Faculty Profiles - MATSUBARA, Shigeki

Generation Timing Estimation of Responsive Utterances to Show Attentive Listening to Narrative Speech Reviewed

ITO Koichiro, MURATA Masaki, OHNO Tomohiro, MATSUBARA Shigeki

Vol. J106-A ( 3 ) page： 136 - 145 2023.3

Authorship：Last author Language：Japanese Publishing type：Research paper (scientific journal) Publisher：The Institute of Electronics, Information and Communication Engineers

Conversational agents, e.g., communication robots and smart speakers, are expected to listen to human narratives. For realizing such agents, it is necessary to generate responsive utterances at appropriate timings. This paper proposes a method to estimate the degree of appropriateness as generation timings of responsive utterances. The proposed method estimates the degree by encoding acoustic and linguistic narrative information with a transformer-based approach. To evaluate the performance, an experiment was conducted using 131,616 responsive utterances. The experimental results demonstrated that the proposed method was effective. Additionally, we confirmed that the degree of appropriateness was useful for response timing detection.

DOI： 10.14923/transfunj.2022hap0013

Bottom-up Japanese Word Ordering using BERT Reviewed

Masato Yamazoe, Tomohiro Ohno and Shigeki Matsubara

Proceedings of the 15th International Conference on Agents and Artificial Intelligence (ICAART-2023) Vol. 3 page： 673 - 681 2023.2

Japanese Word Reordering based on Topological Sort Reviewed

Peng Sun, Tomohiro Ohno and Shigeki Matsubara

Proceedings of the 15th International Conference on Agents and Artificial Intelligence (ICAART-2023) Vol. 3 page： 768 - 775 2023.2

Question Difficulty Decision for Automated Interview in Foreign Language Test Reviewed

Minghao Lin, Koichiro Ito and Shigeki Matsubara

Proceedings of the 15th International Conference on Agents and Artificial Intelligence (ICAART-2023) Vol. 3 page： 768 - 775 2023.2

Question Difficulty Decision for Automated Interview in Foreign Language Test Open Access

Lin M., Ito K., Matsubara S.

International Conference on Agents and Artificial Intelligence Vol. 3 page： 760 - 767 2023

Language：English Publishing type：Research paper (scientific journal) Publisher：International Conference on Agents and Artificial Intelligence

As globalization continues to progress all over the world, demand is growing for objective and rapid assessment of language proficiency in foreign language learners. While automated assessment of listening, reading, and writing skills has been proposed, little research has been done to automate assessment of speaking skills. In this paper, we propose a method of deciding the difficulty of questions generated for interview tests of skills assessment in speaking a foreign language. To address question difficulty flexibly according to the abilities of test takers, our method considers the appropriateness of responses from test takers. We implemented this method using the large-scale pre-trained language model BERT (Bidirectional Encoder Representation from Transformers). Experiments were conducted using simulated test data from the Japanese Learner’s Conversation Database to confirm the effectiveness of our method in deciding difficulty.

DOI： 10.5220/0011774900003393

Japanese Word Reordering Based on Topological Sort Open Access

Sun P., Ohno T., Matsubara S.

International Conference on Agents and Artificial Intelligence Vol. 3 page： 768 - 775 2023

Language：English Publishing type：Research paper (scientific journal) Publisher：International Conference on Agents and Artificial Intelligence

In Japanese, some sentences are grammatically well-formed, but not easy to read. This paper proposes a method for Japanese word reordering, which first adapts a Japanese BERT model to predict suitable pairwise orderings between two words in a sentence, and then converts the predicted results into a graph. The vertices in the graph represent the words in the sentence. The edges represent the pairwise orderings between two words. Finally, topological sort is applied to find the correct word ordering in a sentence by visiting each graph vertex. We conducted an evaluation experiment with uneasy-to-read Japanese sentences created from newspaper article sentences.

DOI： 10.5220/0011775000003393

Bottom-up Japanese Word Ordering Using BERT Open Access

Yamazoe M., Ohno T., Matsubara S.

International Conference on Agents and Artificial Intelligence Vol. 3 page： 673 - 681 2023

Language：English Publishing type：Research paper (scientific journal) Publisher：International Conference on Agents and Artificial Intelligence

Although Japanese is widely regarded as a relatively free word order language, word order in Japanese is not entirely arbitrary and has some sort of preference. As a result, it is an important technique to produce a sentence that is not only grammatically correct but also easy to read. This paper proposes a method for the word ordering of a whole Japanese sentence when the dependency relations between words are given. Using BERT, the method identifies the easy-to-read word order for a syntactic tree based on bottom-up processing. We confirmed the effectiveness of our method through an experiment on word ordering using newspaper articles.

DOI： 10.5220/0011743600003393

Using Citation Contexts for Identifying Meaningful References in Scholarly Papers Open Access

MATSUBARA Shigeki, IKOMA Tomoki

Proceedings of the Annual Conference of JSAI Vol. JSAI2023 ( 0 ) page： 1E3GS604 - 1E3GS604 2023

Language：Japanese Publishing type：Research paper (scientific journal) Publisher：The Japanese Society for Artificial Intelligence

Citation count is the most recognized metric for measuring the impact of scholarly papers. However, cited papers contribute to a citing paper in a variety of ways. In order to accurately measure the individual impact of cited papers, it is crucial to quantify their respective degree of contribution to a citing paper. This study proposes a method to automatically identify which papers are highly meaningful to a citing paper by extracting the description of a cited paper from a citing sentence for use in identifying meaningful references. Experimental results demonstrated the effectiveness of the extracted description.

