|
Position |
Professor |
|
Research Field |
Informatics / Human interface and interaction, Informatics / Intelligent informatics, Informatics / Database, Informatics / Web informatics and service informatics |
|
Homepage URL |
|
|
External Link |
|
|
|
|
Graduating School 【 display / non-display 】
Graduate School 【 display / non-display 】
-
Kobe University Graduate School, Division of Science and Technology Doctor's Course Completed
- 2002.3
Campus Career 【 display / non-display 】
-
KONAN UNIVERSITY Faculty of Intelligence and Informatics 学部長
2022.4 - 2024.3
-
KONAN UNIVERSITY Faculty of Intelligence and Informatics Professor
2011.4
External Career 【 display / non-display 】
-
独立行政法人 通信総合研究所（現 情報通信研究機構）
2002.4 - 2008.3
Country：Japan
Papers 【 display / non-display 】
-
Two-Stage Fine-Tuning for Dialogue Generation with Small Community Prominent Leaders’ Philosophies Reviewed
Tetsuya Kitahata, Kazuhiro Seki, and Akiyo Nadamoto
Proceedings of The 27th International Conference on Information Integration and Web Intelligence (iiWAS2025) 427 - 442 2025.12
Authorship：Last author
-
Speech-Scenario Generation Based on the Philosophy of a Prominent Leader Within a Small Community Reviewed
Tetsuya Kitahata, Kazuhiro Seki, and Akiyo Nadamoto
Proceedings of The 36th International Conference of Database and Expert Systems Applications(DEXA 2025) 291 - 306 2025.8
Authorship：Last author
-
読み手の感情に着目したコロナ禍における流言ツイートの特徴分析 Reviewed
西岡 竜生，灘本 明代
日本データベース学会データドリブンスタディーズ 3 ( 3 ) 8pages 2025.3
Authorship：Last author
-
Amon Shimozaki, Yousuke Tsuge, Tatsuya Kitamura, Tomohiro Umetani, Akiyo Nadamoto
The DASFAA 2024 Workshop on Emerging Results in Data Science and Engineering (ERDSE 2024) 187 - 202 2025.1
Joint Work
Authorship：Last author Publisher：Springer
In recent years, the widespread use of large-scale language models, such as ChatGPT, has facilitated the generation of various documents. Moreover, numerous studies have been conducted on automatic dialogue generation using large-scale dialogue models, with accuracy improving daily. Most automatic dialogue generation targets chats, Q&A, manuals, and so on. However, automatically generating dialogues incorporating humor, such as those in Manzai scenarios, remains challenging. In this study, we explore the potential for generating dialogues that include humor by employing several existing large-scale language models. Specifically, we focus on Manzai, a form of Japanese comedic content, as a case study for humorous dialogue. For this purpose, we utilized models such as Llama2, Llama2-Chat, and ChatGPT to generate a Manzai scenario automatically. Additionally, we fine-tuned Llama2 and Llama2-Chat with various datasets to automatically generate humorous dialogues and compare the outcomes.
Review Papers (Misc) 【 display / non-display 】
-
末吉 将也, 灘本 明代
甲南大学紀要.知能情報学編 17 ( 1 ) 21 - 34 2024.7
Publisher：甲南大学知能情報学部
インターネット上には様々なニュースがあるが，そのニュースの中にはフェイクニュースが多数存在する．このようなフェイクニュースにより，ユーザはそのニュースを信じてしまい，意図せずともフェイクニュースを拡散してしまう問題がある．フェイクニュースには，一見してフェイクとわかるものとわからないものがある．また，フェイクニュースの一部分を訂正することにより，そのフェイクニュースが正しくなる場合がある．本研究では，このように部分的に誤っているフェイクニュースの誤り箇所を抽出し訂正する手法の提案を行う．具体的には，フェイクニュースの主人公に着目し，その主人公に対する「いつ」，「どこで」，「何をしたか」の構造に着目したフェイクニュースの誤り箇所を抽出し，その訂正を行う．
DOI： 10.14990/0002000467
-
Survival Information Science: Evolving Informatics for Well-Being of Humanity and Society
HASHIMOTO Takako, NADAMOTO Akiyo
TRENDS IN THE SCIENCES 27 ( 5 ) 5_77 - 5_82 2022.5
Publisher：Japan Science Support Foundation
DOI： 10.5363/tits.27.5_77
Other Link： https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-18K11443/
-
生存情報学 : 人類と社会がより良く生き延びるために進化する情報学—特集 未来社会を生き延びるための生存情報学
橋本 隆子, 灘本 明代
学術の動向 = Trends in the sciences / 『学術の動向』編集委員会 編 27 ( 5 ) 77 - 82 2022.5
Publisher：東京 : 日本学術協力財団
Other Link： https://ndlsearch.ndl.go.jp/books/R000000004-I032204116
-
Generating Funny Dialogue by Word Transformation based Phoneme Similarity
TAKAYAMA Noriyuki, KITAMURA Tatsuya, NADAMOTO Akiyo
Proceedings of the Technical Committee on Speech Communication 1 ( 1 ) n/a 2021.9
Publisher：The Technical Committee on Speech Communication,the Acoustical Society of Japan
We have proposed a manzai robot system that generates scenario and performs it automatically. In this study, we proposed a "word-mistake boke" method that replaces one syllable in a word to other character. Though there are many possible ways to generate a misspoken word, it is necessary to generate a funny one for manzai scenario. In this paper, we propose a method for generating a funny misspoken word using a word transformation method based on phoneme similarity. Furthermore，we propose an experimental method for evaluating the relationship between dialogue with a funny misspoken word and phoneme similarity.
