| Tsunenori Mine | Last modified date：2024.04.26 |
Associate Professor /
Advanced Software Engineering
Department of Advanced Information Technology
Faculty of Information Science and Electrical Engineering
Department of Advanced Information Technology
Faculty of Information Science and Electrical Engineering
Graduate School
システム情報科学府 高度ICT人材教育開発センター 教育基盤研究部門
Undergraduate School
Administration Post
Other
E-Mail *Since the e-mail address is not displayed in Internet Explorer, please use another web browser:Google Chrome, safari.
Homepage
https://kyushu-u.elsevierpure.com/en/persons/tsunenori-mine
Reseacher Profiling Tool Kyushu University Pure
https://www.m.ait.kyushu-u.ac.jp/mine/index_e.html
Tsunenori Mine .
Academic Degree
Dr Engineering
Country of degree conferring institution (Overseas)
No
Field of Specialization
Artificial Intelligence, Multi-Agent Systems, Data Mining, Text Mining, Information Sharing, Information Retrieval, Information Recommendation, Personalized System
ORCID(Open Researcher and Contributor ID)
https://orcid.org/0000-0002-7462-8074
Total Priod of education and research career in the foreign country
01years00months
Outline Activities
My research interests are natural language processing, information retrieval, information extraction,
collaborative filtering, semantic Web and multi-agent systems.
collaborative filtering, semantic Web and multi-agent systems.
Research
Research Interests
Membership in Academic Society
- Discovery and Recommendation of Lesser Known, but Attractive POI (Point of Interests) for Local Tourism Revitalization
keyword : Lesser Known but Attractive POI, Local Tourism Revitalization
2024.04～2024.04. - Research and Development of Intelligent Chatbot systems
keyword : Chatbot, dialogue system, deep learning, machine learning, text mining, natural language processing
2019.07. - Prediction of Travel Time or Delay Time using Probe Data
keyword : Data Mining, Travel Time, Delay Time, Probe Data, Intelligent Transport Systems
2014.01. - Study on Open Data Mining to Proceed Government 2.0
keyword : Open Data, Data Mining, Text Mining, Government 2.0
2016.04～2022.03. - Study on Improvement of Learners Learning Performance using Comment Mining
keyword : Estimation of Learning Situation, Improvement of Learning Performance, Feedback Generation, Comment Mining
2012.04. - ItoCamLife : An Information Sharing and Recommendation Platform to Support Smart Mobility at Ito Campus, Kyushu University
keyword : Information Sharing, Information Recommendation, Smart Mobility, ItoCamLife, Data Mining
2016.06～2021.03. - Development of Reciprocal Recommendation System Toward Intelligent Job Matching
keyword : Reciprocal Recommendation, Job Matching, Personalization, MultiAgent
2011.05～2016.03. - Development of Personalized Transit Recommendation based on User Situation
keyword : Information Recommendation, Transit Recommendation, Personalization, User Context
2012.11～2021.03. - a study of effective and quantitative measurement, refinement and assesment of students' process for their project-based-learning
keyword : Project-based-learning, quantitative assessment, system development process
2010.09～2013.03. - Research on Agent-Community-based Personalized Information Retrieval
keyword : MultiAgents, Information Retrieval, Personalization, Agent Community
2007.04～2012.03. - Agent-Community-based Peer-to-Peer Information Retrieval
keyword : Peer-to-Peer, Information Retrieval, Multiagent Systems
2003.04～2012.03. - Research on Multi-agent-based Intelligent Information Retrieval Systems
keyword : Peer-to-Peer, Information Retrieval, Multiagent Systems
2004.04～2012.03.
