended to remain silent when facing high scientific uncertainty related to the pandemic. Considering the high quality of the COVID-19-related articles on Japanese Wikipedia, this research also suggested that the sidelining of the science and medicine editors in discussions is not necessarily a problem. Instead, the social and political context of the issues with high scientific uncertainty is more important than the scientific discussions that support accuracy.

TextCNN-based ensemble learning model for Japanese Text Multi-classification

In this paper, we aim at improving Japanese text classification using TextCNN-based ensemble learning model. Specifically, we first construct three different sub-classifiers, combining ALBERT, RoBERTa, DistilBERT with TextCNN, respectively; and then explore the effectiveness of ensemble learning model to leverage complementary information from different sub-classifiers for better text classification. We also conduct a series of experiments with the dataset collected from Japanese Wikipedia pages, which was divided into 31 categories. The experimental results show that the proposed approach achieves a good performance. The accuracy, precision, recall and F1 scores reach 0.881, 0.884, 0.880 and 0.881, respectively, which shows that the TextCNN-based ensemble learning model can be used for Japanese Text Multi-Classification effectively.

Introducing an “invisible enemy”: A case study of knowledge construction regarding microplastics in Japanese Wikipedia

Microplastics, a novel global environmental problem, has become a fast-developing research field with worldwide investments. With scientific development, social science studies are expected to contribute to developing more effective communication on microplastics. Among various media, this article focused on Wikipedia and discussed one particular situation—when the language of Wikipedia differs from the language of mainstream information sources. Inspired by actor network theory, we analyzed the revision history of the Japanese Wikipedia article on microplastics from 2014 to 2020 and elucidated how knowledge of microplastics was textually constructed. We found editors’ reluctance to disprove and update the latest scientific knowledge in edit actions, as well as problems and flaws existing in the created knowledge that shows limited enforcement of Wikipedia policies. Based on the findings, we urge more attention to improving the public knowledge of fast-developing science, especially when the knowledge construction might encounter a cross-language context.

Public Interest in Birds and Congruence with Seasonal Observational Data: A Case Study Using Japanese Wikipedia Pageview Data

Public Interest in Birds and Congruence with Seasonal Observational Data: A Case Study Using Japanese Wikipedia Pageview Data

A clinical specific BERT developed using a huge Japanese clinical text corpus.

Generalized language models that are pre-trained with a large corpus have achieved great performance on natural language tasks. While many pre-trained transformers for English are published, few models are available for Japanese text, especially in clinical medicine. In this work, we demonstrate the development of a clinical specific BERT model with a huge amount of Japanese clinical text and evaluate it on the NTCIR-13 MedWeb that has fake Twitter messages regarding medical concerns with eight labels. Approximately 120 million clinical texts stored at the University of Tokyo Hospital were used as our dataset. The BERT-base was pre-trained using the entire dataset and a vocabulary including 25,000 tokens. The pre-training was almost saturated at about 4 epochs, and the accuracies of Masked-LM and Next Sentence Prediction were 0.773 and 0.975, respectively. The developed BERT did not show significantly higher performance on the MedWeb task than the other BERT models that were pre-trained with Japanese Wikipedia text. The advantage of pre-training on clinical text may become apparent in more complex tasks on actual clinical text, and such an evaluation set needs to be developed.

Semantic Textual Similarity in Japanese Clinical Domain Texts Using BERT.

Background Semantic textual similarity (STS) captures the degree of semantic similarity between texts. It plays an important role in many natural language processing applications such as text summarization, question answering, machine translation, information retrieval, dialog systems, plagiarism detection, and query ranking. STS has been widely studied in the general English domain. However, there exists few resources for STS tasks in the clinical domain and in languages other than English, such as Japanese. Objective The objective of this study is to capture semantic similarity between Japanese clinical texts (Japanese clinical STS) by creating a Japanese dataset that is publicly available. Materials We created two datasets for Japanese clinical STS: (1) Japanese case reports (CR dataset) and (2) Japanese electronic medical records (EMR dataset). The CR dataset was created from publicly available case reports extracted from the CiNii database. The EMR dataset was created from Japanese electronic medical records. Methods We used an approach based on bidirectional encoder representations from transformers (BERT) to capture the semantic similarity between the clinical domain texts. BERT is a popular approach for transfer learning and has been proven to be effective in achieving high accuracy for small datasets. We implemented two Japanese pretrained BERT models: a general Japanese BERT and a clinical Japanese BERT. The general Japanese BERT is pretrained on Japanese Wikipedia texts while the clinical Japanese BERT is pretrained on Japanese clinical texts. Results The BERT models performed well in capturing semantic similarity in our datasets. The general Japanese BERT outperformed the clinical Japanese BERT and achieved a high correlation with human score (0.904 in the CR dataset and 0.875 in the EMR dataset). It was unexpected that the general Japanese BERT outperformed the clinical Japanese BERT on clinical domain dataset. This could be due to the fact that the general Japanese BERT is pretrained on a wide range of texts compared with the clinical Japanese BERT.