-
吉川, 正俊, 灘本, 明代, 難波, 英嗣
情報処理学会論文誌データベース（TOD） 14 ( 2 ) i - ii 2021.4
Academic Awards Received 【 display / non-display 】
-
Highly Commended PaperAward
2016.6 EmeraldLiteratiNetwork
Akiyo Nadamoto
-
Outstanding Paper Award
2014.6 Emerald LiteratiNetwork
Akiyo Nadamoto, Yuki Hattori
-
Best paper award
2011.11 the 13th International Conference on Information Integration and Web-Based Applications & Services
Koichi Takaoka, Akiyo Nadamoto
-
Outstanding Paper Award
2011.11 Emerald
Akiyo Nadamoto, Eiji Aramaki, Takeshi Abekawa, Yohei Murakami
Grant-in-Aid for Scientific Research 【 display / non-display 】
-
User-centric Behavioral Facilitation Information in the Disaster Situation
2019.4 - 2022.3
JSPS Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research(B)
Nadamoto Akiyo
In the event of a large-scale disaster, there is a large amount of information on social networking services (SNS) that make promote or inhibit actions for readers. We called this information "behavioral facilitation information". We have investigated methods (rule-based and many deep learning methods (LSTM, BLSTM, BERT, and RoBERTa)) for automatically extracting the behavioral facilitation information. We found that RoBERTa gave the best results, and succeeded in automatically extracting behavioral facilitation information from a large amount of SNS. Furthermore, we analyzed the extracted behavioral facilitation information from the viewpoints of the sending area (disaster-stricken area and outside of the disaster-stricken area) and the senders (disaster-stricken victims and others).
The results of our research are 8 peer-reviewed journals, 10 peer-reviewed international conferences, 23 domestic conferences. -
A Study on High Quality Dataset Construction for Multitask Learning
2024.4 - 2028.3
JSPS Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research(B)
-
ツイート投稿者の感情を推測するための統合的な基盤技術に関する研究
2020.4 - 2023.3
JSPS Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research(C)
熊本忠彦
テキストベースのメッセージ交換を主とするコミュニケーション手段では，相手の表情や仕草を見たり，声を聴いたりすることができないため，メッセージの感性的側面を正確に捉えることができないことも多い．このような問題を回避するために，顔文字のような非言語表現が用いられることもあるが，顔文字がメッセージの感性的側面にどのような影響を及ぼしているのかに関しては不明な点も多い．そこで本研究では，ツイッターを対象に，顔文字が付与されることでツイートの感性的側面（ツイートを読んだ人がツイート投稿者の感情をどう受け取るか）がどのように変化するかをアンケート調査に基づいて調べ，ツイート投稿者の感情推測手法を提案する． -
Research on large-scale creative work by cooperation in crowdsourcing
2019.4 - 2023.3
JSPS Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research(B)
-
A Study on Integrated Kansei Information Mining Techniques for Individual Writers and Readers
2017.4 - 2020.3
JSPS Grants-in-Aid for Scientific Research Grant-in-Aid for Scientific Research(C)
Kumamoto Tadahiko
The purpose of this study is to extract three kinds of affective information from sentences such as tweets: emotions of the writers who wrote the sentences, impressions which readers felt from reading the sentences, and the writers' emotions which the readers estimated by reading the sentences.
First, we conduct questionnaire surveys, and analyze the relationships between tweets and emotions and impressions extracted from the tweets. And then, we formulate these relationships using multivariate analysis methods and machine learning methods, resulting in high accuracy and robustness. We have developed a method that can extract the emotions and impressions of tweets.
Note that, in this study, ten kinds of emotions of "joy, love, relief, sad, dislike, fear, anger, embarrassment, uplifting, and surprise" and eight kinds of impressions of "aggressive/unpleasant, negative, pleasant, pleasant/happy, positive, heartwarming, annoying, and scary" are used.
Other External funds procured 【 display / non-display 】
-
作り手，売り手，消費者のおいしさの表現比較分析に関する研究
2014.10 - 2016.3
浦上財団 浦上財団
Preferred joint research theme 【 display / non-display 】
-
漫才台本自動生成に基づく漫才ロボットの研究
-
SNSからの防災情報マイニング・分析
Committee Memberships 【 display / non-display 】
-
2019.7 - 2023.6 ACM SIGMOD日本支部 支部長
-
2018.6 日本データベース学会 副会長
-
2018.4 日本学術会議 連携会員
-
2017.6 - 2019.5 電子情報通信学会 データ工学研究会委員長
-
2014.6 日本データベース学会 理事
Social Activities 【 display / non-display 】
-
日本学術会議
2017.10
-
International Workshop on Information-explosion and Next Generation search(INGS2008) publication/publicity chair
2008.11
-
情報処理学会論文誌データベース編集委員
2006.4
-
振興調整費 評価委員
2005.11
-
電子通信学会データ工学専門委員
2003.4