研究者詳細 - 松原　茂樹

論文 - 松原　茂樹

分割表示 239 件中 41 - 60 件目／全件表示 >>

A Model-Theoretic Formalization of Natural Language Inference Using Neural Network and Tableau Method 査読有り Open Access

加藤芳秀, 松原茂樹

Proceedings of the 36th Pacific Asia Conference on Language, Information and Computation - 巻頁： 430 - 437 2022年

Estimating the Generation Timing of Responsive Utterances by Active Listeners of Spoken Narratives 査読有り

Koichiro Ito, Masaki Murata, Tomohiro Ohno, and Shigeki Matsubara

Proceeding of 2021 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU 2021) 頁： 495 - 502 2021年12月

Using Citation Contexts in Scholarly Papers for Research Data Search 査読有り

Masaya Tsunokake and Shigeki Matsubara

Proceedings of the 16th International Joint Symposium on Artificial Intelligence and Natural Language Processing (iSAI-NLP 2021) 2021年12月

Identification of Important Utterances in Narrative Speech Using Attentive Listening Responses 査読有り

Shunsuke Kosugi, Koichiro Ito, Masaki Murata, Tomohiro Ohno, and Shigeki Matsubara

Proceedings of the 16th International Joint Symposium on Artificial Intelligence and Natural Language Processing (iSAI-NLP 2021) 2021年12月

A New Representation for Span-based CCG Parsing 査読有り

Yoshihide Kato and Shigeki Matsubara

Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP-2021) 頁： 10579 - 10584 2021年11月

Natural Language Inference using Neural Network and Tableau Method 査読有り

Ayahito Saji, Daiki Takao, Yoshihide Kato and Shigeki Matsubara

Proceedings of the 35th Pacific Asia Conference on Language, Information and Computation (PACLIC 35) 頁： 406 - 414 2021年11月

同時通訳における順送り方略の適用に影響する起点言語の特徴査読有り

蔡仲煕, 笠浩一朗, 松原茂樹

通訳翻訳研究 20 巻頁： 25 - 40 2021年2月

Word reordering and comma insertion integrated with shift-reduce dependency parsing 査読有り

Miyachi K., Ohno T., Matsubara S.

ICAART 2021 - Proceedings of the 13th International Conference on Agents and Artificial Intelligence 2 巻頁： 1144 - 1150 2021年2月

記述言語：日本語掲載種別：研究論文（国際会議プロシーディングス）出版者・発行元：ICAART 2021 - Proceedings of the 13th International Conference on Agents and Artificial Intelligence

Japanese has widely recognized as relatively free word order language. However, since Japanese word order is not completely arbitrary and has some sort of preference, even native Japanese writers often write Japanese sentences that are grammatically well-formed but not easy to read. Furthermore, in Japanese sentences, a comma plays an important role in explicitly separating the constituents such as words and phrases. This paper proposes a method of word reordering and comma insertion for hard-to-read Japanese sentences so that they become easier to read, as basic technique. Our contribution is to show the feasibility of concurrently performing word reordering, comma insertion, and dependency parsing.

Scopus

A New Representation for Span-based CCG Parsing

Kato, Y; Matsubara, S

2021 CONFERENCE ON EMPIRICAL METHODS IN NATURAL LANGUAGE PROCESSING (EMNLP 2021) 頁： 10579 - 10584 2021年

Classification of URLs Citing Research Artifacts in Scholarly Documents based on Distributed Representations 国際誌 Open Access

Tsunokake M., Matsubara S.

Ceur Workshop Proceedings 3004 巻頁： 20 - 25 2021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：Ceur Workshop Proceedings

This paper describes methods for classifying URLs referring to research artifacts in scholarly papers, and examines their classification performance. The methods discriminate whether a URL refers to a research artifact or not and classify the identified URL into “tool” or “data.” The methods use distributed representations obtained from citation contexts of the URL. Each component of a URL can be regarded as a word, and the meaning of the entire URL can be generated by synthesizing the distributed representation of each component using compositional functions. This paper evaluates several types of compositional functions from the viewpoint of classification performance. Experiments with using URLs in international conference papers showed the effectiveness of our proposed compositional functions.

Open Access

Scopus

Estimating the Generation Timing of Responsive Utterances by Active Listeners of Spoken Narratives

Ito K., Murata M., Ohno T., Matsubara S.