Proposal of Japanese Vocabulary Difficulty Level Dictionaries for Automated Essay Scoring Support System Using Rubric

We are developing a Moodle plug-in, which is an AES (automated essay scoring) support system for the basic education of university students. Our system evaluates essays based on rubric, which has five evaluation viewpoints “Contents, Structure, Evidence, Style, and Skill”. Vocabulary level is one of the scoring items of Skill. It is calculated using Japanese Language Learners’ Dictionaries constructed by Sunakawa et al. Since this does not fully cover the words used in the student-level essays, we found that there is a problem with the accuracy of the vocabulary level scoring. In this paper, we propose to construct comprehensive Japanese vocabulary difficulty level dictionaries using Japanese Wikipedia as the corpus. We apply Latent Dirichlet Allocation (LDA) to the Wikipedia corpus and find the word appearance probability as one of the indexes of word difficulty. We use the TF-IDF value instead of the LDA value of the words, which rarely appears. As a result, we constructed highly comprehensive Japanese vocabulary difficulty level dictionaries. We confirmed that the vocabulary level can be scored for all words in the test dataset by using the constructed dictionaries.

"Anonymous calling": The WikiScanner scandals and anonymity on the Japanese Wikipedia

The Wikiscanner tool, which traced the origin of edits on Wikipedia, stirred media scandals throughout the world. Relying on a “trace ethnography” method, following the discussion on Wikipedia articles, this article deals with the Japanese edition reaction to the scandals. I argue that this reaction represents a unique form of online publicity that facilitates anonymous normative discussion. In addition, by contextualizing the reaction to the historical and cultural conditions of the Japanese Web, the article contends that Wikipedia enables a rare model of anonymous public debate which bridges earlier Japanese conceptions of anonymity and publicity.

Relationship between word2vec representation of Japanese Wikipedia and Word properties

Relationship between word2vec representation of Japanese Wikipedia and Word properties

Extracting location and creator-related information from Wikipedia-based information-rich taxonomy for ConceptNet expansion

Our research goal is to generate new assertions suitable for introduction to the Japanese part of the ConceptNet common sense knowledge ontology. In this paper we present a method for extracting IsA assertions (hyponymy relations), AtLocation assertions (informing of the location of an object or place), LocatedNear assertions (informing of neighboring locations) and CreatedBy assertions (informing of the creator of an object) automatically from Japanese Wikipedia XML dump files. We use the Hyponymy extraction tool v1.0, which analyzes definition, category and hierarchy structures of Wikipedia articles to extract IsA assertions and produce an information-rich taxonomy. From this taxonomy we extract additional information, in this case AtLocation, LocatedNear and CreatedBy types of assertions, using our original method. The presented experiments prove that we achieved our research goal on a large scale: both methods produce satisfactory results, and we were able to acquire 5,866,680 IsA assertions with 96.0% reliability, 131,760 AtLocation assertion pairs with 93.5% reliability, 6217 LocatedNear assertion pairs with 98.5% reliability and 270,230 CreatedBy assertion pairs with 78.5% reliability. Our method surpassed the baseline system in terms of both precision and the number of acquired assertions.

A Comparison of Concept-base Model and Word Distributed Model as Word Association System

We construct Concept-base based on concept chain model and word vector spaces based on Word2Vec using EDR-electronic- dictionary and Japanese Wikipedia data. This paper describes verification experiments of these models regarding the word association system based on the association-frequency-table. In these experiments, we investigate the tendency using associative words of evaluation basis words obtained by these models. In Concept-base model, we observed a tendency that synonyms, superordinate words, and subordinate words are obtained as associative words. Furthermore we observed a tendency that words, which can be compounds or co-occurrence phrases after connecting headwords of the association-frequency-table, are used as associative words in the Word2Vec model. Moreover evaluation result showed the tendency that associative words mostly have category words in the Word2Vec model.

