Award type：International academic award (Japan or overseas)  Country：United States

The 2017 ASSISTments Longitudinal Data Mining Competitionで2位を獲得．このCompetitionで取った方法については，下記論文で発表．
Jihed Makhlouf and 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)（招待論文）
Jihed Makhlouf and Tsunenori Mine, "Predicting if students will pursue a STEM career using School-Aggregated Data from their usage of an Intelligent Tutoring System," EDM 2018 poster, 552-555, (2018)

山口晃平, 峯恒憲. "地域問題レポートにおける担当課の自動分類"

佐野優太, 峯恒憲，"経験的パターンと機械学習手法を利用した現状，要望の抽出"，合同エージェントワー
クショップ＆シンポジウム2016(JAWS2016), (2016)（奨励論文賞）

ICAAIの中の一つの会議(LTLE 2016)で発表した下記論文に対する賞
Building an Interpretable Model of Predicting Student Performance Using Comment Data Mining
Shaymaa Sorour, Tsunenori Mine (Kyushu University, Japan)
http://www.iaiai.org/conference/aai2016/award/

■受賞理由 「デジタルプラクティスへの査読貢献」

1991年10月に発行された情報処理学会論文誌に掲載された論文に対する優秀賞

Language：English   Publishing type：Research paper (international conference proceedings)  

DOI： https://doi.org/10.1145/3583780.36150

Language：English   Publishing type：Research paper (scientific journal)   Publisher：IEEE Access  

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.

DOI： 10.1109/ACCESS.2022.3225558

Web of Science

Scopus

researchmap

Other Link： https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9966597

Language：English   Publishing type：Research paper (scientific journal)   Publisher：IEEE Access  

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.

DOI： 10.1109/ACCESS.2022.3220781

Web of Science

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：ACM  

With the tremendous success of Graph Convolutional Networks (GCNs), they have been widely applied to recommender systems and have shown promising performance. However, most GCN-based methods rigorously stick to a common GCN learning paradigm and suffer from two limitations: (1) the limited scalability due to the high computational cost and slow training convergence; (2) the notorious over-smoothing issue which reduces performance as stacking graph convolution layers. We argue that the above limitations are due to the lack of a deep understanding of GCN-based methods. To this end, we first investigate what design makes GCN effective for recommendation. By simplifying LightGCN, we show the close connection between GCN-based and low-rank methods such as Singular Value Decomposition (SVD) and Matrix Factorization (MF), where stacking graph convolution layers is to learn a low-rank representation by emphasizing (suppressing) components with larger (smaller) singular values. Based on this observation, we replace the core design of GCN-based methods with a flexible truncated SVD and propose a simplified GCN learning paradigm dubbed SVD-GCN, which only exploits K-largest singular vectors for recommendation. To alleviate the over-smoothing issue, we propose a renormalization trick to adjust the singular value gap, resulting in significant improvement. Extensive experiments on three real-world datasets show that our proposed SVD-GCN not only significantly outperforms state-of-the-arts but also achieves over 100x and 10x speedups over LightGCN and MF, respectively.

DOI： 10.1145/3511808.3557462

Web of Science

Scopus

researchmap

Other Link： https://dl.acm.org/doi/pdf/10.1145/3511808.3557462

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：Springer  

Performance prediction models have been proposed countless times due to the benefits that they can provide to educational stakeholders. While many factors have been taken into account when predicting student performance, teachers’ assessment or observation reports have not been commonly used. A teacher’s assessment is a fundamental part of the educational process and has a direct impact on students’ success. In this study, we analyze the topics, and psychological features in teachers’ daily written reports and apply them to the student performance prediction model. Experimental results show the capability of this approach in contributing to the accuracy of performance prediction models.

DOI： 10.1007/978-3-031-11644-5_58

Web of Science

Scopus

researchmap

Other Link： https://www.researchgate.net/profile/Menna-Fateen/publication/362275925_Extraction_of_Useful_Observational_Features_from_Teacher_Reports_for_Student_Performance_Prediction/links/659617a92468df72d3f953ba/Extraction-of-Useful-Observational-Features-from-Teacher-Reports-for-Student-Performance-Prediction.pdf

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：ACM  

As much as Graph Convolutional Networks (GCNs) have shown tremendous success in recommender systems and collaborative filtering (CF), the mechanism of how they, especially the core components (\textiti.e., neighborhood aggregation) contribute to recommendation has not been well studied. To unveil the effectiveness of GCNs for recommendation, we first analyze them in a spectral perspective and discover two important findings: (1) only a small portion of spectral graph features that emphasize the neighborhood smoothness and difference contribute to the recommendation accuracy, whereas most graph information can be considered as noise that even reduces the performance, and (2) repetition of the neighborhood aggregation emphasizes smoothed features and filters out noise information in an ineffective way. Based on the two findings above, we propose a new GCN learning scheme for recommendation by replacing neihgborhood aggregation with a simple yet effective Graph Denoising Encoder (GDE), which acts as a band pass filter to capture important graph features. We show that our proposed method alleviates the over-smoothing and is comparable to an indefinite-layer GCN that can take any-hop neighborhood into consideration. Finally, we dynamically adjust the gradients over the negative samples to expedite model training without introducing additional complexity. Extensive experiments on five real-world datasets show that our proposed method not only outperforms state-of-the-arts but also achieves 12x speedup over LightGCN.

DOI： 10.1145/3477495.3532014

Web of Science

Scopus

researchmap

Other Link： https://dl.acm.org/doi/pdf/10.1145/3477495.3532014

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English  

DOI： 10.5281/zenodo.4008050

Other Link： https://jedm.educationaldatamining.org/index.php/JEDM/article/view/410

Language：English  

DOI： 10.5220/0009580300770086

Other Link： https://www.insticc.org/Primoris/Resources/PaperPdf.ashx?idPaper=95803

Language：English  

Other Link： https://sites.google.com/view/lak19datachallenge

Language：Japanese  

Language：English  

Other Link： https://drive.google.com/file/d/1XS1spxOdbFkFfRsjTAo-r0_hTTe6mkTO/view

Language：English  

Other Link： https://educationaldatamining.org/files/conferences/EDM2018/EDM2018_Preface_TOC_Proceedings.pdf

Language：Japanese   Publishing type：Research paper (scientific journal)  

DOI： http://doi.org/10.1527/tjsai.AG16-B

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.2197/ipsjip.23.192

Other Link： https://www.jstage.jst.go.jp/article/ipsjjip/23/2/23_192/_pdf

Language：Japanese   Publishing type：Research paper (scientific journal)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (scientific journal)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：Japanese   Publishing type：Research paper (scientific journal)  

Language：English   Publishing type：Research paper (scientific journal)  

Language：English   Publishing type：Research paper (scientific journal)  

Language：Japanese   Publishing type：Research paper (scientific journal)  

Language：Japanese   Publishing type：Research paper (scientific journal)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Tsunenori Mine, Daisuke Matsuno, Akihiro Kogo, Makoto Amamiya: Design and Implementation of Agent Community based Peer-to-Peer Information Retrieval Method, CIA 2004, LNAI 3191, pp. 31-46, September 27, 2004

Language：Japanese   Publishing type：Research paper (other academic)  

Language：Japanese   Publishing type：Research paper (scientific journal)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Tsunenori Mine, Daisuke Matsuno, Akihiro Kogo, Makoto Amamiya: ACP2P : Agent Community based Peer-to-Peer Information Retrieval (Draft) , the Third International Workshop on Agents and Peer-to-Peer Computing (AP2PC)

Authorship：Last author   Language：English   Publishing type：Research paper (other academic)  

DOI： https://doi.org/10.1145/3718488

Language：English   Publishing type：Research paper (scientific journal)  

DOI： https://doi.org/10.1145/3707204

Language：English   Publishing type：Research paper (international conference proceedings)  

Authorship：Last author   Language：English   Publishing type：Research paper (scientific journal)   Publisher：IEEE Access  

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.

DOI： 10.1109/ACCESS.2024.3508747

Language：English   Publishing type：Research paper (other academic)  

DOI： https://doi.org/10.48550/arXiv.2409.20042

Language：English   Publisher：International Institute of Applied Informatics  

DOI： 10.52731/liir.v005.302

CiNii Research

Publisher：Applied System Innovation  

Capturing patients’ medical histories significantly influences clinical decisions. Errors in this process lead to clinical errors, which increase costs and dissatisfaction among physicians and patients. Physicians in developing countries are overloaded with patients and cannot always follow the proper history-taking procedure. The challenges have been acknowledged; however, a comprehensive understanding of the status and the remedies has remained unexplored. This paper aims to investigate the workload, history-taking challenges, and the willingness of the physicians to accept digital solutions. A cross-sectional online survey was conducted on 104 physicians across Bangladesh, featuring 22 questions regarding their professional environment, workload, digitization status of health records, challenges in history taking, and attitudes toward adopting digital solutions for managing patient histories; 92.67% of the physicians face high workloads, 88.46% struggle in medical history taking, and only 4.81% use digital medical records. About 70% struggle to complete the necessary history-taking steps, emphasizing the urgent need for solutions. A novel visualization system, the Smart Health Gantt Chart (SHGC), has been introduced for their instant feedback. A total of 93.27% of physicians expressed their willingness to use such a system. The proposed SHGC has the potential to enhance healthcare efficiency in developing nations, benefit physicians, and improve patient-centered care.

DOI： 10.3390/asi7040069

Web of Science

Scopus

Language：English   Publishing type：Research paper (scientific journal)   Publisher：IEEE Access  

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.

Language：English   Publishing type：Research paper (international conference proceedings)  

Other Link： https://ieeecompsac.computer.org/2024/

Publisher：Journal of Soft Computing and Data Mining  

The metaverse is a virtual world with immense potential and is increasingly adopted as an online learning platform for students. Students can build virtual worlds that simulate real-life experiences using interactive activities. In this study, we implemented student learning in the virtual world and evaluated the effects of students’ achievements. Online learning formats were conducted in Second Life (SL) technology for one semester by three groups (classes), and the data was collected, including several categories, i.e., project work, interaction logs, feedback forms, online quizzes, and attendance. We then employed the robust Multiple Linear Regression (MLR) technique to classify student performance during their virtual learning experience. The results showed significant improvements in student learning across several categories, with substantial variations observed among the different classes. Interestingly, we also found that students in each class had unique experiences and faced distinct challenges in the virtual learning environment.

DOI： 10.30880/jscdm.2024.05.01.006

Scopus

Language：English   Publishing type：Research paper (other academic)  

Language：Japanese   Publishing type：Research paper (bulletin of university, research institution)  

Language：English   Publishing type：Research paper (scientific journal)   Publisher：Sustainability (Switzerland)  

In the pursuit of Universal Health Coverage (UHC), developing countries with limited resources face significant hurdles. Traditional top-down approaches, spearheaded by governments, often entail lengthy planning periods and substantial costs to reach underserved communities. This paper proposes a bottom-up, cost-effective, and preventive approach aimed at empowering local human resources and using advanced technology. We call it the Micro Healthcare Entrepreneurship (MHE) model. The MHE model comprises four primary components: (1) utilization of local human resources, (2) disruptive healthcare technology, (3) continuous training, and (4) available funds. Extensive testing of the digital health and social business elements across diverse contexts underscores their efficacy. Furthermore, the viability of social business funds has been established. To solicit feedback and refine the MHE concept, three workshops, conducted in phases, engaged 60 community healthcare service providers. This paper elucidates both the potential advantages and the challenges inherent in implementing the MHE model. By shifting the paradigm toward grassroots empowerment and leveraging existing community assets, this approach offers promise in advancing healthcare accessibility and sustainability in resource-constrained settings to achieve UHC.

DOI： 10.3390/su16062268

Web of Science

Scopus

Other Link： https://www.mdpi.com/2071-1050/16/6/2268

Language：English   Publishing type：Research paper (scientific journal)  

Sequential recommendation has attracted a lot of attention from both academia and industry, however the privacy risks associated to gathering and transferring users' personal interaction data are often underestimated or ignored. Existing privacy-preserving studies are mainly applied to traditional collaborative filtering or matrix factorization rather than sequential recommendation. Moreover, these studies are mostly based on differential privacy or federated learning, which often leads to significant performance degradation, or has high requirements for communication. In this work, we address privacy-preserving from a different perspective. Unlike existing research, we capture collaborative signals of neighbor interaction sequences and directly inject indistinguishable items into the target sequence before the recommendation process begins, thereby increasing the perplexity of the target sequence. Even if the target interaction sequence is obtained by attackers, it is difficult to discern which ones are the actual user interaction records. To achieve this goal, we propose a CoLlaborative-cOnfusion seqUential recommenDer, namely CLOUD, which incorporates a collaborative confusion mechanism to edit the raw interaction sequences before conducting recommendation. Specifically, CLOUD first calculates the similarity between the target interaction sequence and other neighbor sequences to find similar sequences. Then, CLOUD considers the shared representation of the target sequence and similar sequences to determine the operation to be performed: keep, delete, or insert. We design a copy mechanism to make items from similar sequences have a higher probability to be inserted into the target sequence. Finally, the modified sequence is used to train the recommender and predict the next item.

DOI： https://doi.org/10.48550/arXiv.2401.04423

Language：English   Publishing type：Research paper (international conference proceedings)  

As human-computer interaction continues to evolve, the importance of emotion recognition in conversation is becoming more apparent. For applications like chatbots to provide more human-like responses, it is essential for machines to understand the emotions embedded in conversations. Although most of the recent research has focused on extracting contextual information from conversations, the subtleties of emotion shifts (ES) within local conversations are often overlooked. However, acquiring knowledge about ES between interactions can reduce the rate of emotion recognition errors in dialogues with fluctuating emotions. As a solution for ES detection, we define ES between the same speaker as well as between different speakers using Emotion Recognition in Conversation (ERC) datasets. We propose a novel multi-task learning model, called MtlERC-ES, which identifies three tasks simultaneously: Emotion Recognition in Conversation (ERC), Emotion Shift (ES), and Sentiment Classification (SC). Our approach provides high-quality performance on the ERC task, consistently ranking among the top performers across multiple datasets. Our approach also demonstrates the effectiveness of custom ES tasks.