2021 IEEE Automatic Speech Recognition and Understanding Workshop Asru 2021 Proceedings 頁： 495 - 502 2021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：2021 IEEE Automatic Speech Recognition and Understanding Workshop Asru 2021 Proceedings

Narrating is a fundamental human desire. In order to satisfy the desire, it is more effective to have multiple listeners listen to narratives than to have one listener does. Conversational agents, e.g., communication robots and smart speakers, are expected to listen to narratives, instead of human listeners. For such agents to be recognized as narrative listeners, it is necessary to generate responsive utterances at appropriate timings. This paper proposes a method to estimate the generation timings of responsive utterances by active listeners of spoken narratives. The proposed method estimates them by encoding acoustic and linguistic information extracted from narratives with a transformer-based approach. To evaluate the performance of the proposed method, an experiment was conducted using 131, 616 responsive utterances to narratives. The experimental results demonstrated that the proposed method was effective for the estimation and both the acoustic and linguistic information was useful.

DOI： 10.1109/ASRU51503.2021.9688132

Web of Science

Scopus

A New Representation for Span-based CCG Parsing

Kato Y., Matsubara S.

EMNLP 2021 - 2021 Conference on Empirical Methods in Natural Language Processing, Proceedings 頁： 10579 - 10584 2021年

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：EMNLP 2021 - 2021 Conference on Empirical Methods in Natural Language Processing, Proceedings

This paper proposes a new representation for CCG derivations. CCG derivations are represented as trees whose nodes are labeled with categories strictly restricted by CCG rule schemata. This characteristic is not suitable for span-based parsing models because they predict node labels independently. In other words, span-based models may generate invalid CCG derivations that violate the rule schemata. Our proposed representation decomposes CCG derivations into several independent pieces and prevents the span-based parsing models from violating the schemata. Our experimental result shows that an off-the-shelf span-based parser with our representation is comparable with previous CCG parsers.

Scopus

Natural Language Inference using Neural Network and Tableau Method Open Access

Saji A., Takao D., Kato Y., Matsubara S.

Proceedings of the 35th Pacific Asia Conference on Language Information and Computation Paclic 2021 頁： 303 - 311 2021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 of the 35th Pacific Asia Conference on Language Information and Computation Paclic 2021

Natural language inference (NLI) is the task of identifying the inferential relation between a text pair. Neural-based approaches recently achieved high performance on the NLI task, but they cannot explain their reasoning processes. On the other hand, symbolic approaches have an advantage that the reasoning process is understandable to humans. This paper proposes a method to integrate a neural NLI model and a tableau proof system, the latter of which explains the reasoning processes. The standard tableau method consists of decomposing logical formulas by applying inferential rules, and checking whether or not there exists a valuation that satisfies the given constraints. Unlike the standard tableau method, our method adopts not logical formulas but dependency structures as its ingredients and uses a neural NLI model for the latter process. In order to analyze the behavior of a neural NLI model, we conducted an experiment on the neural NLI model alone and the proposed method using the SNLI corpus.

Open Access

Scopus

Word Reordering and Comma Insertion Integrated with Shift-Reduce Dependency Parsing

Miyachi, K; Ohno, T; Matsubara, S

ICAART: PROCEEDINGS OF THE 13TH INTERNATIONAL CONFERENCE ON AGENTS AND ARTIFICIAL INTELLIGENCE - VOL 2 頁： 1144 - 1150 2021年

What Affects the Word Order of Target Language in Simultaneous Interpretation

Cai Z., Ryu K., Matsubara S.

2020 International Conference on Asian Language Processing, IALP 2020 頁： 135 - 140 2020年12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：2020 International Conference on Asian Language Processing, IALP 2020

A simultaneous interpretation has to be conducted to shorten delay and reduce working memory load. The interpreters adopt a strategy that generates a target language in a word order which is similar to the source language. To make it clear how the interpreters decide the word order in practice, we conduct a statistical study based on the comparison of the word order between translation and simultaneous interpretation. The analysis reveals that in 10.5% of all cases, interpreters maintain the word order of source language, instead of adopting inversion as the translation, and that syntactic factors, including dependencies of chunk pairs, length of post-modifier, affect the word order decision more than non-syntactic factors.