Integrated Parallel Sentence and Fragment Extraction from Comparable Corpora

Parallel corpora are crucial for statistical machine translation (SMT); however, they are quite scarce for most language pairs and domains. As comparable corpora are far more available, many studies have been conducted to extract either parallel sentences or fragments from them for SMT. In this article, we propose an integrated system to extract both parallel sentences and fragments from comparable corpora. We first apply parallel sentence extraction to identify parallel sentences from comparable sentences. We then extract parallel fragments from the comparable sentences. Parallel sentence extraction is based on a parallel sentence candidate filter and classifier for parallel sentence identification. We improve it by proposing a novel filtering strategy and three novel feature sets for classification. Previous studies have found it difficult to accurately extract parallel fragments from comparable sentences. We propose an accurate parallel fragment extraction method that uses an alignment model to locate the parallel fragment candidates and an accurate lexicon-based filter to identify the truly parallel fragments. A case study on the Chinese--Japanese Wikipedia indicates that our proposed methods outperform previously proposed methods, and the parallel data extracted by our system significantly improves SMT performance.

日本語版WikipediaにおけるDOIリンクの予備的分析

本研究では，日本語版Wikipedia に含まれるDOIリンクの予備的分析を行った．2015年3月時点での日本語版Wikipedia には28,546件のDOIリンクがあり，そのうち標準名前空間ページに含まれる27,201件について，DOI登録機関はCrossRefが97%，JaLCが2%であった．日本国外の大手出版社が多く，雑誌タイトルレベルでは，Nature，Science，PNASなどの自然科学分野の有力誌が多く含まれていた．また，日本国内の出版者のDOIリンクが含まれていた．

Building up a class hierarchy with properties by refining and integrating Japanese Wikipedia Ontology and Japanese WordNet

Previously, we constructed the Japanese Wikipedia Ontology JWO via a semi-automatic process using the Japanese Wikipedia, but it had problems due to a lack of upper classes and appropriate definitions of properties. Thus, the aim of the current study was to complement the upper classes in JWO by refining and integrating JWO and Japanese WordNet JWN to build a class hierarchy with defined properties based on the considerations of property inheritance. To achieve this, we developed tools that help users to refine the class-instance relationships, to identify the JWO classes that need to be aligned with JWN synsets, and to align the JWO classes with the JWN synsets via user interaction. We also integrated JWO and JWN using a domain ontology development environment, DODDLE-OWL. We also propose a method for building a class hierarchy with defined properties by elevating the common properties defined in sibling classes to higher classes in JWO.

Building up Ontologies with Many Properties from Japanese Wikipedia

Here is discussed how to build up ontologies with many properties from Japanese Wikipedia. The ontologies include is-a relationship (rdfs:subClassOf), class-instance relationship (rdf:type) and synonym relation (skos:altLabel) moreover it includes property relations and types. Property relations are triples, property domain (rdfs:domain) and property range (rdfs:range). Property types are object (owl:ObjectProperty), data (owl:DatatypeProperty), symmetric (owl:SymmetricProperty), transitive (owl:TransitiveProperty), functional (owl:FunctionalProperty) and inverse functional (owl:InverseFunctionalProperty). Experimental case studies show us that the built Japanese Wikipedia Ontology goes better than DBpedia from utility when we use, such as Hub of Linked Data, especially in Japan.

日本語版Ｗｉｋｉｐｅｄｉａからの外部リンクの特徴とリンク切れの発生状況

本研究では情報探索行動の起点としてのWikipediaの有効性について検討することを目的に，日本語版Wikipediaの外部リンク約119万件について，URLの詳細やアクセス障害・リンク切れの状況について調査した．結果から，1)日本語版Wikipediaからのリンクの11%程度でアクセスに障害がある，2) edu，co.jp，go.jpドメインの外部リンクでアクセス障害が多い，3) 新聞社が運営するニュースサイトで特にアクセス障害が多いこと等を明らかにした．

Learning a Large Scale of Ontology from Japanese Wikipedia

Here is discussed how to learn a large scale of ontology from Japanese Wikipedia. The learned ontology includes the following properties: rdfs:subClassOf (IS-A relationship), rdf:type (class-instance relationship), owl:Object/DatatypeProperty (Infobox triple), rdfs:domain (property domain), and skos:altLabel (synonym). Experimental case studies show us that the learned Japanese Wikipedia Ontology goes better than already existing general linguistic ontologies, such as EDR and Japanese WordNet, from the points of building costs and structure information richness.