Other Link： https://aclanthology.org/2023.paclic-1.26/

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：ArabicNLP 2023 - 1st Arabic Natural Language Processing Conference, Proceedings  

Delegating short answer grading to automated systems enhances efficiency, giving teachers more time for vital human-centered aspects of education. Studies in automatic short answer grading (ASAG) approach the problem from instance-based or reference-based perspectives. Recent studies have favored instance-based methods, but they demand substantial data for training, which is often scarce in classroom settings. This study compares both approaches using an Arabic ASAG dataset. We employ in-context meta-learning for instance-based and semantic score-based similarity for reference-based grading. Results show both methods outperform a baseline and occasionally even surpass human raters when grading unseen answers. Notably, the semantic score-based similarity approach excels in zero-shot settings, outperforming in-context meta-learning. Our work contributes insights to Arabic ASAG and introduces a prompt category classification model, leveraging GPT3.5 to augment Arabic data for improved performance.

DOI： 10.18653/v1/2023.arabicnlp-1.28

Scopus

researchmap

Other Link： https://aclanthology.org/2023.arabicnlp-1.28/

Publishing type：Research paper (international conference proceedings)   Publisher：2023 14th International Conference on Mobile Computing and Ubiquitous Network, ICMU 2023  

Chatbots are widely used in education, medicine, business, and other fields, among which natural language processing such as ChatGPT, especially chatbots based on large-scale language models (LLMs), are expected to flexibly answer users' complex questions. In this study, we present iTOChat, a mobile bot that uses ChatGPT to recommend departments and open campus events to students visiting Kyushu University, an open campus of the School of Engineering, by inputting students' personal information such as hobbies and future plans. Although iTOChat is expected to make accurate recommendations to students, this is not easy because ChatGPT does not contain detailed information about engineering departments and open campus events. Therefore, we developed a method to supplement the missing information in ChatGPT by embedding reasons and conditions for recommending departments in the ChatGPT prompts. This paper describes the features and objectives of the developed system and evaluates the satisfaction of students who used iTOChat at an open campus.

DOI： 10.23919/ICMU58504.2023.10412254

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/icmu/icmu2023.html#EguchiNSJMM23

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

<p>In 2014, the above paper entitled ‘Quasi-Linear Support Vector Machine for Nonlinear Classification’ was published by Zhou, et al. [1]. They proposed a quasi-linear kernel function for support vector machine (SVM). However, in this letter, we point out that this proposed kernel function is a part of multiple kernel functions generated by well-known multiple kernel learning which is proposed by Bach, et al. [2] in 2004. Since then, there have been a lot of related papers on multiple kernel learning with several applications [3]. This letter verifies that the main kernel function proposed by Zhou, et al. [1] can be derived using multiple kernel learning algorithms [3]. In the kernel construction, Zhou, et al. [1] used Gaussian kernels, but the multiple kernel learning had already discussed the locality of additive Gaussian kernels or other kernels in the framework [4], [5]. Especially additive Gaussian or other kernels were discussed in tutorial at major international conference ECCV2012 [6]. The authors did not discuss these matters.</p>

DOI： 10.1587/transfun.2022eal2051

Web of Science

Scopus

CiNii Research

researchmap

Publishing type：Research paper (international conference proceedings)   Publisher：Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)  

Automatic short answer scoring (ASAS) has received considerable attention in the field of education. However, existing methods typically treat ASAS as a standard text classification problem, following conventional pre-training or fine-tuning procedures. These approaches often generate embedding spaces that lack clear boundaries, resulting in overlapping representations for answers of different scores. To address this issue, we introduce a novel metric learning (MeL)-based pre-training method for answer representation optimization. This strategy encourages the clustering of similar representations while pushing dissimilar ones apart, thereby facilitating the formation of a more coherent same-score and distinct different-score answer embedding space. To fully exploit the potential of MeL, we define two types of answer similarities based on scores and rubrics, providing accurate supervised signals for improved training. Extensive experiments on thirteen short answer questions show that our method, even when paired with a simple linear model for downstream scoring, significantly outperforms prior ASAS methods in both scoring accuracy and efficiency.

DOI： 10.1007/978-981-99-7022-3_21

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/pricai/pricai2023-2.html#WangDIM23

Publishing type：Research paper (international conference proceedings)   Publisher：ISKE 2023 - 18th International Conference on Intelligent Systems and Knowledge Engineering  

Local events are important for the vitality and dynamism of a city. However, some events or sites are lesser-known or only known by the local community. The information about them is then limited or not available. To address this issue, we propose an approach to automatically generate knowledge graphs from information posted on social networks and data on human mobility. This graph is generated from the tweets using the syntactic relationships between the verbs, nouns, and adjectives. Its exploitation would allow to estimate the categories of lesser-known Points of Interest (POIs) or events. It can also allow to generate sentences and to answer questions that tourists may ask. The specificity of our approach lies in the Graph Convolutional Network (GCN)-based encoding of POIs with a graph of verbs, nouns, and adjectives and the estimation of POI categories using the proposed encoding. Experimental results show that our approach performs better than approaches in the literature.

DOI： 10.1109/ISKE60036.2023.10481229

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/iske/iske2023.html#RajaonarivoMA23

Publishing type：Research paper (international conference proceedings)   Publisher：Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)  

Understanding user emotion is essential for Human-AI Interaction (HAI). Thus far, many approaches have been studied to recognize emotion from signals of various physiological modalities such as cardiac activity and skin conductance. However, little attention has been paid to the fact that physiological signals are influenced by and reflect various factors that have little or no association with emotion. While emotion is a cross-modal factor that triggers responses across multiple physiological modalities, features used in existing approaches also reflect modality-specific factors that affect only a single modality and have little association with emotion. To address this, we propose an approach to extract features that exclusively reflect cross-modal factors from multimodal physiological signals. Our approach introduces a multilayer RNN with two types of layers: multiple Modality-Specific Layers (MSLs) for modeling physiological activity in individual modalities and a single Cross-Modal Layer (CML) for modeling the process by which emotion affects physiological activity. By having all MSLs update their hidden states using the CML hidden states, our RNN causes the CML to learn cross-modal factors. Using real physiological signals, we confirmed that the features extracted by our RNN reflected emotions to a significantly greater extent than the features of existing approaches.

DOI： 10.1007/978-981-99-7019-3_40

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/pricai/pricai2023-1.html#IshikawaKNNIMA22

Language：English   Publishing type：Research paper (international conference proceedings)  

Other Link： https://www.wi-iat.com/wi-iat2023/index.html

Language：English   Publishing type：Research paper (international conference proceedings)  

Publishing type：Research paper (international conference proceedings)   Publisher：Communications in Computer and Information Science  

Student performance prediction continues to be a focus of research in educational data mining due to its many potential benefits. While teachers’ assessment reports are a crucial part of the educational process, they have not been commonly used in performance prediction. We propose a model that uses similarity learning as an embedding-enhancing technique. Results outperform earlier research with an average accuracy of 73% for detecting strong performance.

DOI： 10.1007/978-3-031-36336-8_111

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/aied/aied2023p.html#FateenM23

Language：English   Publishing type：Research paper (international conference proceedings)  

Other Link： https://www.wi-iat.com/wi-iat2022/index.html

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

Since the effectiveness of a method depends on various contexts and varies with time, the timely selection from among many recommendation methods appropriate to contexts has become a critical issue for EC vendors. In this paper, we propose a practical recommendation system that continuously selects recommendation methods suitable for contexts and corresponding to temporal change using a multi-armed bandit policy. Our proposed system quickly detects changes in the effectiveness of recommendation methods using aggressive exploration. The experimental results using a real-world dataset collected from an EC site demonstrate that the proposed system outperforms baseline methods.

DOI： 10.14923/transinfj.2021sgp0002

CiNii Research

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Other Link： https://hal.science/hal-04445624/document

Language：English   Publishing type：Research paper (international conference proceedings)  

Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

In many cram schools, instructors write reports on students' progress after each class. The generation of these reports is a heavy burden for instructors, and there is a need to reduce this burden. Therefore, in this paper, we propose a system that automatically generates a student learning status reports. Students' learning status is often evaluated from several specific items, and rule-based sentence generation can be considered for those items. However, since viewpoints other than the specific items are often incorporated into the report document, a keyword-based sentence generation function is required to incorporate expressions that are difficult to be generated by the rule-base methods. Here we consider two keyword-based methods: the Sequence-to-Sequence-based method, which learns the correspondence between keywords and sentences, and the Information Retrieval-based method, which directly retrieves and reuses past reports. In this paper, we compare and evaluate the two methods and implement the model with better performance into our report generation system. We evaluated the two methods based on actual data of about 200,000 reports written by instructors, and confirmed that the Seq2Seq-based model with Attention had the best performance, and was able to generate more accurate sentences by learning positive and negative expressions separately.

DOI： 10.1109/IIAIAAI55812.2022.00019

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/iiaiaai/iiaiaai2022.html#KobashiM22

Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

In-house training includes opportunities for employees to reflect on the training through the writing of written reflections. Its purpose is to create opportunities to look back and think about "how the content learned in the training could be utilised in subsequent work,"and to allow employees to discover their own issues and set future action goals. In addition, it is possible to measure the effectiveness of the training, and the planners can also discover issues in the training itself, such as the appropriateness of the content, which is useful for improving future training programs. However, there has been little research on these employee reflection texts, especially text analysis using natural language processing (NLP), and there is a need for methods to find useful uses for the texts. By taking a NLP approach to the accumulated data of reflective statements, it is possible to estimate the degree of understanding of the training content and to grasp the degree of growth of employees over time. In addition, there is a potential for proposing training content suited to each employee and improving the efficiency of training as a further development of this approach. As a first step towards the realization of these possibilities, this study proposes automatic modality estimation methods focusing on sentence-final expressions in retrospective self-reflection sentences. Modality in Japanese includes the speaker's subjective semantic content, which indicates the speaker's judgement of facts and his/her attitude towards speech and communication to the listener. Therefore, we believe that the modality part of self-reflection sentences on training texts contains important information that leads to the understanding of each employee's behavior and intentions. Experimental results show that the proposed methods enables modality estimation with higher accuracy than manually formulated rules based on sentence structure. The results also show the possibility of developing the proposed methods to be applied to a deeper analysis approach to reflection sentences.

DOI： 10.1109/IIAIAAI55812.2022.00032

Scopus

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/iiaiaai/iiaiaai2022.html#YamadaM22

Language：English   Publishing type：Research paper (international conference proceedings)  

Other Link： https://www.sigapp.org/sac/sac2022/index.html

Language：English   Publisher：Institute of Industrial Science The University of Tokyo  

<p>Detecting and predicting abnormal driving behaviors such as sudden braking counts for preventing traffic accidents. Keeping distances from other vehicles is considered one of the most common situations where sudden braking occurs. However, to detect distances, state-of-the-art 3D object detection usually requires advanced sensors such as LiDAR. In this work, we use a monocular visual odometry system to estimate distances between vehicles with only a common drive recorder. As a preparation for predicting sudden braking events, we train a mining model to detect them with the estimated distance information along with probe data. The experiments show that with the estimated distances, the model gains an improved performance on braking detection.</p>

DOI： 10.11188/seisankenkyu.74.129

CiNii Research

Language：English   Publishing type：Research paper (international conference proceedings)  

Other Link： https://www.wi-iat.com/wi-iat2021/

Language：English   Publishing type：Research paper (international conference proceedings)  

Other Link： https://sites.google.com/view/ieeeica2021/home

Language：English   Publishing type：Research paper (scientific journal)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Other Link： https://ieeexplore.ieee.org/abstract/document/9638959

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English  

DOI： https://doi.org/10.52731/ijscai.v5.i1.575

Language：English   Publishing type：Research paper (international conference proceedings)  

Other Link： https://eric.ed.gov/?id=ED615587

Language：English   Publishing type：Research paper (scientific journal)  

DOI： https://doi.org/10.1007/s13177-021-00253-6

Language：English   Publishing type：Research paper (scientific journal)  

DOI： https://doi.org/10.1007/s13177-021-00251-8

Language：English   Publishing type：Research paper (scientific journal)  

DOI： https://doi.org/10.11188/seisankenkyu.73.131

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English  

DOI： 10.1109/QRS-C5114.2020.00026

Other Link： https://qrs20.techconf.org/QRSC2020_FULL/#!/home

Language：English  

Other Link： http://wi2020.vcrab.com.au/index.html

Language：Japanese  

Language：Others   Publishing type：Research paper (other academic)  

In this paper, the authors present SMArch, an inter-organizational software architecture for smart mobility, which can be defined as services coordinating multi-modal mobility. The architecture is extended from the Japanese system architecture for intelligent transportation systems (ITS), which is defined in 1999 and used for a reference giving a shared vision among stakeholders involved in ITS. Although the Japanese ITS architecture focuses road traffic only, SMArch focuses any mobility including not only road vehicles but also airplanes and ships. Moreover, SMArch introduces the service path concept, handles the itinerary as a series of mobility services and staying services such as lodging, dining and parking. Service coordinating agents present service path candidates satisfying the given traveler's requirements. The agents and other services interact to search for such service path candidates with the service path query language defined based on the service path concept and the information model defined in the SMArch.