DOI： 10.11517/pjsai.jsai2023.0_1e3gs604

Extracting Information about Research Resources from Scholarly Papers

Extracting Information about Research Resources from Scholarly Papers Reviewed

Ayahito Saji and Shigeki Matsubara

Proceedings of the 24th International Conference on Asia-Pacific Digital Libraries (ICADL-2022) 2022.12

Classification of URL Citations in Scholarly Papers for Promoting Utilization of Research Artifacts Reviewed

Masaya Tsunokake and Shigeki Matsubara

Proceedings of AACL-IJCNLP-2022 The 1st Workshop on Information Extraction from Scientific Publications 2022.11

A Model-Theoretic Formalization of Natural Language Inference Using Neural Network and Tableau Method Reviewed

Ayahito Saji, Yoshihide Kato, Shigeki Matsubara

Proceedings of the 36th Pacific Asia Conference on Language, Information and Computation (PACLIC 35) 2022.10

Saji, A; Matsubara, S

FROM BORN-PHYSICAL TO BORN-VIRTUAL: AUGMENTING INTELLIGENCE IN DIGITAL LIBRARIES, ICADL 2022 Vol. 13636 page： 440 - 448 2022

Language：English Publishing type：Research paper (scientific journal) Publisher：Lecture Notes in Computer Science Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics

This paper presents a method to extract information about research resource metadata from scholarly papers to expand research resource repositories using an academic knowledge graph. Here, we considered the hypothesis that information extracted from scholarly papers is beneficial to expand research resource repositories from two perspectives: (1) the amount of information in existing metadata, and (2) the number of entries in existing repositories. We constructed an academic knowledge graph, where the nodes and directed edges in the graph correspond to the entities and their relations in scholarly papers, respectively. To verify our hypothesis, we constructed a knowledge graph using 15,721 papers published in international conferences. We then investigated the expandability of a language resource metadata repository using the constructed knowledge graph. The experimental results demonstrated that the constructed knowledge graph could be used to enrich the descriptions of metadata and increase the number of entries in the repository.

DOI： 10.1007/978-3-031-21756-2_35

Web of Science

Construction of Responsive Utterance Corpus for Attentive Listening Response Production Reviewed

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

2022 Language Resources and Evaluation Conference, LREC 2022 page： 7244 - 7252 2022

Authorship：Last author Language：English Publishing type：Research paper (international conference proceedings) Publisher：2022 Language Resources and Evaluation Conference, LREC 2022

In Japan, the number of single-person households, particularly among the elderly, is increasing. Consequently, opportunities for people to narrate are being reduced. To address this issue, conversational agents, e.g., communication robots and smart speakers, are expected to play the role of the listener. To realize these agents, this paper describes the collection of conversational responses by listeners that demonstrate attentive listening attitudes toward narrative speakers, and a method to annotate existing narrative speech with responsive utterances is proposed. To summarize, 148,962 responsive utterances by 11 listeners were collected in a narrative corpus comprising 13,234 utterance units. The collected responsive utterances were analyzed in terms of response frequency, diversity, coverage, and naturalness. These results demonstrated that diverse and natural responsive utterances were collected by the proposed method in an efficient and comprehensive manner. To demonstrate the practical use of the collected responsive utterances, an experiment was conducted, in which response generation timings were detected in narratives.

Construction of Responsive Utterance Corpus for Attentive Listening Response Production

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

LREC 2022: THIRTEEN INTERNATIONAL CONFERENCE ON LANGUAGE RESOURCES AND EVALUATION page： 7244 - 7252 2022

Use of empty categories for conversion of phrase structure into dependency structure Open Access

KATO Yoshihide, MATSUBARA Shigeki Kobayashi, KOBAYASHI Mikiteru

Proceedings of the Annual Conference of JSAI Vol. JSAI2022 ( 0 ) page： 1P4GS603 - 1P4GS603 2022

Language：Japanese Publishing type：Research paper (scientific journal) Publisher：The Japanese Society for Artificial Intelligence

Syntactic conversion methods are useful for comparing the performance of parsers that are based on different syntactic formalism. Most of the methods, however, do not treat empty categories, which the current phrase structure parsers can identify. This paper proposes a method that uses empty category information for conversion from phrase structures into dependency structures. Our proposed method converts a phrase structure of Penn Treebank (PTB) into a dependency structure of Enhanced Universal Dependencies (EUD). The method copies the phrase structures corresponding to empty categories and replaces each empty category with the corresponding phrase structure. This treatment of empty categories enables us to derive dependency relations that are represented by empty categories.

DOI： 10.11517/pjsai.jsai2022.0_1p4gs603

A Model-Theoretic Formalization of Natural Language Inference Using Neural Network and Tableau Method

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

加藤芳秀, 松原茂樹

Proceedings of the 36th Pacific Asia Conference on Language, Information and Computation Vol. - page： 430 - 437 2022

Saji A., Kato Y., Matsubara S.

Paclic Pacific Asia Conference on Language Information and Computation ( 2022 ) page： 430 - 437 2022

Language：English Publishing type：Research paper (scientific journal) Publisher：Paclic Pacific Asia Conference on Language Information and Computation

Saji et al. (2021) integrated a neural-based NLI model and a symbolic one into a framework to get the best of both worlds. This framework is based on a tableau method, which is a proof system for formal logic; however, it has a remaining issue that its theoretical limitations have not been clarified. To solve this issue, this paper formalizes the framework model-theoretically. On the basis of the formalization, we demonstrate that a certain kind of soundness holds for this framework, while the completeness does not.