Papers
| 1. | Shaowen Peng, Kazunari Sugiyama, Tsunenori Mine, Less is More: Removing Redundancy of Graph Convolutional Networks for Recommendation, ACM Transactions on Information Systems, https://doi.org/10.1145/3632751, 42, 3, 1-26, Article No. 85, 2023.11, [URL], While Graph Convolutional Networks (GCNs) have shown great potential in recommender systems and collaborative filtering (CF), they suffer from expensive computational complexity and poor scalability. On top of that, recent works mostly combine GCNs with other advanced algorithms which further sacrifice model efficiency and scalability. In this work, we unveil the redundancy of existing GCN-based methods in three aspects: (1) Feature redundancy. By reviewing GCNs from a spectral perspective, we show that most spectral graph features are noisy for recommendation, while stacking graph convolution layers can suppress but cannot completely remove the noisy features, which we mostly summarize from our previous work; (2) Structure redundancy. By providing a deep insight into how user/item representations are generated, we show that what makes them distinctive lies in the spectral graph features, while the core idea of GCNs (i.e., neighborhood aggregation) is not the reason making GCNs effective; and (3) Distribution redundancy. Following observations from (1), we further show that the number of required spectral features is closely related to the spectral distribution, where important information tends to be concentrated in more (fewer) spectral features on a flatter (sharper) distribution. To make important information be concentrated in as few features as possible, we sharpen the spectral distribution by increasing the node similarity without changing the original data, thereby reducing the computational cost. To remove these three kinds of redundancies, we propose a Simplified Graph Denoising Encoder (SGDE) only exploiting the top-K singular vectors without explicitly aggregating neighborhood, which significantly reduces the complexity of GCN-based methods. We further propose a scalable contrastive learning framework to alleviate data sparsity and to boost model robustness and generalization, leading to significant improvement. Extensive experiments on three real-world datasets show that our proposed SGDE not only achieves state-of-the-art but also shows higher scalability and efficiency than our previously proposed GDE as well as traditional and GCN-based CF methods.. |
| 2. | Bo Wang and Tsunenori Mine, Optimizing Upstream Representations for Out-of-Domain Detection with Supervised Contrastive Learning, 32nd ACM International Conference on Information and Knowledge Management (CIKM2023), 2585-2595, 2023.10, [URL], Out-of-Domain (OOD) text detection has attracted significant research interest. However, conventional approaches primarily employ Cross-Entropy loss during upstream encoder training and seldom focus on optimizing discriminative In-Domain (IND) and OOD representations. To fill this gap, we introduce a novel method that applies supervised contrastive learning (SCL) to IND data for upstream representation optimization. This effectively brings the embeddings of semantically similar texts together while pushing dissimilar ones further apart, leading to more compact and distinct IND representations. This optimization subsequently improves the differentiation between IND and OOD representations, thereby enhancing the detection effect in downstream tasks. To further strengthen the ability of SCL to consolidate IND embedding clusters, and to improve the generalizability of the encoder, we propose a method that generates two different variations of the same text as "views". This is achieved by applying a twice "dropped-out" on the embeddings before performing SCL. Extensive experiments indicate that our method not only outperforms state-of-the-art approaches, but also reduces the requirement for training a large 354M-parameter model down to a more efficient 110M-parameter model, highlighting its superiority in both effectiveness and computational economy.. |