DOI： 10.1145/3436829.3436868

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English  

Other Link： https://icce2020.apsce.net/

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English  

Other Link： https://sites.google.com/view/lak20datachallenge

Language：English   Publishing type：Research paper (scientific journal)  

This paper addresses the problem of predicting bus travel time over unstable intervals between two adjacent bus stops using two types of machine learning techniques: ANN and SVR methods. Our model considers the variability of travel time because the travel time is often influenced by stochastic factors, which increase the variance of travel time over an interval between inter-time periods. The factors also affect the variance of the travel time over the interval at the same time period between inter-days. In addition, the factors show some correlation of travel time over the interval between time periods in a day. The performance of the proposed model is validated with real bus probe data collected from November 21st to December 20th, 2013, provided by Nishitetsu Bus Company, Fukuoka, Japan. We demonstrated the impact of two types of input variables for the prediction in off- and on-peak (rush hour) periods. The results show that the two types of inputs can effectively improve the prediction accuracy. Moreover, we compared the proposed method with our previous methods. The experimental results show the validity of our proposed method.

DOI： 10.1007/s13177-018-0169-3

Language：English   Publishing type：Research paper (other academic)  

Food quality inspection is an essential factor in our daily lives. Food inspection is analyzing heterogeneous food data from different sources for perception, recognition, judgment, and monitoring. This study aims to provide an accurate system in image processing techniques for the inspection and classification of sashimi food damage based on detecting external data. The external texture was identified based on the visible and invisible system that was acquired using multispectral technology. We proposed the Grey Level Co-occurrence Matrix (GLCM) model for analysis of the texture features of images and the classification process was performed using Artificial Neural Network (ANN) method. This study showed that multispectral technology is a useful system for the assessment of sashimi food and the experimental also indicates that the invisible channels have the potential in the classification model, since the hidden texture features that are not clearly visible to the human eye.

DOI： 10.1145/3379310.3379317

Language：English  

DOI： 10.1109/AIT49014.2019.9144929

Other Link： https://ieeexplore.ieee.org/document/9144929

Language：English   Publishing type：Research paper (other academic)  

Understanding conditions and situations causing sudden braking is important for preventing traffic accidents. Previous studies have used probe vehicle data to detect risky positions where sudden braking frequently occurred. However, they have mainly focused on vehicle-related factors.In this paper, we propose a novel method for discriminating sudden braking. Unlike previous studies, the method exhaustively explores probe data including temporal factors, constructs a large number of features combining pairs of feature names and their values, and applies the Support Vector Machine classifier and Feature Selection method to the features. To conduct the experiments, we used probe data provided by the Aizu-Wakamatsu City Open Data Utilization Verification Project. The proposed method discriminated sudden braking quite accurately, with a discrimination performance averaging an F1 measure of 93.2%. We also found that the probability of the occurrence of sudden braking is not always high at locations where sudden braking frequently occurred, but rather, temporal factors such as date and time, or day of week are strongly related to performance in discriminating sudden braking with high probability.

DOI： 10.1109/ITSC.2019.8917480

Language：English   Publishing type：Research paper (other academic)  

This paper presents a novel mining and visualizing tool that detects features to estimate sudden braking. The tool uses a machine learning and feature selection method to find the features exhaustively from combinations of the features which include not only vehicle-related factors, but also outer circumstances or temporal factors. The tool also obtains the locations inferred by the features detected. A normal way would first search for locations where sudden braking behavior frequently occurred, but it is not always true that the occurrence probability of sudden braking at the locations is high. On the other hand, our tool finds the locations related to sudden braking with high probability, more than 98%. Through the visualizing process, the features can be used as clues to find new factors which affect sudden braking.

DOI： 10.1109/IIAI-AAI.2019.00024

Language：English   Publishing type：Research paper (other academic)  

It is known as Benford's law that the distribution of the first digits forms a specific shape for natural numerical datasets. Deviation from the Benford's distribution indicates the irregularity of the dataset. However, it does not tell any clue to interpret the reason of irregularity. The present paper constructs a search engine of cells that appear in tables by correlating a cell with the words in the title of row or column or in the explanation of the table. We generate an exhaustive dataset of cells for testing irregularity by enumerating the search conditions. We applied the method to the number of applicants, the number of candidates, and the number of successful applicants in each department of 565 private universities in Japan. We confirmed the effectiveness of the proposed method by extracting the characteristics of the irregular datasets.

DOI： 10.1109/IIAI-AAI.2019.00207

Language：English   Publishing type：Research paper (scientific journal)  

Prediction of Travel Time over Unstable Intervals between Adjacent Bus Stops Using Historical Travel Time in both previous and current time periods

Language：English  

DOI： 10.1007/978-981-13-7434-0

Language：English   Publishing type：Research paper (other academic)  

Crime is one of the most important social problems for administrative region. Ascertaining the detailed characteristics of crime and preparing countermeasures are important to keep community life safe and secure. A lot of studies using crime data and geographical data have been carried out with a view to crime prevention. These studies include analyzing geographical features of crime, mapping crime-related information and crime hotspots on the map, predicting crime rate and so on. In addition, police stations have recently begun emailing notifications regarding crime to citizens to help them avoid crime. The e-mail messages include rich information about regional crime; they are actively used by services providing guidance to people in how to avoid crime. These services map the messages onto regional maps using the location information in the messages and show the relations between the locations and crime on the map. In addition, some services send alarms to their users when the GPS information of the users indicates that they are passing by the places where crime has occurred. However, these services only use the location and crime information extracted from the messages. Thus, we cannot say the messages have been fully used to clarify characteristics of regional crime. Therefore, in this paper, we investigate whether or not the crime messages sent by e-mail can be further exploited as a valid source for analyzing the criminal characteristics of a region, i.e., whether or not they include the characteristics of regional crime. To this end, in this research, we conducted experiments to make clear whether or not the crime messages sent by e-mail can help to distinguish regions. Experimental results illustrate that the contents of e-mail crime messages helped to distinguish regions having greater than or equal to 100 reports, with an average F-measure of about 90.3%, while only using the names of the areas where crime has occurred cannot match that F-measure.

DOI： 10.1007/978-3-030-19063-7_46

Language：English   Publishing type：Research paper (other academic)  

As the software systems are becoming more and more complicated, developers have an increasing dependency on code recommendation tools to assist them to fulfill their development tasks. However, the current historical-code-based recommendation methods are directly affected by the quality of the historical codes and the program-environment-information-based recommendation methods cannot provide satisfactory recommendation results for static methods because it is difficult to know all possible static members only using the program context, and even if we know all the static members, we still cannot add all of them to the entry point for search because its large number may cause a space explosion. In this paper, we propose a type conversion sequence recommendation method based on program environment information. Combing with the reachability analysis using semantic Web technology, the proposed method tries to reduce the searching entry points to solve the space explosion problem caused by the recommendation of static methods. We implemented an Eclipse plug-in based on the proposed method and conducted experiments on Tomcat source code. The experimental results showed that the proposed method can not only recommend type conversion sequences with static methods effectively, but also has a higher accuracy for the recommendation of object methods compared with the Eclipse Code Recommenders.

DOI： 10.1109/SmartWorld.2018.00076

Language：English   Publishing type：Research paper (other academic)  

Prediction of bus travel time and/or delay time is a useful tool for passengers who want to plan their journey, e.g., when they should leave from the origin bus stop, what they will do after arriving at the destination bus stop, and so on. Many studies have tackled this task using probe data and/or the real time data provided by automatic vehicle location (AVL) systems. Most of them only targeted a small number of routes, short time periods, e.g. less than one week, and used few machine learning models to evaluate their methods. However, different routes generally show different characteristics. In fact, there are big differences between urban routes and rural routes. Furthermore, the performance of machine learning models also varies according to the data dealt with by the models. In this paper, we propose prediction models for bus delay over all intervals between pairs of adjacent bus stops. To build the models, we use one month of bus probe data, which includes more than 80 routes, and apply several machine learning models: linear regression (LR), artificial neural network (ANN), support vector regression (SVR), random forest (RF), and gradient boosting decision tree (GBDT). Experimental results demonstrate the superiority of the GBDT-based prediction model and the effects of considering travel time over prior intervals.

DOI： 10.1109/BDCAT.2018.00020

Language：English   Publishing type：Research paper (scientific journal)  

Prediction of Bus Travel Time over Unstable Intervals between Two Adjacent Bus Stops

Language：English   Publishing type：Research paper (other academic)  

Prediction of bus travel times is of crucial importance for passengers in letting them know their departure time from an origin and arrival time at a destination and allowing them to make decisions (e.g., postpone departure time at certain hours) and to reduce their waiting time at bus stops. To predict bus travel times, it is important to know whether the target routes are stable or not. In this paper, we propose a time series approach to predict the travel time over an interval between two adjacent bus stops. We build Artificial Neural Network (ANN) models to predict the travel time over the interval. To make accurate predictions, we divide a day into 8 time-periods in calculating travel time over the interval and classify unstable intervals into three types: weak, medium and strong unstable. We use bus probe data collected from November 21st to December 20th 2013 and provided by Nishitetsu Bus Company, Fukuoka, Japan. Experimental results show that our models can effectively improve the prediction accuracy of travel times over intervals by focusing on the three unstable classes and calculating travel times for each interval at each of 8 time-periods in a day.

DOI： 10.1109/ICAdLT.2017.8547006

Language：English   Publishing type：Research paper (other academic)  

With increasing popularity of location-based social networks, POI recommendation has received much attention recently. Unlike most of the current studies which provide recommendations from perspective of users, in this paper, we focus on the perspective of Point-of-Interest (POI) for predicting potential users for a given POI. We propose a novel vector representation model for the prediction. Many current matrix factorization-based methods only pay attention to combining new information and basic matrix factorization, while in our model, we improve the matrix factorization model itself by replacing dot product with cosine similarity. We also address the problem of randomness of user’s check-in behavior by applying deep neural network to modeling the relationships between the user’s current check-in and context information of current check-in. Extensive experiments conducted on two real-world datasets demonstrate the superior performance of our proposed model and the effectiveness of the factors incorporated in our model.

DOI： 10.1007/978-3-030-03098-8_5

Language：English   Publishing type：Research paper (other academic)  

People do not always know about everything in the whole area even where they go about their daily life. Visitors to the area still more do not know about the area such as facilities and/or when and where to hold events in the area, means of transport to/from the area, and means of movements in the area if the area is too large for visitors to walk around, and so on. Although the main Web site of a local area such as a region, district, park or an organization usually works as a portal and provides the information above, the information is not always linked to with each other, in particular, on the map of the area except for facility names. Therefore people often feel difficult to find information related to their current or specified positions. Considering these problems, in this paper, we discuss an information sharing and recommendation platform specialized to a local area. The platform provides a variety of information associated with a map of the area, which means the information is linked with each other through the map. As a case study, we present 'ItoCamLife,' which targets Ito Campus of Kyushu University. Ito Campus is one of four campuses in Kyushu University and the largest campus in Japan as one campus. More than 10 thousand people including students, members of faculty and staffs are commuting, and many visitors are also coming there especially when a big event such as university festival or a large scale conference is held. Therefore, Ito Campus needs such a platform of sharing and recommending information. In this paper, we present an overview of ItoCamLife, including its concept and current functions.

DOI： 10.1109/IIAI-AAI.2018.00030

Language：English   Publishing type：Research paper (other academic)  

Smart mobility systems, which include Intelligent Transportation Systems(ITS) and smart energy systems, become more important. There is, however, lack of its platform studies. We started a sustainable information infrastructure project for smart mobility systems. The project pursues issues that establish an information infrastructure architecture and seamless development method chain for it. The project has mainly two features; 1) applying life-cycle-oriented methods, which are a cycle from system development to operations, to a real world, and 2) dealing with an uncertainty in system design phase. This paper describes current status of the project.

DOI： 10.1109/IIAI-AAI.2018.00025

Language：English   Publishing type：Research paper (other academic)  

Smart mobility services evolve rapidly in excessive competition with rivals in the market. Meanwhile, service suppliers cooperate partially to avoid exhaustion and concentrate on their core business. Therefore, homebrew and external services have dependency one another. This paper proposes some software engineering practices to develop and operate such smart mobility services. An interorganizational architecture is imported from the ITS domain as a shared view for service suppliers in competition and cooperation. Each service supplier construct an intraorganizational architecture for its home-brew services in microservice architecture style. Technical practices are basically feature oriented. Services are comprehended by the feature model and their boundaries are managed based on it. Supply chain management of services is conducted based on the data flow diagram traceable from the feature model. The contract and its variation on service supply is managed and accounted by the assurance case. Risks on sustainable service supply are studied with failure modes and effects analysis.

DOI： 10.1109/IIAI-AAI.2018.00031

Language：English   Publishing type：Research paper (other academic)  

Agent-based traffic simulation has become more and more attractive and important to develop new ITS (Intelligent Transport Systems) services. So far a variety of studies and developments that combine simulators and evaluate ITS services on the combined simulators have been conducted. In this paper, we introduce a simulation environment, called Agent-based USE (Agent-based Unified Simulation Environment), and some application examples for ITS services. The Agent-based USE provides an easy-to-build simulation environment for ITS-related services. In particular, by connecting simulators with ITS services, the Agent-based USE determines behaviors to be changed on the simulators using the data of the services such as recommendation results generated by the services, tells the decisions to simulators; the Agent-based USE then obtains the data representing the current situation on the simulators and sends the data to the services as feedback so as to enable the services to generate the next recommendation. In addition, by using the Agent-based USE, it is possible to construct a co-simulation environment where simulation is performed by synchronizing some kinds of simulators and services and by sharing each simulation information. In this paper, we introduce the overview and architecture of the Agent-based USE for traffic simulation, and discuss its usefulness through some application examples.