DOI： 10.1109/IALP51396.2020.9310461

Scopus

論文で引用された研究データの同定と分類

角掛正弥, 松原茂樹

電気学会論文誌Ｃ（電子・情報・システム部門誌） 140 巻 ( 12 ) 頁： 1357 - 1364 2020年12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：一般社団法人電気学会

This paper proposes a method for identifying and classifying the research data cited in scholarly papers, aiming at automatic generation of metadata stored in data repository. This study focuses on URL citations in the scholarly papers. That is, the targets are to identify the URLs referring to the research data and to classify them into tool and data. The method is realized as a multi-class classification (tool/data/others). The method acquires the distributed representations of the URLs from the context around them, and uses them as the input feature. There exists an advantage in that the meanings of URLs can be given based on their surrounding words. This study adopts an approach of computing the meaning of the entire URL from those of the components of the URL. In order to evaluate the performance of the proposed method, experiments on URL classification were conducted. The scholarly papers included in the proceedings of the international conference were used as experimental data. Experimental results have shown the effectiveness of the proposed method for identifying and classifying URLs referring to research data.

DOI： 10.1541/ieejeiss.140.1357

Scopus

CiNii Research

Parsing Gapping Constructions Based on Grammatical and Semantic Roles 査読有り

Yoshihide Kato, Shigeki Matsubara

Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP-2020) 頁： 2747 - 2752 2020年11月

Identification of Research Data References Based on Citation Contexts

Ikoma T., Matsubara S.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 12504 LNCS 巻頁： 149 - 156 2020年

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：Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

In this paper, a method for the automatic identification of research data references in publications is proposed for automatically generating research data repositories. The International Conference on Language Resources and Evaluation (LREC) requires authors to list research data references separately from other publication references. The goal of our research is to automate the discrimination process. We investigated the reference lists in LREC papers and the citation contexts to find characteristic features that are useful for identifying research data references. We confirmed that key phrases appeared in the citation contexts and the bibliographical elements in the reference lists. Our proposed method uses the presence or absence of key phrases to identify research data references. Experiments on LREC proceedings papers proved the effectiveness of using key phrases in the citation context.

DOI： 10.1007/978-3-030-64452-9_13

Scopus

Relation between degree of empathy for narrative speech and type of responsive utterance in attentive listening

Ito K., Murata M., Ohno T., Matsubara S.

LREC 2020 - 12th International Conference on Language Resources and Evaluation, Conference Proceedings 頁： 696 - 701 2020年

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：LREC 2020 - 12th International Conference on Language Resources and Evaluation, Conference Proceedings

Nowadays, spoken dialogue agents such as communication robots and smart speakers listen to narratives of humans. In order for such an agent to be recognized as a listener of narratives and convey the attitude of attentive listening, it is necessary to generate responsive utterances. Moreover, responsive utterances can express empathy to narratives and showing an appropriate degree of empathy to narratives is significant for enhancing speaker's motivation. The degree of empathy shown by responsive utterances is thought to depend on their type. However, the relation between responsive utterances and degrees of the empathy has not been explored yet. This paper describes the classification of responsive utterances based on the degree of empathy in order to explain that relation. In this research, responsive utterances are classified into five levels based on the effect of utterances and literature on attentive listening. Quantitative evaluations using 37,995 responsive utterances showed the appropriateness of the proposed classification.

Scopus

同時通訳における順送り方略の適用に影響する起点言語の特徴査読有り

蔡仲煕, 笠浩一朗, 松原茂樹

通訳翻訳研究 20 巻 ( 0 ) 頁： 25 - 40 2020年

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：日本通訳翻訳学会

Simultaneous interpretation has to be conducted to shorten delay and reduce working memory load. Interpreters adopt a strategy that involves the generation of a target language in the word order that is similar to that of the source language. To clarify how the interpreters decide the word order in practice, we conduct a statistical study on the word order of simultaneous interpretation, based on comparison with written translation. The word order comparisons indicate that the strategy of maintaining the original word order is more likely to be adopted in cases where the word order is not consistent in different translations. We analyze the features of source languages, including dependencies of chunk pairs, length of post modifier, and distance between two chunks, and confirm that such features affect the word-order decision in simultaneous interpretation.

DOI： 10.50837/its.2002

CiNii Research

　前のページ - 次のページ　

<松原 茂樹>

論文 - 松原 茂樹

<松原　茂樹>

論文 - 松原　茂樹