| 3. | LANDY RAJAONARIVO, TSUNENORI MINE, Few-shot learning-based lesser-known POI category estimation based on syntactic and semantic information, IEEE Access, 10.1109/ACCESS.2023.3327636, 141100-141111, 2023.10, The estimation of points of interest (POI) categories is essential in several contexts, such as land use estimation, POI and itinerary recommendation in the tourism sector, and so on. Most of these approaches are based on well-known POIs and use information such as people’s mobility or check-in data. This information is not or rarely available for lesser-known POIs. However, these lesser-known POIs cannot be ignored because of this lack of information, as they may be important to local people in terms of their culture and history and worth discovering by tourists or local authorities. To address this challenge, we propose an approach based on the techniques of coupling the syntactic and semantic analysis of data via a knowledge graph using Few-shot learning (FSL) and Light Graph Convolution Network (LightGCN). The FSL technique allows us to work with very little data, which not only works with lesser-known POIs but also reduces the complexity in terms of tasks and execution time. The results show that our approach outperforms the baseline approaches and that considering the semantic aspect of the data via Linked Open Data (LOD) significantly improves the results of the approach based on the syntactic analysis alone.. |
| 4. | Ristu Saptono, Tsunenori Mine, Distribution-adapted Model For Helpful Vote Prediction, IEEE Access, 10.1109/ACCESS.2022.3225558, 10, 125194-125211, 2022.12, [URL], The number of helpful votes on a review is an essential indicator of how much impact the review has on other customers in electronic commerce. Therefore, predicting the number of helpful votes is an important task. Regression analysis and Tobit modeling are typical methods of prediction. Those methods come from the same initial assumption that the number of helpful votes follows a normal distribution on any dataset. However, the assumption is not usually confirmed, and the distribution of the helpful votes often follows other distributions. This paper proposes a framework for investigating the feasibility of building a model that predicts the number of helpful votes according to the distribution of the number of helpful votes. On top of that, considering the review age, we propose an adaptive window size sampling method to evaluate the model on review datasets sorted chronologically. The experimental results validated that the model adapting to the best approximate distribution gives a significant improvement compared to the baseline models. In addition, model evaluation using the adaptive window size sampling method has significant impacts on the performance on large datasets.. |
| 5. | Yuichi Ishikawa, Roberto Legaspi, Kei Yonekawa, Yugo Nakamura, Shigemi Ishida, Tsunenori Mine, and Yutaka Arakawa, Unsupervised Learning of Domain-Independent User Attributes, IEEE Access, 10.1109/ACCESS.2022.3220781, 10, 119649-119665, 2022.11, Learning user attributes is essential for providing users with a service. In particular, for e-commerce portals which deal in variety of goods ranging from clothes to foods to home electronics, it is especially important to learn “domain-independent” attributes such as age, gender, and personality that affect people’s behavior across various domains of daily life (e.g., clothing, eating and housing) because these attributes can be used for personalization in diverse domains their service covers. Thus far, researchers have proposed approaches to learn user representation (UR) from user-item interactions, trying to embed rich information about user attributes in UR. However, very few can learn URs that are domain-independent without confounding them with domain-specific attributes (e.g., food preferences). This could consequently undermine the former’s utility for personalizing services in other domains from which the URs are not learned. To address this, we propose an approach to learn URs that exclusively reflect domain-independent attributes. Our