DOI： 10.1109/IIAI-AAI.2018.00027

Language：English   Publishing type：Research paper (other academic)  

Prediction of bus travel time is one of crucial issues for passengers in letting them know their departure time from an origin and arrival time at a destination and allowing them to make decisions (e.g., postpone departure time at certain hours) and to reduce their waiting time at bus stops. This paper proposes a time series approach to predict travel time over an interval between two adjacent bus stops. We build an Artificial Neural Network (ANN) model to predict travel time over the interval. To make accurate prediction, we divide a day into 8 time-periods in calculating travel time over the interval at each time-period and also use the travel time condition at right before the target time-period in order to apply the dynamical change of travel time as well as the historical average travel time at the same time-period during the past several days. To validate the proposed method, we used bus probe data collected from November 21st to December 20th in 2013, provided by Nishitetsu Bus Company, Fukuoka, Japan. Experimental results show that our models can effectively improve prediction accuracy of travel time on the route compared to a method only using the historical average travel time.

DOI： 10.1145/3164541.3164630

Language：English   Publishing type：Research paper (other academic)  

Prediction of bus travel time is one of crucial issues for passengers in letting them know their departure time from an origin and arrival time at a destination and allowing them to make decisions (e.g., postpone departure time at certain hours) and to reduce their waiting time at bus stops. This paper proposes a time series approach to predict travel time over an interval between two adjacent bus stops. We build an Artificial Neural Network (ANN) model to predict travel time over the interval. To make accurate prediction, we divide a day into 8 time-periods in calculating travel time over the interval at each time-period and also use the travel time condition at right before the target time-period in order to apply the dynamical change of travel time as well as the historical average travel time at the same time-period during the past several days. To validate the proposed method, we used bus probe data collected from November 21st to December 20th in 2013, provided by Nishitetsu Bus Company, Fukuoka, Japan. Experimental results show that our models can effectively improve prediction accuracy of travel time on the route compared to a method only using the historical average travel time.

DOI： 10.1145/3164541.3164630

Language：English  

Smart mobility systems, which include Intelligent Transportation Systems (ITS) and smart energy ones, become more important. There is, however, lack of its platform studies. This paper proposes a sustainable information infrastructure project for smart mobility systems. The project pursues issues that establish an information infrastructure architecture and a seamless development method chain for it. The project has mainly two features: (1) applying life-cycle-oriented methods, which are a cycle from system development phase to operation phase. In addition, these methods are applied to real world applications, and (2) dealing with an uncertainty that occurs in system development upper phase. This paper describes an overview of the project.

DOI： 10.1007/978-3-319-70636-8_3

Language：English   Publishing type：Research paper (other academic)  

Government 2.0 activities have become very attractive and popular. Using the platforms to support the activities, anyone can anytime report issues in a city on the Web and share the reports with other people. Since a variety of reports are posted, officials in the city management section have to give priorities to the reports. However, it is not easy task to judge the importance of the reports since importance judgments vary depending on the officials and consequently the agreement rate becomes low. To remedy the low agreement rate problem of human judgment, it is necessary to create an automatic method to find reports with high priorities. Hirokawa et al. employed the Support Vector Machine (SVM) with word feature selection method (SVM+FS) to detect signs of danger from posted reports because signs of danger is one of high priority issues to be dealt with. However they did not compare the SVM+FS method with other conventional machine learning methods and it is not clear whether or not the SVM+FS method has better performance than the other methods. This paper compared the results of the SVM+FS method with conventional machine learning methods: SVM, Random Forest, and Naïve Bayse with conventional word vectors, an LDA-based document vector, and word embedding by Word2Vec. Experimental results illustrate the validity and effectiveness of the SVM+FS method.

DOI： 10.1007/978-3-319-72550-5_26

Language：English  

Other Link： https://www.semanticscholar.org/paper/Rubric-based-Automated-Japanese-Short-answer-and-to-Ishioka/ee5f864fc9afbbd2727a6dfbfb79a8ca24abcba9?p2df

Language：English   Publishing type：Research paper (other academic)  

Connections between people are important in social life. The connections are usually recorded using business cards in business situations. However, opportunities for encounters to create a connection are rarely recorded, if people do not take the trouble of recording them. This is especially true concerning information regarding the place or situation that afforded the connection opportunity. That makes it difficult to reuse the connection information. Much research has been conducted so far on automatically detecting situations of a connection between two people. However, this research usually relies on special sensors or tools to recognize the connection. Such sensors or tools are unfortunately not usually readily available in daily life, and people do not want to carry additional sensors on a daily basis if they are only specialized to serve the purpose of recognizing connections. This paper proposes a method to detect situations in the connections between people using the BLE functions, acceleration sensors, microphone and recorder built into an ordinary smartphone. The objective in this research is to make clear the possibilities and the limitations of the proposed method in detecting peoples connection situations. Experimental results suggest that the proposed method can detect the conversation situation between people with their smartphones in their pocket or bag who are standing within a short distance of each other of less than one meter.

DOI： 10.1109/IIAI-AAI.2017.157

Language：English   Publishing type：Research paper (other academic)  

Connections between people are important in social life. The connections are usually recorded using business cards in business situations. However, opportunities for encounters to create a connection are rarely recorded, if people do not take the trouble of recording them. This is especially true concerning information regarding the place or situation that afforded the connection opportunity. That makes it difficult to reuse the connection information. Much research has been conducted so far on automatically detecting situations of a connection between two people. However, this research usually relies on special sensors or tools to recognize the connection. Such sensors or tools are unfortunately not usually readily available in daily life, and people do not want to carry additional sensors on a daily basis if they are only specialized to serve the purpose of recognizing connections. This paper proposes a method to detect situations in the connections between people using the BLE functions, acceleration sensors, microphone and recorder built into an ordinary smartphone. The objective in this research is to make clear the possibilities and the limitations of the proposed method in detecting peoples connection situations. Experimental results suggest that the proposed method can detect the conversation situation between people with their smartphones in their pocket or bag who are standing within a short distance of each other of less than one meter.

DOI： 10.1109/IIAI-AAI.2017.157

Language：English   Publishing type：Research paper (international conference proceedings)  

Other Link： https://www.m.ait.kyushu-u.ac.jp/mine/Paper/Mine-SECON-2017.pdf

Language：English   Publishing type：Research paper (other academic)  

Government 2.0 activities have become very attractive and popular these days. Using platforms to support the activities, anyone can anytime report issues or complaints in a city with their photographs and geographical information on the Web, and share them with other people. Since a variety of reports are posted, officials in the city management section have to check the importance of each report and sort out their priorities to the reports. However, it is not easy task to judge the importance of the reports. When several officials work on the task, the agreement rate of their judgments is not always high. Even if the task is done by only one official, his/her judgment sometimes varies on a similar report. To remedy this low agreement rate problem of human judgments, we propose a method of detecting signs of danger or unsafe problems described in citizens' reports. The proposed method uses a machine learning technique with word feature selection. Experimental results clearly explain the low agreement rate of human judgments, and illustrate that the proposed machine learning method has much higher performance than human judgments.

DOI： 10.1145/3106426.3106520

Language：English   Publishing type：Research paper (other academic)  

Government 2.0 activities have become very attractive and popular these days. Using platforms to support the activities, anyone can anytime report issues or complaints in a city with their photographs and geographical information on the Web, and share them with other people. Since a variety of reports are posted, officials in the city management section have to check the importance of each report and sort out their priorities to the reports. However, it is not easy task to judge the importance of the reports. When several officials work on the task, the agreement rate of their judgments is not always high. Even if the task is done by only one official, his/her judgment sometimes varies on a similar report. To remedy this low agreement rate problem of human judgments, we propose a method of detecting signs of danger or unsafe problems described in citizens' reports. The proposed method uses a machine learning technique with word feature selection. Experimental results clearly explain the low agreement rate of human judgments, and illustrate that the proposed machine learning method has much higher performance than human judgments.

DOI： 10.1145/3106426.3106520

Language：English  

Language：English   Publishing type：Research paper (scientific journal)  

The purpose of this study is to examine different formats of comment data to predict student performance. Having students write comment data after every lesson can reflect students' learning attitudes, tendencies and learning activities involved with the lesson. In this research, Latent Dirichlet Allocation (LDA) and Probabilistic Latent Semantic Analysis (pLSA) are employed to predict student grades in each lesson. In order to obtain further improvement of prediction results, a majority vote method is applied to the predicted results obtained in consecutive lessons. The research findings show that our proposed method continuously tracked student learning situations and improved prediction performance of final student grades.

Other Link： http://www.ifets.info/

Language：English   Publishing type：Research paper (scientific journal)  

Government 2.0 activities have become attractive and popular these days. Using tools of their activities, anyone can report issues or complaints in a city on the Web with their photographs and geographical information, and share their information with other people. On the other hand, unlike telephone calls, the concreteness of a report depends on its reporter. Thus, the actual status and demand to the status may not be described clearly or either one may be miss-described in the report. It may accordingly happen that officials in the city management section can not grasp the actual status or demand to the status of the report. To solve the problems, automatic finding incomplete reports and completing missing information are indispensable. In this paper, we propose methods to detect parts related to an actual status or demand to the status in a report using empirical patterns, dependency relations, and several machine learning techniques. Experimental results show that an average F-score and an average accuracy score our methods achieved were 0.798 and 0.893, respectively. In addition, in our methods, RF achieved better results than SVM for both F-score and accuracy scores.

DOI： 10.1527/tjsai.AG16-B

Language：English   Publishing type：Research paper (scientific journal)  

The purpose of this study is to examine different formats of comment data to predict student performance. Having students write comment data after every lesson can reflect students' learning attitudes, tendencies and learning activities involved with the lesson. In this research, Latent Dirichlet Allocation (LDA) and Probabilistic Latent Semantic Analysis (pLSA) are employed to predict student grades in each lesson. In order to obtain further improvement of prediction results, a majority vote method is applied to the predicted results obtained in consecutive lessons. The research findings show that our proposed method continuously tracked student learning situations and improved prediction performance of final student grades.

Language：English   Publishing type：Research paper (other academic)  

Government 2.0 activities have become very attractive and popular these days. Using platforms to support the activities such as FixMyStreet, SeeClickFix, or CitySourced, anyone can anytime report issues or complaints in a city with their photographs and geographical information on the Web, and share them with other people. On the other hand, unlike telephone calls, the concreteness of a report depends on its reporter; the actual status and demand to the status may not be described clearly or either one may be miss-described in the report. It may accordingly happen that officials in the city management section can not understand the actual sta-tus or a demand to the status from the report. To solve the problems, it is indispensable to complement missing in-formation and estimate the actual status or the demand to the status from ambiguous information in the report. This paper proposes novel methods to detect segments related to an actual status and the demand to the status in a report. The methods combine empirical rules with several machine learning techniques that actively use dependency relation between words. Experimental results illustrate the validity of the proposed methods.

DOI： 10.1145/3011141.3011201

Language：English   Publishing type：Research paper (other academic)  

Today, the use of learning analytics is becoming more crucial in the learning environment for the purpose of understanding and optimizing students' learning situations. The purpose of this paper is to examine the impacts of Teacher Interventions (TIs) on students' attitudes and achievements involved with the lesson by analyzing their freestyle comment data after every lesson. The current study proposes a new method for building an accessible prediction model, which represents students' activities, situations and viewpoints; the method classifies words in the student comments into six attribute types and indicates the most important types that affect the prediction results. Further, the prediction results are compared with the topic-based statistical method that uses Latent Dirichlet Allocation and Support Vector Machine models. The results proved that there were positive correlations between TIs and the quality of writing comments that affect on improving the prediction results.

DOI： 10.1109/FIE.2016.7757736

Language：English   Publishing type：Research paper (other academic)  

This paper proposes a new approach for predicting final student grade with high accuracy. It builds an attribute dictionary (AD) automatically from students’ comments collected after every lesson. Furthermore, it combines white-box models: Decision Tree (DT) and Random Forest (RF), and a black-box model: Support Vector Machine (SVM) to construct an interpretable prediction model and carry out eclectic rule-extraction. First, the AD is built from students’ comments, which are converted to attribute vectors. Second, the output decision is generated by SVM using the attribute vectors in the training phase and then DT and RF are applied to the output decision to extract symbolic rules. Experimental results illustrate the validity of the AD constructed automatically and the superiority of the proposed approach compared to single machine learning techniques: DT, RF and SVM.

DOI： 10.1007/978-3-319-47440-3_18

Language：English   Publishing type：Research paper (other academic)  

Estimation of bus travel time is one of crucial issues for passengers to let them know their departure time and arrival time at the destination, and to reduce their waiting time at the bus stop. Furthermore, it can also help to promote the development of city public transportation. To estimate bus travel time, it is worth of knowing whether target routes for passenger's travel is unstable or not, i.e. route with frequent delay or not. In this paper, we present an approach of distinguishing stable and unstable routes using bus probe data. The approach includes two phases; first phase estimates travel time variabilities on routes of a day, and second one distinguishes stable and unstable routes. We divide a day into eight inter-time periods: early morning, morning peak, late morning, midday, early afternoon, afternoon peak, evening, and late night, and detect stable and unstable routes considering the variabilities of standard deviation of the routes in a day. We used bus probe data collected from November 21st to December 20th in 2013. Considering the overall distribution of standard deviation of average travel time of all routes, we defined the criterion to distinguish stable and unstable routes and found the rate of unstable routes on weekends tended to become more than that on weekdays.

DOI： 10.1109/ICA.2016.31

Language：English   Publishing type：Research paper (other academic)  

Software libraries have become more and more complex in recent years. Developers usually have to rely on search engines to find API documents and then select suitable APIs to do relevant development when working on unfamiliar functions. However, the traditional search engines do not focus on searching APIs that make this process inconvenient and time consuming. Although a lot of efforts have been made on API understanding and code search in industry and academia, work and tools that can recommend API methods to users based on their description of API's functionality are still very limited. In this paper, we propose a search-based recommendation algorithm on API methods. We call the algorithm APIBook and implement an API method recommendation tool based on the proposed algorithm. The algorithm can recommend relevant API methods to users based on user input written in natural language. This algorithm combines semantic relevance, type relevance and the extent of degree that API method is used to sort these API methods and rank those that are highly relevant and widely used in the top positions. Examples of codes in real projects are also provided to help users to learn and to understand the API method recommended. The API recommendation tool selects the Java Standard Library as well as 100 popular open source libraries as API recommending material. Users can input the API description via the Web interface, and view the search results with sample codes on screen. The evaluation experiment is performed and the result shows that APIBook is more effective for finding APIs than traditional search models and it takes on average 0.7 seconds for finding relevant API methods which we think to be reasonable for satisfying daily query requirements.

DOI： 10.1145/2993717.2993727

Language：English   Publishing type：Research paper (other academic)  

This paper identifies metrics that can be used in the operation phase to realize a platform to facilitate the development of services for smart mobility. The paper analyzes requirements of the platform forsmart mobility using the I^∗ framework which facilitates identification of goals between stakeholders. We conduct goal-question-metric(GQM) identifying metrics, which can evaluate whether the application or platform meets the goals that we have identified through the I^∗ analysis.

DOI： 10.1109/ICA.2016.25

Language：English   Publishing type：Research paper (other academic)  

Smart mobility systems, which include Intelligent Transportation System (ITS) and smart energy systems, become more important. There is, however, lack of its platform studies. This paper proposes a sustainable information infrastructure project for smart mobility systems. The project pursues issues that establish an information infrastructure architecture and seamless development method chain for it. The project has mainly two features, 1) applying life-cycle-oriented methods, which are a cycle from system development to operations, to the real world, and 2) dealing with uncertainties when a system begins to be designed. This paper describes an overview of the project.

DOI： 10.1109/IIAI-AAI.2016.110

Language：English   Publishing type：Research paper (other academic)  

This paper presents a new application of public transit guide system named PATRASH (Personalized Autonomous TRAansit recommendation System considering user context and History). Although existing transit guide systems provide a lot of functions, their users can not always use all the functions under their control because of the complexity of the functions. Under this situation, we developed and supply PATRASH to reduce complicated user operations. PATRASH provides simple functions users can easily use. The concept of PATRASH is "First Show, Second Search." PATRASH displays routes and time table information recommended to the users. If they need, they can search another route. The target users of PATRASH are transit riders who come to Ito campus of Kyushu university. Ito campus locates in a rural area and has four routes run by three bus transportation companies and connected to a railroad or a subway. Since the main route usually used by each transit rider seems to be almost determined, we believe personalized function of PATRASH becomes more and more important. To realize the concept of PATRASH and give simple usability to the users, we propose a new 3D adaptive interface and release PATRASH to the public. We also present a preliminal case study of PATRASH project.

DOI： 10.1109/IIAI-AAI.2016.204

Language：English   Publishing type：Research paper (other academic)  

Most current prediction models are difficult for teachers to interpret. This induces significant problems of grasping characteristics for each grade group of students, which are helpful for giving intervention and providing feedback to them. In this paper, we propose a new method to build a practical prediction model based on comment data mining. The current study classifies students' comments into six attributes (attitudes, finding, cooperation, review the lesson, understanding, and next activity plan), then extracts generic rules 'IF-THEN' about students' activities, attitudes and situations in the learning environment. Decision Tree (DT) and Random Forest (RF) models are applied to discriminate unique features related to each grade group. Evaluation results reported a set of rules for students' performance among with their situations reflected through all the course of a semester.

DOI： 10.1109/IIAI-AAI.2016.114

Language：English   Publishing type：Research paper (other academic)  

Building an understandable student prediction model has an essential role to play in the educational environment. Most current prediction models are difficult for teachers to interpret. This poses problems for model use (e.g., improve student performance, interventions and allow a feedback process). In this paper, we propose a new approach in building a practical model by identifying a number of attributes in comment data that reflect students' learning attitudes, tendencies and activities involved with the lesson. We check the capability of an attributes representation model compared to the Latent Dirichlet Allocation (LDA) model that represents student comments as a statistical latent class 'Topics.' In addition, we employ a Multi-Instance Learning (MIL) method to all available information for each student to improve the efficiency and effectiveness of classical representation for each lesson to predict final student performance. Computational experiments show that when the model is regarded as MIL, the prediction performance achieves better results than those based on single instance representation for each lesson.

DOI： 10.1145/2843043.2843066

Language：English  

Language：English   Publishing type：Research paper (scientific journal)  

Understanding individual students more deeply in the class is the most vital role in educational situations. Using comment data written by students after each lesson helps in the understanding of their learning attitudes and situations. They can be a powerful source of data for all forms of assessment. The PCN method categorizes the comments into three items: P (Previous learning activity), C (Current learning activity), and N (Next learning activity plan). The objective of this paper is to investigate how the three time-series items: P, C, and N, and the difficulty of a subject affect the prediction results of final student grades using two types of machine learning techniques: Support Vector Machine (SVM) and Artificial Neural Network (ANN). The experiment results indicate that the students described their current activities (C-comment) in more detail than previous and next activities (P-and N-comments)
this tendency is reflected in prediction accuracy and F-measure of their grades.

DOI： 10.2197/ipsjjip.23.784

Language：English   Publishing type：Research paper (scientific journal)  

Understanding individual students more deeply in the class is the most vital role in educational situations. Using comment data written by students after each lesson helps in the understanding of their learning attitudes and situations. They can be a powerful source of data for all forms of assessment. The PCN method categorizes the comments into three items: P (Previous learning activity), C (Current learning activity), and N (Next learning activity plan). The objective of this paper is to investigate how the three time-series items: P, C, and N, and the difficulty of a subject affect the prediction results of final student grades using two types of machine learning techniques: Support Vector Machine (SVM) and Artificial Neural Network (ANN). The experiment results indicate that the students described their current activities (C-comment) in more detail than previous and next activities (P-and N-comments); this tendency is reflected in prediction accuracy and F-measure of their grades.

DOI： 10.2197/ipsjjip.23.784

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.2197/ipsjjip.23.744

Language：Japanese  

学習者に，毎回の授業についての感想やコメントを求めることがよく行われている．これらのコメント文には，学習者の授業態度や学習状況が含まれることが多いため，学習者に応じたフィードバックを提供する非常に良い材料となる．我々は，先行研究において，学習者に，授業前の予習内容(P)，授業時の理解度や態度(C)，次授業時までの学習目標(N)の3項目を記述するよう求めてきた．この記述を自己評価記述文と呼ぶ．当該科目担当教師の主観評価により，項目Cの内容について点数化し，その点数と，最終成績との間に相関があることを確認した．しかし上位の学習者の識別が困難であった．この原因として，主観評価では理解度を詳細に点数化することが難しいこと，更に，求められた内容(P，C，N)に沿った記述をしない学習者の存在が考えられた．そこで，これら主観評価の問題と，PCN項目に関連しない内容記述の問題を考慮し，本論文では，PCNの各項目らしさを示すスコアを定義し，機械学習手法の一つであるSVMを適用することで，PCN項目を記述する文の推定を高精度で行うとともに，学習者の自己評価記述文のPCNスコアが高いほど，最終成績の推定精度も高くなることを確認した．

DOI： 10.14923/transinfj.2014JDP7124

Language：English   Publishing type：Research paper (other academic)  

Learning analytics is valuable sources of understanding students' behavior and giving feedback to them so that we can improve their learning activities. Analyzing comment data written by students after each lesson helps to grasp their learning attitudes and situations. They can be a powerful source of data for all forms of assessment. In the current study, we break down student comments into different topics by employing two topic models: Probabilistic Latent Semantic Analysis (PLSA), and Latent Dirichlet Allocation (LDA), to discover the topics that help to predict final student grades as their performance. The objectives of this paper are twofold: First, determine how the three time-series items: P-, C- and N-comments and the difficulty of a subject affect the prediction results of final student grades. Second, evaluate the reliability of predicting student grades by considering the differences between prediction results of two consecutive lessons. The results obtained can help to understand student behavior during the period of the semester, grasp prediction error occurred in each lesson, and achieve further improvement of the student grade prediction.

DOI： 10.1109/ICALT.2015.24

Language：English   Publishing type：Research paper (scientific journal)  

Language：English   Publishing type：Research paper (other academic)  

The paper aims to interpret visitor behaviors by analyzing a large scale Wi-Fi location data obtained at Interop Tokyo 2014. We first detected a situation of a visitor's stay at a booth, calculated the sojourn time of the visitor and represented each visitor as a booth vector whose element value is the total sojourn time of the visitor at a booth. Then, we classified the visitors by k-Means. By using category labels of each booth, we analyzed characteristics of clusters. The results illustrate that some categories appear at the top rank of most clusters, and the area of the booths included in each cluster is mostly some specific small one, not the entire one.

DOI： 10.1109/IIAI-AAI.2015.194

Language：English   Publishing type：Research paper (other academic)  

DOI： 10.1109/IIAI-AAI.2015.153

Language：English   Publishing type：Research paper (other academic)  

Examining student learning behavior is one of the crucial educational issues. In this paper, we propose a new method to predict student performance by using comment data mining. A teacher just asks students after every lesson to freely describe and write about their learning situations, attitudes, tendencies, and behaviors. The method employs Latent Dirichlet Allocation (LDA) and Support Vector Machine (SVM) to predict student grades in each lesson. In order to obtain further improvement of prediction results, we apply a majority vote method to the predicted results obtained in consecutive lessons to keep track of each student's learning situation. Also, we evaluate the reliability of the predicted student grades to know when we can rely prediction results of student grade during the period of the semester. The experiment results show that our proposed method continuously tracked student learning situation and improved prediction performance of final student grades compared to Probabilistic Latent Semantic Analysis (PLSA) and Latent Semantic Analysis (LSA) models. Also, considering the differences of prediction results in the two consecutive lessons helps to evaluate the reliability of the predicted results.

DOI： 10.1109/IIAI-AAI.2015.170

Language：English   Publishing type：Research paper (other academic)  

A lot of actions have recently been taken to support Government 2.0 movement. As the number of the actions increase, many people submit greater number of complaint reports by phone or mobile devices, and make sure the situation reported with each other. According to the actions, the delay in taking action of the government side becomes more clearly identified due to overloads of the government side to deal with the activities. To remedy the above situations, it increases of importance to develop an efficient approach to deal with the complaint reports. Automatic classification of the complaint reports, or estimation and extraction of demanding sentences from the reports are contributory to the approach. In this paper, we propose a method of automatically estimating categories of the complaint reports as a first step. We conducted experiments of estimating categories of the complaint reports. The experiment results showed the following findings: (1) Feature selection is a key to improve the F-score of estimating categories of complaint reports. The percentage of the words strongly effective for the category estimation is about 3.9% of the entire distinct words. (2) Proposed Mutual-Information-based methods outperform the F-score of a conventional Random-Forest-based method. (3) The F-score performance of estimating a category depends on the ambiguity level of the category. In particular, the F-score of estimating categories of a complaint report assigned multiple categories is 1.5 times worse than that of a complaint report assigned single category.

DOI： 10.1109/IIAI-AAI.2015.195

Language：Japanese  

DOI： 10.14923/transinfj.2014SWP0015

Language：English   Publishing type：Research paper (other academic)  

Assessment of learning progress and learning gain play a pivotal role in education fields. New technologies like comment data mining promote the use of new types of contents; student comments highly reflect student learning attitudes and activities compared to more traditional methods and they can be a powerful source of data for all forms of assessment. A teacher just asks students after every lesson to freely describe and write about their learning situations and behaviors. This paper proposes new methods based on a statistical latent class "Topics" for the task of student grade prediction; our methods convert student comments using latent semantic analysis (LSA) and probabilistic latent semantic analysis (PLSA), and generate prediction models using support vector machine (SVM) and artificial neural network (ANN) to predict student final grades. The experimental results show that our methods can accurately predict student grades based on comment data.

DOI： 10.1145/2676723.2677259

Language：English   Publishing type：Research paper (other academic)  

Assessment of learning progress and learning gain play a pivotal role in education fields. New technologies like comment data mining promote the use of new types of contents
student comments highly reflect student learning attitudes and activities compared to more traditional methods and they can be a powerful source of data for all forms of assessment. A teacher just asks students after every lesson to freely describe and write about their learning situations and behaviors. This paper proposes new methods based on a statistical latent class "Topics" for the task of student grade prediction
our methods convert student comments using latent semantic analysis (LSA) and probabilistic latent semantic analysis (PLSA), and generate prediction models using support vector machine (SVM) and artificial neural network (ANN) to predict student final grades. The experimental results show that our methods can accurately predict student grades based on comment data.

DOI： 10.1145/2676723.2677259

Language：English   Publishing type：Research paper (other academic)  

Predicting students' academic achievement with high accuracy has an important vital role in many academic disciplines. Most recent studies indicate the important role of the data type selection. They also attempt to understand individual students more deeply by analyzing questionnaire for a particular purpose. The present study uses free-style comments written by students after each lesson, to predict their performance. These comments reflect their learning attitudes to the lesson, understanding of subjects, difficulties to learn, and learning activities in the classroom. To reveal the high accuracy of predicting student's grade, we employ (LSA) latent semantic analysis technique to extract semantic information from students' comments by using statistically derived conceptual indices instead of individual words, then apply (ANN) artificial neural network model to the analyzed comments for predicting students' performance. We chose five grades instead of the mark itself to predict student's final result. Our proposed method averagely achieves 82.6&#37; and 76.1&#37; prediction accuracy and F-measure of students' grades, respectively.

DOI： 10.1109/FIE.2014.7044399

Language：English   Publishing type：Research paper (other academic)  

This paper proposes a prediction framework for student performance based on comment data mining. Given the comments containing multiple topics, we seek to discover the topics that help to predict final student grades as their performance. To this end, the paper proposes methods that analyze students’ comments by two topic models: Probabilistic Latent Semantic Analysis (PLSA), and Latent Dirichlet Allocation (LDA). The methods employ Support Vector Machine (SVM) to generate prediction models of final student grades. In addition, Considering the student grades predicted in a range of lessons can deal with prediction error occurred in each lesson, and achieve further improvement of the student grade prediction.