approach introduces a novel multi-layer RNN with two types of layers: Domain Specific Layers (DSLs) for modeling behavior in individual domains and a Domain Independent Layer (DIL) for modeling attributes that affect behavior across multiple domains. By exchanging hidden states between these layers, the RNNs implement the process of domain-independent attributes affecting domain-specific behavior and makes the DIL learn URs that capture domain-independence. Our evaluation results confirmed that the URs learned by our approach have greater utility in predicting behavior in the other domains from which these URs were not learned thereby demonstrating adaptability to various domains.. |
| 6. | Shaowen Peng, Kazunari Sugiyama and Tsunenori Mine, SVD-GCN: A Simplified Graph Convolution Paradigm for Recommendation, 31st ACM International Conference on Information and Knowledge Management (CIKM2022), https://doi.org/10.1145/3511808.3557462, 1625-1634, 2022.10. |
| 7. | Shaowen Peng, Kazunari Sugiyama and Tsunenori Mine, Less is More: Reweighting Important Spectral Graph Features for Recommendation, ACM SIGIR 2022, 1273-1282, 2022.07, [URL]. |
| 8. | Menna Fateen, Kyouhei Ueno, Tsunenori Mine, An Improved Model to Predict Student Performance Using Teacher Observation Reports, 29TH INTERNATIONAL CONFERENCE ON COMPUTERS IN EDUCATION 2021, 1, 31-40, 2021.11. |
| 9. | Jihed Makhlouf, Tsunenori Mine, Analysis of Click-Stream Data to Predict STEM Careers from Student Usage of an Intelligent Tutoring System, Journal of Educational Data Mining, 10.5281/zenodo.4008050, 12, 2, 1-18, 2020.08, [URL]. |
| 10. | Jihed Makhlouf, Tsunenori Mine, Prediction Models for Automatic Assessment to Student Free-written Comments, CSEDU2020, 10.5220/0009580300770086, 77-86, 2020.05, [URL]. |
| 11. | Jihed Makhlouf, Tsunenori Mine, Investigating Reading Behaviors within Student Reading Sessions to Predict their Performance, LAK19 Data Challenge International Workshop on Predicting Performance Based on the Analysis of Reading Behavior: A Data Challenge 2019, 2019.03, [URL]. |
| 12. | Jihed Makhlouf, Tsunenori Mine, Investigating How School-Aggregated Data Can Influence in Predicting STEM Career from Student usage of an Intelligent Tutoring System, EDM 2018 Workshop on Scientific Findings from the ASSISTments Longitudinal Data (2018), 2018.07, [URL]. |
| 13. | Jihed Makhlouf, Tsunenori Mine, Predicting if students will pursue a STEM career using School-Aggregated Data from their usage of an Intelligent Tutoring System, EDM(Educational Data Mining) 2018, 552-555, 2018.07, [URL]. |
| 14. | Shaymaa Sorour, Tsunenori Mine, Kazumasa Goda, Sachio Hirokawa, A Predictive Model to Evaluate Students Performance, Journal of Information Processing, 10.2197/ipsjip.23.192, 23, 2, 192-201, 2015.02, [URL]. |
| 15. | Tsunenori Mine, Hirotake Kobayashi, Applying user feedback and query learning methods to multiple communities, Proceedings of PRIMA 2009, 276-291, 2009.12. |
| 16. | Tsunenori Mine, Kosaku Kimura, Satoshi Amamiya, Ken'ichi Takahashi, Makoto Amamiya, Agent-Community-Network-Based Secure Collaboration Support System, Agent-Based Technologies and Applications for Enterprise Interoperability -- International Workshops, ATOP 2005, Utrecht, The Netherlands, July 25-26, 2005, and ATOP 2008, Estoril, Portugal, May 12-13, 2008, Revised Selected Papers --, Springer, 234-255, LNBIP 25, 2009.05. |
| 17. | Tsunenori Mine, Akihiro Kogo, Satoshi Amamiya, Makoto Amamiya, Refinement of the ACP2P by Sharing User-Feedbacks and Learning Query-Responder-Agent-Relationships , The 8th International Conference on Autonomous Agents and MultiAgent Systems, 1341-1342, 2009.05. |
| 18. | Haibo Yu, Tsunenori Mine, Makoto Amamiya, Agent-Community-based P2P semantic MyPortal information retrieval system architecture, Journal of Embedded Computing (Selected papers of EUC2005), IOS Press, 3, 1, 63-75, 2009.01. |
| 19. | Tsunenori Mine, Akihiro Kogo, Makoto Amamiya, Agent-Community-Based Peer-to-Peer Information Retrieval and Its Evaluation, Systems and Computers in Japan, Wiley Periodicals, Inc., A Wiley Company,, vol.37, no.13, pp.1-10, 2006.11. |