DOI： 10.1007/978-3-319-19773-9_117

Language：English   Publishing type：Research paper (scientific journal)  

Public transportation guidance services, such asYahoo, Jorudan andNAVITIME, are widely used nowadays and support our daily lives. Although they provide useful services, they have not fully been personalized yet. This paper presents a personalized transportation system called PATRASH: Personalized Autonomous TRAnsportation recommendation System considering user context and History. In particular,we discuss anAdaptiveUser Interface (AUI) of PATRASH.Before designing a personalized route recommendation function for PATRASH’s AUI, we investigated possibilities and effectiveness of the function. First, we collected and analyzed 10 subjects’ usage histories of public transportation. Through this investigation, we confirmed the possibilities and effectiveness of the personalized route recommendation function. Second, we investigated the effectiveness of the basic functions of PATRASH’s AUI by comparing with two major transportation guidance systems in Japan. We evaluated those systems from the point of view of usabilities: click costs and time costs. The experimental results illustrate the effectiveness of AUI of PATRASH.

DOI： 10.1007/978-3-662-47227-9_9

Language：English   Publishing type：Research paper (scientific journal)  

In this paper we propose a new approach based on text mining techniques for predicting student performance using LSA (latent semantic analysis) and K-means clustering methods. The present study uses free-style comments written by students after each lesson. Since the potentials of these comments can reflect student learning attitudes, understanding of subjects and difficulties of the lessons, they enable teachers to grasp the tendencies of student learning activities. To improve our basic approach using LSA and k-means, overlap and similarity measuring methods are proposed. We conducted experiments to validate our proposed methods. The experimental results reported a model of student academic performance predictors by analyzing their comments data as variables of predictors. Our proposed methods achieved an average 66.4% prediction accuracy after applying the k-means clustering method and those were 73.6% and 78.5% by adding the overlap method and the similarity measuring method, respectively.

DOI： 10.2197/ipsjjip.23.192

Language：English  

Other Link： http://www.acisinternational.org/journal/full/V15%20N2.pdf

Language：English   Publishing type：Research paper (other academic)  

The present study proposes prediction approaches of student's grade based on their comments data. Students describe their learning attitudes, tendencies and behaviors by writing their comments freely after each lesson. The main difficulty of this research is to predict students' performance by separately using two class data in each lesson. Although students learn the same subject, there exist differences between the comments in the two classes. The proposed methods basically employ latent semantic analysis (LSA) and two types of machine learning technique: SVM (support vector machine) and ANN (artificial neural network) for predicting students' final results in four grades of S, A, B and C. Moreover, an overlap method was proposed to improve the accuracy prediction results, the method allows to accept two grades for one mark to get the correct relation between LSA results and students' grades. The proposed methods achieve 50.7% and 48.7% prediction accuracy of students' grades by SVM and ANN, respectively. To this end, the results of this study reported models of students' academic performance predictors that are valuable sources of understanding students' behavior and giving feedback to them so that we can improve their learning activities.

DOI： 10.1109/IIAI-AAI.2014.17

Language：English   Publishing type：Research paper (other academic)  

Public transportation guidance services, which are widely used nowadays, support our daily lives. However they have not fully been personalized yet. Regarding personalized services, an adaptive user interface plays a crucial role. This paper presents an Adaptive User Interface (AUI) agent of our personalized transportation recommendation system called PATRASH. To design and implement the agent, first, we collected and analyzed public transportation usage histories of 10 subjects so as to confirm the possibilities and effectiveness of the personalized route recommendation function. Then we propose a method to deal with user histories and evaluate the effectiveness of the proposed method based on click costs, comparing with two major transportation guidance systems in Japan. We also propose a decision-tree-based route recommendation method. The experimental results illustrate the effectiveness of the proposed method.

Language：English   Publishing type：Research paper (other academic)  

Public transportation guidance services, such as Yahoo, Jorudan and NAVITIME, are widely used nowadays and support our daily lives. Although they provide useful services, they have not fully been personalized yet. This paper presents a personalized transportation system called Patrash: Personalized Autonomous TRAnsportation recommendation System considering user context and History. In particular, we discuss an Adaptive User Interface (AUI) of Patrash. Before designing a personalized route recommendation function for Patrash's AUI, we investigated possibilities and effectiveness of the function. First, we collected and analyzed 10 subjects' usage histories of public transportation. Through this investigation, we confirmed the possibilities and effectiveness of the personalized route recommendation function. Second, we investigated the effectiveness of the basic functions of Patrash's AUI by comparing with two major transportation guidance systems in Japan. We evaluated those systems from the points of views of usabilities: click costs and time costs. The experimental results illustrate the effectiveness of AUI of Patrash.

DOI： 10.1109/IIAI-AAI.2014.31

Language：English   Publishing type：Research paper (other academic)  

In this paper we propose a new approach based on text mining technique to predict student's performance using LSA (latent semantic analysis) and K-means clustering method. The present study uses free style comments written by students after each lesson. Since the potentials of these comments can reflect students' learning attitudes, understanding and difficulties to the lessons, they enable teachers to grasp the tendencies of students' learning activities.To improve this basic approach, overlap method and similarity measuring technique are proposed. We conducted experiments to validate our proposed methods. The experimental results illustrated that prediction accuracy was 73.6% after applying the overlap method and that was 78.5% by adding the similarity measuring.

DOI： 10.1007/978-3-319-09635-3_15

Language：English   Publishing type：Research paper (other academic)  

Predicting students' academic performance has long been an important research topic in many academic disciplines. The prediction will help the tutors identify the weak students and help them score better marks; these steps were taken to improve the performance of the students. The present study uses free style comments written by students after each lesson. These comments reflect their learning attitudes to the lesson, understanding of subjects, difficulties to learn, and learning activities in the classroom. (Goda and Mine, 2011) proposed PCN method to estimate students' learning situations from their comments freely written by themselves. This paper uses C (Current) method from the PCN method. The C method only uses comments with C item that focuses on students' understanding and achievements during the class period. The aims of this study are, by applying the method to the students' comments, to clarify relationships between student's behaviour and their success, and to develop a model of students' performance predictors. To this end, we use Latent Semantic Analyses (LSA) and K-means clustering techniques. The results of this study reported a model of students' academic performance predictors by analysing their comment data as variables of predictors.

Language：English   Publishing type：Research paper (other academic)  

A personalized system considering user context and social histories is attractive. We present a personalized transportation system called PATRASH: Personalized Autonomous Transportation recommendation System considering user context and History. PATRASH considers not only users' context and transportation histories, but also real bus delay. The bus delay returned by Bus Delay API (BDAPI) is used for making adjustments of recommended routes. In this paper, we proposes a model for dealing with the bus delay information. We also make confirmation of effectiveness or possibilities of personalized route recommendation based on user context and histories. To this end, we investigated 10 examinees' usage histories of public transportation as a preliminal case study. The preliminal investigation results promise us to predict individual public transportation route.

DOI： 10.1109/CSCI.2014.74

Language：English   Publishing type：Research paper (other academic)  

Capturing people's behavior patterns, such as purchase patterns or shop-around-behavior patterns, are big marketing concern since such patterns are good resources to invent new marketing services. To this end, a lot of studies using POS (Points Of Sales) data have been conducted to extract purchase behaviors for CRM (Customer Relation Management). In addition, since a lot of widespread sensor devices, smart cards or mobile phones have become available recently, anyone can try to analyze pedestrians' trajectories obtained from GPS data and extract people's behavior patterns. However, only few research has been conducted to analyze relationships between boarding and purchase behaviors based on usage histories of smart IC cards. This paper presents case studies of purchase behavior analysis based on usage histories of smart IC cards, especially focusing on purchase behaviors of passengers who just get off trains.

DOI： 10.1109/IIAI-AAI.2014.63

Language：English   Publishing type：Research paper (other academic)  

This paper proposes a novel reciprocal recommendation method for job matching with bi-directional feedback. The proposed method uses, as mutual feedback, bilateral messages between job seekers and recruiters, such as applying to a job, scout of a seeker, and reply to the offer, on the seekers-recruiters' user network. During job matching process, user agents, as delegate of their owners, send and receive those messages with each other. From those feedback messages, each user agent computes the popularity degree of its owner user: seeker or recruiter, and evaluation degree of each other from the popularity degree, considering both the popularity and evaluation degrees, and the similarity between a condition provided by its user and a profile of each candidate user, the agent dynamically updates a ranking list for recommendation of its owner user after every matching action. Preliminary experiments illustrate the validity of the proposed method.

DOI： 10.1109/IIAI-AAI.2013.91

Language：English   Publishing type：Research paper (other academic)  

To grasp a student's lesson attitude and learning situation and to give a feed back to each student are educational foundations. Goda et al. proposed the PCN method to estimate a learning situation from a comment freely written by students[6, 7]. The PCN method categorizes comments into three items of P (previous), C(current) and N(next). They pointed out a correlation between the student's final results and the validity of a descriptive content of item C, that is something related to understanding of the lesson and learning attitudes to the lesson. However, a problem left in their work is the badness of performance in prediction for upper grade students. This paper proposes two manners of utilization of PCN scores: the validity level determination for assessment, and for prediction performance of students' final grades. In order to validate the proposed manners of utilization, we conducted two experiments. First, we employed multiple regression analysis to calculate PCN scores that determine the validity level with respect to each viewpoint. Students who wrote comments with a high PCN score are considered as those who describe their learning attitude appropriately. We also applied a machine learning method SVM (support vector machine) to students' comments for predicting their final results in five grades of S, A, B, C and D. Experimental results illustrated that as comments of students get higher PCN scores, the prediction performance of the students' grades becomes higher.

DOI： 10.1145/2512276.2512294

Language：English   Publishing type：Research paper (scientific journal)  

Learning comments are valuable sources of interpreting student status of understanding. The PCN method introduced in [Gouda2011] analyzes the attitudes of a student from a view point of time series. Each sentence of a comment is manually classified as one of P,C,N or O sentence. P(previous) indicates learning activities before the classtime, C(current) represents understanding or achievements during the classtime, and N(next) means a learning activity plan or goal until next class. The present paper applies SVM(Support Vecotor Machine) to predict the category to which a given sentence belongs. Empirical evaluation using 4,086 sentences was conducted. By selecting feature words of each category, the prediction performance was satisfactory with F-measures 0.8203, 0.7352, 0.8416 and 0.8612 for P,C,N and O respectively.

DOI： 10.1007/978-3-642-41175-5_14

Language：Japanese   Publishing type：Research paper (scientific journal)  

Language：Japanese   Publishing type：Research paper (scientific journal)  

Language：English   Publishing type：Research paper (other academic)  

Balance is very important in all areas from art down to the details of our daily life. There is no exception in research work especially when we want to develop an application which will be used in the real world. Currently more and more semantic Web applications are emerging. Although there are some researches on the evaluation and benchmarking of semantic Web applications, they mainly focused on the semantic Web technology itself or specific types of applications. The evaluation of general features of Semantic Web applications is not sufficient. In this paper, we will focus on the evaluation of one of the general characteristics - balance of semantic Web applications. An analysis of current semantic Web applications will be given first in order to summarize their statistical features and get some hints for identifying the aspects that are critical for balance evaluation. Then we summarize the main aspects of balance evaluation, and present the key factors for each balance evaluation aspect. Finally, suggestions for future semantic Web application development and evaluation will be made based on the analysis of semantic Web applications and the summarizing of balance evaluation aspects.

DOI： 10.1109/IIAI-AAI.2012.49

Language：English   Publishing type：Research paper (other academic)  

Peer-to-peer (P2P) communication and computing frameworks are important for constructing robust large-scale distributed systems. Overlay network systems use distributed hash-table (DHT) to provide scalable and efficient node search capabilities. However, the DHT-based method has a problem for the maintenance cost of dynamically changing large-scale-network, in which nodes are frequently joining and leaving. This paper proposes a novel technique of P2P communication path management. The proposed technique devises a robust semi-structured overlay network called Ordered Tree with Tuft (OTT for short). OTT provides not only efficient node searching, but also low-cost self-maintenance capabilities for the dynamically changing network. In this method, joining and leaving of a node are managed in O(1) with high probability. Furthermore, the proposed OTT-based technique can find and construct a path shorter than that on the normal ordered tree, by setting up bypass links between remote nodes on OTT.

DOI： 10.1007/978-3-642-31809-2_8

Language：English   Publishing type：Research paper (other academic)  

According to user's intent, Broder classified Web queries into three classes: navigational, informational and transactional. However, current search engines do not provide relevant search results based on the user's intent. In this paper, we propose a user intent based searching mechanism (UIBS for short) in order to enable precise discovery of a Web site (for navigational searching) as well as aggregating of their information (for informational searching) and performing further activities (for transactional searching). We first propose a Web site capability description interface for the explicit description of a Web site's capability and a search interface that enables explicit description of user's search intents, preferences, and query on the information. And then we describe the process of the query engine which provides relevant search results based on the user's intent, preferences, and requirements. Finally, a test example is given to show the availability of the proposed UIBS mechanism.

DOI： 10.1109/TrustCom.2011.191

Language：Others  

Protection of Personal Information based on User Preference

Language：English   Publishing type：Research paper (other academic)  

These days, many university teachers are concerned about the increasing number of students whose motivation is declining. Some of them fall into a situation that they cannot recover from by themselves, and require assistance, but they hesitate to call for help. In order to recognize such students quickly and give guidance to them in class, we have collected time-series comments in the classroom and analyzed them. In the analysis, we divided the comments into the three time slots: P (Previous), C (Current), and N (Next), and quantify them so that we can infer the learning behaviors between the previous and the current classes. We call this analysis method the PCN method. The PCN method is useful for grasping students' learning status in the class. Some of our case studies illustrate the validity of the PCN method.

DOI： 10.1007/978-3-642-23863-5_16

Language：English   Publishing type：Research paper (other academic)  

Some of Internet services require users to provide their sensitive information such as their name, address, credit card number, and an ID-password pair. In these services, the manner in which the provided information is used is solely determined by the service providers. As a result, even when the manner in which information is used by a service provider appears vulnerable, users have no choice but to allow such usage. Therefore, we have proposed a user customized data protection framework that enables users to select the manner in which their sensitive information is protected. In our framework, a user selects a policy that defines the manner in which his/her information is to be protected and its manner defined by the policy is incorporated into a program. By allowing a service provider to the information provided by a user through the program, the user can protect his/her sensitive information in a manner selected by him/her. This framework works well when existing a manner (protection policy) which is tolerant to the alteration of the program, otherwise, a program alteration might be a concern. Therefore, in this paper, we attempts to protect a customized program by using program obfuscation and sanitizable signature techniques.

DOI： 10.1109/CSAE.2011.5953301

Language：English   Publishing type：Research paper (other academic)  

Some of Internet services require users to provide their sensitive information such as credit card number, and an ID-password pair. In these services, the manner in which the provided information is used is solely determined by the service providers. As a result, even when the manner in which information is used by a service provider appears vulnerable, users have no choice but to allow such usage. In this paper, we propose a framework that enables users to select the manner in which their sensitive information is protected. In our framework, a policy, which defines the type of information protection, is offered as a Security as a Service. According to the policy, users can incorporate the type of information protection into a program. By allowing a service provider to use their sensitive information through this program, users can protect their sensitive information according to the manner chosen by them.

DOI： 10.1007/978-3-642-22191-0_27

Language：English   Publishing type：Research paper (scientific journal)  

This paper proposes a novel Peer-to-Peer Information Retrieval (P2PIR) method using user feedback and query-destination-learning. The method uses positive feedback information effectively for getting documents relevant to a query by giving higher score to them. The method also utilizes negative feedback information actively so that other agents can filter it out with itself. Using query-destination-learning, the method can not only accumulate relevant information from all the member agents in a community, but also reduce communication loads by caching queries and their sender-responder agent addresses in the community. Experiments were carried out on both single and multiple communities constructed with multi-agent framework Kodama. The experimental results illustrated that the proposed method effectively increased retrieval accuracy.

DOI： 10.1527/tjsai.26.97

Language：English   Publishing type：Research paper (other academic)  

Learning activity plays important role in enhancing one's knowledge and skill. There are many ways to acquire and extract learning activities of students from their learning information; we focus on comments handwritten in their attendance sheets. It is easy for teachers to collect the sheets every class and for students to write their activities as comments. The sheets consequently provide time-series text data related to students; such the data are treasures because the comments and the questionnaire reflect their learning activities directly and indirectly. We propose a method called a PCN method for quantifying the comments into triple showing inferred learning activities student by student. Case studies illustrate the validity of the PCN method.

Language：English   Publishing type：Research paper (other academic)  

When searching for information a user wants, search engines often return lots of results unintended by the user. Query expansion is a promising approach to solve this problem. In the query expansion research, one of the biggest issues is to generate appropriate keywords representing the user's intention. The Related Word Extraction Algorithm (RWEA) we proposed extracts such keywords for the query expansion. In this paper, we evaluate the RWEA through several experiments considering the types of queries given by the users. We compare the RWEA, Robertson's Selection Value (RSV) which is one of the famous relevance feedback methods, and the combination of RWEA and RSV. The results show that as queries become more ambiguous, the advantage of the RWEA becomes higher. From the points of view of query types, the RWEA is appropriate for informational queries and the combined method is for navigational queries. For both query types, RWEA helps to find relevant information.

DOI： 10.1007/978-3-642-16098-1_3

Language：English   Publishing type：Research paper (other academic)  

This paper proposes a web retrieval system with extended queries generated from the contents of Wikipedia.By using the extended queries, we aim to assist user in retrieving Web pages and acquiring knowledge. To extract extended query items, we make much of hyperlinks in Wikipedia in addition to the related word extraction algorithm. We evaluated the system through experimental use of it by several examinees and the questionnaires to them. Experimental results show that our system works well for user's retrieval and knowledge acquisition.

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (other academic)  

In this paper, a Semantic Agent-Community-based Peer-to-Peer information retrieval method called SACP2P method is proposed for reliable community Web information sharing. The evaluation experiment is performed and the result has shown that SACP2P method can aggregate information from different sources published through different methods (including Web content and Web services) and be effectiveness on reducing communication loads in a P2P network.

DOI： 10.1109/SSIRI.2009.57

Language：English   Publishing type：Research paper (other academic)  

This paper proposes a novel Peer-to-Peer Information Retrieval (P2PIR) method using user feedback and query-learning. The method actively utilizes negative feedback information so that other agents can filter it out when retrieving it. The proposed method effectively increases retrieval accuracy and decreases communication loads required for document retrieval in communities. The experiments were carried out on multiple communities constructed with multi-agent framework Kodama [1]. The experimental results illustrated the validity of our proposed method.

DOI： 10.1007/978-3-642-11161-7_19

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (other academic)  

Recently, several community-based systems have been developed; however, almost all such systems have been developed as Webserver- based systems. Thus, server administrator can easily eavesdrop on user communications, since they have to send/receive information through the server. Therefore, we propose multi-agent-based peer-topeer (P2P) system wherein each peer manages his/her information and exchanges it with other peers directly. This, thus, resolves the problems posed byWeb-server-based systems; however, we have to consider attacks from malicious third parties. This study designs and implements security protocols/mechanisms for a hierarchical community-based multi-agent system. Furthermore, if we consider a practical use case, we should be able to demonstrate that the proposed system can be implemented by combining it with existing security techniques for more reliable and rapid deployment. Finally, we evaluate the performance of the proposed security system and present an example application.

DOI： 10.1007/978-3-642-01639-4_12

Language：Japanese   Publishing type：Research paper (scientific journal)  

Language：English   Publishing type：Research paper (other academic)  

This paper proposes two methods for improving the retrieval accuracy of the Agent-Community-based Peer-to-Peer information retrieval (ACP2P) method. One uses user feedbacks exchanged in a community. The other uses query-learning methods that make a middle agent to learn query-responder agent relationships. The latter methods are useful not only for improving the retrieval accuracy, but also for reducing communication loads. We conduct several experiments with test collections so that the evaluation can be done in an objective manner. The experimental results illustrate the validity of our proposed methods.

Language：English   Publishing type：Research paper (other academic)  

Community-based collaboration support systems are useful for exchanging information on topics that community members are interested in. Most of them developed so far are based on server-client architecture and provide their services on Web servers. They require special administrative facilities, and ask users to upload their data on the systems. Furthermore, security mechanisms are not often provided for the communities. Considering these problems, we have been developing an Agent-Community-Network-based collaboration support system: in particular, a business-matching support system. Our system requires neither any special administrative facilities nor the need to upload user data to a special server. Furthermore, it supports secure peer-to-peer communication between users. It is implemented with a multi-agent Kodama framework.

DOI： 10.1007/978-3-642-01668-4_13

Language：English   Publishing type：Research paper (other academic)  

The progress of Information Technology, which is the infrastructure of an advanced information society, is remarkable and has the enormous impact on our daily life. On the other hand, it has been pointed out by the industry in Japan that there is a lack of highly skilled Information and Communication Technology personnel who can lead the next generation. In order to address this issue, the Graduate School of Information Science and Electrical Engineering in Kyushu University has established Social Information System Engineering education course. Since April 2007, we have started practical education program with an objective to foster world class leaders who have extraordinary technical skill, basic knowledge and sense of ethics. This effort is steadily progressing by the collaboration with various companies through Nippon Keidanren with support of Ministry of Education, Culture, Sports, Science and Technology. There have been several findings in 1) Project Based Learning, 2) omnibus courses, 3) long term internship, and 4) curriculum improvements during the planning and execution of this program. This paper describes the content, method, interim result and evaluation of this education course. We also discuss the challenges that need to be resolved.

Language：English   Publishing type：Research paper (other academic)  

When searching for information a user wants, search engines often return lots of results unintended by the user. Query expansion is a promising approach to solve this problem. In the query expansion research, one of big issues is to generate appropriate keywords representing the user's intention. This paper proposes the related word extraction algorithm (RWEA) which pays attention to the distance between sentences where keywords in the original query appear. The RWEA is based on the idea that a word nearby important words is also important. We conducted several experiments to evaluate the RWEA. The results promise the effectiveness of the RWEA for improving search results.

DOI： 10.1109/WIIAT.2008.308

Language：English   Publishing type：Research paper (scientific journal)  

A multi-agent system that consists of agents distributed in a network requires an application-level routing method using abstract identifiers such as agent names. We have accordingly proposed a routing method based on an Ordered-Tree-with-Tuft-shaped (OTT-shaped) overlay network. The method can maintain the network robustly and at low cost, and also make use of short-cuts between nodes to find shorter paths than those only found by the ordered tree of OTT. However it has a problem that the path-search cost by shortcuts is too expensive to achieve a reasonable success rate. This paper proposes our newly improved method that finds shorter-paths with both reasonable cost and high success rate of path-search by short-cuts and discusses some of experimental results achieved by software simulation. The results illustrate the validity of our proposed method.

Language：English   Publishing type：Research paper (other academic)  

When searching for intended pages on the Internet, users often have a hard time to find the pages because the pages do not always come at the higher rank of searched results. The Personalized Search is a promising approach to solve this problem. In the Personalized Search, User Profiles (UPs in short) that represent interests of the users, are well used and often created from personal documents of the users. This paper proposes (1) a method for making UPs based on Open Directory Project (ODP) and shows (2) a Personalized Search system using the UPs made from Book Marks. Some of experimental results illustrate the validity of our method for making the UPs, and show the precision enhancement of this system.

DOI： 10.1007/978-3-540-89197-0_78

Language：Japanese   Publishing type：Research paper (scientific journal)  

Language：Japanese   Publishing type：Research paper (scientific journal)  

Publishing type：Research paper (scientific journal)  

Language：Japanese   Publishing type：Research paper (scientific journal)  

Language：English   Publishing type：Research paper (other academic)  

Business matching and collaboration support systems are useful, in particular for small-and-medium sized companies. Most of them developed so far are based on the server-client architecture and provide their services on Web servers. They require special administrative facilities, ask users to upload their data for matching between business needs and seeds, and leave to themselves peer-to-peer communication or negotiation between matched companies. Considering these problems, we have been developing an agent-community-network- based business matching and collaboration support system. Our system requires neither any special administrative facilities nor uploading user data to a special server. It furthermore supports secure peer-to-peer communication between users. It is implemented with multi-agent Kodama framework.

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.11309/jssst.25.4_3

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.11309/jssst.25.4_1

Language：English   Publishing type：Research paper (other academic)  

In web retrieval, it is often the case that the results given by search engines are not just what we want. To solve this problem there have been many studies on improving queries to be submitted to search engines. However, they are still insufficient due to lack of consideration on time information. To remedy this we propose a user-schedule-based web page recommendation method. The method makes use of a user's schedule to make the recommendation suite to his/her requests. In addition, an algorithm for extracting related words is introduced as a key technique in this method. Some preliminary experiments show very promising results in recommending web pages relevant to users requests. We also confirmed that this algorithm outper-forms a method using mutual information.

DOI： 10.1109/WI.2007.4427188

Language：English   Publishing type：Research paper (other academic)  

The Agent-Community-based Peer-to-Peer Information Retrieval (ACP2P) method[1],[2] uses agent communities to manage and look up information of interest to users. An agent works as a delegate of its user and searches for information that the user wants by communicating with other agents. The communication between agents is carried out in a peer-to-peer computing architecture. Retrieving information relevant to a user query is performed with content files which consist of original and retrieved documents, and two histories: a query/retrieved document history and a query/sender agent history. The ACP2P is implemented using the Multi-Agent Kodama framework. In this paper, we present some mathematical aspects of the ACP2P method with respect to the relationships between communication loads and the number of records that are stored both in the two histories and retrieved document content files, and discuss the experimental results, for which illustrate the validity of this approach. The results confirm the mathematical conjectures we presented and show that the two histories are more useful for reducing the communication load than a naive method employing 'multicast' techniques, and lead to a higher retrieval accuracy than the naive method.

DOI： 10.1007/11925941_12

Language：English   Publishing type：Research paper (other academic)  

The Agent-Community-based Peer-to-Peer Information Retrieval (ACP2P) method uses agent communities to manage and look up information of interest to users. An agent works as a delegate of its user and searches for information that the user wants by communicating with other agents. The communication between agents is carried out in a peer-to-peer computing architecture. The ACP2P is implemented using the Multi-Agent Kodama framework. This paper presents how the ACP2P method works in an agent community network and show the experimental results to illustrate the validity of this approach.

DOI： 10.1145/1160633.1160878

Language：English   Publishing type：Research paper (scientific journal)  

The Agent-Community-based Peer-to-Peer (ACP2P) information retrieval method uses agent communities to manage and look up information of interest to users. An agent works as a delegate of its user and searches for information that the user wants by communicating with other agents. The communication between agents is carried out in a peer-to-peer computing architecture. Retrieving information relevant to a user query is performed with content files which consist of original and retrieved documents, and two histories: a query/retrieved document history and a query/sender agent history. Making use of the histories has a collaborative filtering effect, which gradually creates virtual agent communities, where agents with the same interests stay together. Our hypothesis is that a virtual agent community reduces communication loads necessary to perform a search. As an agent receives more queries, then more links to new knowledge are acquired. From this behavior, a "give-and-take" (or positive feedback) effect for agents seems to emerge. However, we have only shown parts of the effects of reducing communication loads and making "give-and-take" through preliminary experimental results. This paper discusses more detail of experimental results and shows that the two histories help in reducing communication loads among agents, facilitating bidirectional communications between them and thus creating virtual agent communities, where agents with the same interests stay together.