| 20. | Tsunenori Mine, Daisuke Matsuno, Akihiro Kogo, Makoto Amamiya, Design and Implementation of Agent Community based Peer-to-Peer Information Retrieval Method, Eighth International Workshop CIA 2004 on Cooperative Information Agents, 3191, 31-46, LNAI volume 3191 Klusch, M.; Ossowski, S.; Kashyap, V.; Unland, R. (Eds.) pp. 31-46, 2004.09. |
| 21. | T. Mine, T. Shoudai, A. Suganuma, Automatic Exercise Generator with Tagged Documents Considering Lerner's Performance, Proceedings of the WebNet2000, pp.779-780, 2000.11. |
| 22. | T. Mine, A. Suganuma, T. Shoudai, The Design and Implementation of Automatic Exercise Generator with Tagged Documents based on the Intelligence of Students:AEGIS, Proceedings of the ICCE/ICCAI 2000, pp.651-658, 2000.11. |
| 23. | T. Shoudai, A. Suganuma, T. Mine, AEGIS: Automatic Exercise Generator with Tagged Documents based on the Intelligence of the Students, Proceedings of the Fourth Joint Conference on Knowledge-Based Software Engineering (JCKBSE) 2000, 62, 311-314, pp.311--314, 2000.09. |
| 24. | G. Zhong, T. Mine, T. Helmy, M. Amamiya, The Design and Application of the KODAMA System, Proceedings of the Fourth Joint Conference on Knowledge-Based Software Engineering (JCKBSE) 2000, 62, 43-50, pp.43--50, 2000.09. |
| 25. | T. Helmy, T. Mine, M. Amamiya, Adaptive exploiting User Profile and Interpretation Policy for Searching and Browsing the Web on KODAMA system, Proceedings of the 2nd International Workshop on Natural Language and Information Systems(NLIS 2000),which is one of Eleventh International Workshops on Database and Expert Systems Applications, 120-124, pp.120--124, 2000.09. |
| 26. | T. Mine, M. Higashi, M. Amamiya, Case Frame Acquisition and Verb Sense Disambiguation on a Large Scale Electronic Dictionary, Proc. of NLPRS(Natural Language Processing Pacific Rim Sympo.)'97 in Phuket, Thailand, pp.221-226, 1997.12. |
| 27. | Tsunenori Mine, Daisuke Matsuno, Akihiro Kogo, Makoto Amamiya, ACP2P : Agent Community based Peer-to-Peer Information Retrieval, the Third International Workshop on Agents and Peer-to-Peer Computing (AP2PC) (joint Workshop of AAMAS), Agents and Peer-to-Peer Computing 2004 LNCS Volume Editors Gianluca Moro, Sonia Bergamaschi and Karl Aberer to appear. |
Presentations
| 1. | Takuya Kawatani, Eisuk Itoh, Sachio Hirokawa, Tsunenori Mine, Location does not always determine sudden braking, The 22nd IEEE International Conference on Intelligent Transportation Systems, 2019.10, [URL]. |
| 2. | Shaymaa Sorour, Tsunenori Mine, Kazumasa Goda, Correlation of Topic Model and Student Grades Using Comment Data Mining, SIGCSE2015, 2015.03, [URL]. |
| 3. | Shaymaa Sorour, Tsunenori Mine, Sachio Hirokawa, Kazumasa Goda, Predicting Students' grades based on free style Comments Data by Artificial Neural Network, The 44th Annual Frontiers in Education (FIE) Conference, 2014.10, [URL]. |
| 4. | Shaymaa Sorour, Tsunenori Mine, Sachio Hirokawa, Kazumasa Goda, Efficiency of LSA and K-means in predicting Students' academic performance based on comments data, The 6th International Conference on Computer Supported Education (CSEDU2014), 2014.04. |
| 5. | Hiroyuki Nakamura, Tsunenori Mine, Dealing with Bus Delay and User History for Personalized Transportation Recommendation, The 2014 International Conference on Computational Science and Computational Intelligence, 2014.03. |
| 6. | Kazumasa Goda, Sachio Hirokawa, Tsunenori Mine, Correlation of Grade Prediction Performance and Validity of Self-Evaluation Comments, SIGITE, 2013.10. |
- IEEE
- ACM
- IPSJ
- IEICE
- JSAI
- NLP
Educational
Educational Activities
Programming I (to the 2nd year undergraduate students),
and Advanced Design of Software (to graduates).
Fundamentals of EECS I
Other Educational Activities
and Advanced Design of Software (to graduates).
Fundamentals of EECS I
- 2023.09.
- 2024.08.
- 2023.08.
- 2023.07.
- 2022.12.
- 2022.08.
- 2023.03.
- 2022.01.
- 2021.07.
- 2021.02.
- 2021.06.
- 2018.03.
- 2017.08.
- 2017.04.
- 2016.09.
- 2012.08.
- 2006.07.
Unauthorized reprint of the contents of this database is prohibited.