DOI： 10.1002/scj.20625

Language：English   Publishing type：Research paper (scientific journal)  

In this paper, we mainly focus on a communication mechanism which enables efficient information publishing and sharing among semantic desktops. We propose MyPortal as a "one stop" for all the information relevant to the user and further propose the conceptual architecture of a P2P community Web information retrieval system based on MyPortal. This architecture enables not only precise location of My-Portal instances and their Web resources but also the automatic or semiautomatic integration of hybrid semantic information delivered through Web content and Web services, and it also ensures that the semantics will not be lost during any part of the lifecycle of the information retrieval process.

Language：English   Publishing type：Research paper (other academic)  

The semantic Web and Web service technologies have provided both new possibilities and challenges to automatic information processing. There are a lot of researches on applying these new technologies into current personal Web information retrieval systems, but no research addresses the semantic issues from the whole life cycle and architecture point of view. Web services provide a new way for accessing Web resources, but until now, they have been managed separately from traditional Web contents resources. In this poster, we propose a conceptual architecture for a personal semantic Web information retrieval system. It incorporates semantic Web, Web services and multi-agent technologies to enable not only precise location of Web resources but also the automatic or semi-automatic integration of hybrid Web contents and Web services.

DOI： 10.1145/1062745.1062825

Language：English   Publishing type：Research paper (other academic)  

The semantic Web and Web services technologies have provided both new possibilities and challenges to automatic information processing. There are a lot of researches on applying these new technologies to current personal Web information retrieval systems, but no research addresses the semantic issues from the whole life cycle and architecture point of view. Web services provide a new way for accessing Web resources, but until now, they have been managed separately from conventional Web contents resources. In this paper, we point out new system requirements and propose a conceptual architecture for a personal semantic Web information retrieval system. It incorporates semantic Web, Web services and multi-agent technologies to enable not only precise location of Web resources but also the automatic or semi-automatic integration of hybrid Web contents and Web service resources.

DOI： 10.1109/ICWS.2005.23

Language：English   Publishing type：Research paper (other academic)  

In this paper, we propose a conceptual architecture for a personal semantic Web information retrieval system. It incorporates semantic Web, Web service, P2P and multi-agent technologies to enable not only precise location of Web resources but also the automatic or semi-automatic integration of Web resources delivered through Web contents and Web services. In this architecture, the semantic issues concerning the whole lifecycle of information retrieval were considered consistently and the integration of Web contents and Web services is enabled seamlessly.

Language：English   Publishing type：Research paper (other academic)  

This paper proposes an agent community based information retrieval method, which uses agent communities to manage and look up information related to users. An agent works as a delegate of its user and searches for information that the user wants by communicating with other agents. The communication between agents is carried out in a peer-to-peer computing architecture. In order to retrieve information relevant to a user query, an agent uses two histories: a query/retrieved document history(Q/RDH) and a query/sender agent history(Q/SAH). The former is a list of pairs of a query and retrieved document information, where the queries were sent by the agent itself. The latter is a list of pairs of a query and the address of a sender agent and shows "who sent what query to the agent". This is useful for finding a new information source. Making use of the Q/SAH is expected to have a collaborative filtering effect, which gradually creates virtual agent communities, where agents with the same interests stay together. Our hypothesis is that a virtual agent community reduces communication loads involved in performing a search. As an agent receives more queries, then more links to new knowledge are acquired. From this behavior, a "give and take" (or positive feedback) effect for agents seems to emerge. We implemented this method with Multi-Agent Kodama, and conducted experiments to test the hypothesis. The empirical results showed that the method was much more efficient than a naive method employing 'multicast' techniques only to look up a target agent.

DOI： 10.1007/11574781_6

Language：English   Publishing type：Research paper (scientific journal)  

Due to the problem of information overload, locating relevant Web portals precisely based on user requirements is quite an essential task. As the need for application-to-application communication and interoperability grows, providing Web portal services that satisfy human as well as machine requirements is becoming a new challenge for Web portals. However, a Web portal capability expressing mechanism, which enables the precise location of Web portals as well as the automated discovery and invocation of Web portal services, is lacking. In this paper, we investigate how to incorporate Semantic Web technology with Web service technologies to describe the capabilities of Web portals. We also discuss the possibilities of using these descriptions for discovering and using the distributed existing portal resources.

Language：English   Publishing type：Research paper (scientific journal)  

This paper proposes an agent community based information retrieval method, which uses agent communities to manage and look up information related to users. An agent works as a delegate of its user and searches for information that the user wants by communicating with other agents. The communication between agents is carried out in a peer-to-peer computing architecture. In order to retrieve information related to a user query, an agent uses two histories: a query/retrieved document history(Q/RDH) and a query/sender agent history(Q/SAH). The former is a list of pairs of a query and retrieved documents, where the queries were sent by the agent itself. The latter is a list of pairs of a query and sender agents and shows "who sent what query to the agent". This is useful to find a new information source. Making use of the Q/S AH is expected to cause a collaborative filtering effect, which gradually creates virtual agent communities, where agents with the same interests stay together. Our hypothesis is that a virtual agent community reduces communication loads to perform a search. As an agent receives more queries, then more links to new knowledge are achieved. From this behavior, a "give and take"(or positive feedback) effect for agents seems to emerge. We implemented this method with Multi-Agents Kodama which has been developed in our laboratory, and conducted preliminary experiments to test the hypothesis. The empirical results showed that the method was much more efficient than a naive method employing 'broadcast' techniques only to look up a target agent.

DOI： 10.1527/tjsai.19.421

Language：Japanese   Publishing type：Research paper (scientific journal)  

Tsunenori Mine, Daisuke Matsuno and Makoto Amamiya:
Agent-Community-based Peer-to-Peer Information Retrieval(in Japanese),
Japanese Society of Artificial Intelligence Journal J-STAGE, http://tjsai.jstage.jst.go.jp/ja/, vol. 19, no. 5, pp. 421--428, 2004

Language：English   Publishing type：Research paper (other academic)  

This paper presents an agent community based peer-to-peer information retrieval method called ACP2P method[16] and discusses the experimental results of the method. The ACP2P method uses agent communities to manage and look up information related to users. An agent works as a delegate of its user and searches for information that the user wants by communicating with other agents. The communication between agents is carried out in a peer-to-peer computing architecture. In order to retrieve information relevant to a user query, an agent uses a content file, which consists of retrieved documents and two histories: a query/retrieved document history(Q/RDH) and a query/sender agent history(Q/SAH). The former is a list of pairs of a query and the address of an agent that returned documents relevant to the query. The latter is a list of pairs of a query and the address of a sender agent and shows "who sent what query to the agent". This is useful for finding a new information source. Making use of Q/SAH is expected to have a collaborative filtering effect, which gradually creates virtual agent communities, where agents with the same interests stay together. Our hypothesis is that a virtual agent community reduces communication loads necessary to perform a search. As an agent receives more queries, then more links to new knowledge are acquired. From this behavior, a "give and take"(or positive feedback) effect for agents seems to emerge. We implemented this method with Multi-Agent Kodama, and conducted experiments to test the hypothesis. The experimental results showed that the method employing two histories was much more efficient than a naive method employing 'multicast' techniques only to look up a target agent. Further, making use of Q/SAH facilitates bidirectional communications between agents and thus creates virtual agent communities.

Language：English   Publishing type：Research paper (other academic)  

This paper proposes an Agent Community based Peer-to-Peer information retrieval method called ACP2P method, which uses agent communities to manage and look up information related to users. An agent works as a delegate of its user and searches for information that the user wants by communicating with other agents. The communication between agents is carried out in a peer-to-peer computing architecture. In order to retrieve information related to a user query, an agent uses a content file, which consists of retrieved documents, and two histories: a query/retrieved document history(Q/RDH) and a query/sender agent history(Q/SAH). We implemented this method with Multi-Agents Kodama, and conducted preliminary experiments to test the hypothesis. The empirical results showed that the method was much more efficient than a naive method employing 'multicast' techniques only to look up a target agent.

Language：English   Publishing type：Research paper (other academic)  

Next generation of information systems will rely on cooperative intelligent agents for playing a fundamental role in actively searching and finding relevant information on behalf of their users in complex and open environments, such as the Internet. User Interface Agents {VIA) are semi-intelligent systems, which help the users to access, manage, share and exchange information. Recently, various researchers have proposed a learning approach towards building such agents and some working prototypes have been demonstrated. Such agents learn by watching over the shoulder of the user and detect patterns and regularities in the user's behavior. We present a new approach of the collaborative VIA that helps the user to retrieve information that is consistent to the user's need. The model provides tools and utilities for the user to manage his/her information repositories with dynamic organization and adaptation views. In order to investigate the performance of the UIA, we carried out several experiments. Through the experiments, the results ensure that the techniques of personalization, clustering the user's preferences, and making use of the preferences promise to achieve more relevant information to the user's queries.

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (other academic)  

From the 1990s on, one of the most important challenges facing computer science researchers has been the design and construction of software tools to exploit Internet computing. At the same time, the development of agent technology has gone hand in hand with the explosion of the Internet. As worldwide network computing environments become more and more complex, software agents are believed to have the potential to help present and manage the Internet in an autonomous or semi-autonomous way. Yet, to date, a number of fundamental questions about the theory and practice of this new software engineering paradigm have remained unanswered. Here we explore the features that make the agent-based approach such an appealing and evolutionary computational model. In particular, we envision a global agent-based distributed computing architecture that provides a convenient programming abstraction and sufficient transparency. This paper gives a general introduction to the underlying concepts of our research and development both at the level of design philosophy and in practical implementation techniques. It is argued that the shift from computer networks to agent networks is a significant extension of network programming technology because agents are well suited to modeling, designing and implementing scalable, flexible and secure distributed systems over a worldwide computing environment.

DOI： 10.1109/HICSS.2003.1174210

Language：English   Publishing type：Research paper (scientific journal)  

Due to the spread of distributed systems, a number of agent systems have been developed. Since most existing multiagent systems implement different agent platforms, it is extremely hard to get heterogeneous agents to work together. In this paper, we introduce a new network protocol called Agent Platform Protocol (APP), which is designed to meet the exact demands of agent interaction over world-wide networks. The most significant contribution of this protocol is that it supports Peer-to-Peer communication among agent platforms and hides the network-level implementation details from agents. To demonstrate its usefulness, APP has been used to realize the interaction between independent agent systems from two projects, JADE and KODAMA. JADE is an agent system in compliance with the FIPA specifications, and KODAMA is an agent system one being developed in our laboratory. The experimental results show that APP can help JADE and KODAMA agents to communicate with each other without getting the agents being aware of he difference in their agent platforms.

DOI： 10.1541/ieejeiss.123.1503

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (scientific journal)  

Language：English  

Language：English   Publishing type：Research paper (scientific journal)  

The academic staff research and education activity database of Kyushu University is a document database system which enables each academic staff member to input and update his or her activity report, administrators to maintain the database, and visitors to search and read reports via the Internet. Those databases and user interfaces can handle mixed report formats by storing metadata of formats and constraints. Through an analysis of the update history of the system and e-mail records between users and developers, we confirmed that distributed administration style reduced human work load and that the majority of bugs are confined to small components. © 2002 Wiley Periodicals, Inc. Syst. Comp. Jpn.

DOI： 10.1002/scj.10043

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English  

Language：English  

Language：English  

Language：English   Publishing type：Research paper (scientific journal)  

Language：English   Publishing type：Research paper (scientific journal)  

Agent-based computing has been advocated by many researchers and software developers as a promising and innovative way to model, design and implement complex Web-related applications. KODAMA (Kyushu university Open & Distributed Autonomous MultiAgent) project, which is described in this paper, is our endeavour to advance both the technology and engineering of well-known multiagent systems. In particular, we have noted that software agents might be the potential solution to many problems faced by today's Web. However, building a high-quality, large-scale multiagent system, which can operate in open environments. is a great challenge. So far. we have devoted a lot of effort to design and implement a generic agent architecture, a hierarchical agent community structure, and an independent network communication middleware. To ensure that KODAMA can be used to create Web-agent applications, its network communication performance and a prototype distributed database retrieval system have been tested. The result shows that KODAMA is suitable for developing network-aware applications,

Language：English   Publishing type：Research paper (other academic)  

With increasing access to the Internet and the wealth of material online, a Web-based self-teaching system has considerable educational value. Accordingly, we developed AEGIS (Automatic Exercise Generator based on the Intelligence of Students), which automatically generates questions whose difficulty level fits the achievement level of a student. However, it was implicitly assumed that all the questions were already categorized according to their subjects. In practice, this is not the case, but it is unreasonable (because of time and cost) to expect teachers to categorize each question into a suitable subject domain. Therefore, we need a method for categorizing questions automatically according to specified teaching concepts. This paper presents an automatic question categorization mechanism according to both a list of teaching concepts, called a Navigation List (NaviList for short), and the meaning of questions. We define an XML tag called a CONCEPT tag, which indicates a concept in a question, and an ontology, which is a hierarchical cluster of concepts. The method uses the tags and the ontology to categorize questions, based on the similarity between each category in a NaviList pre-composed by a teacher and an ontological concept specified by a CONCEPT tag in a question.

DOI： 10.1109/CIE.2002.1186203

Language：English   Publishing type：Research paper (other academic)  

This paper presents a text mining system that obtains the relationships between the topics of international conferences. The system at first gathers Web pages related to the query 'call for paper' with a search engine and filters out pages irrelevant to the query with SVMs. Next it extracts the topics of conferences and clusters them. Each cluster embodies the relationships between topics and between topics and conferences. The clustered topics are shown through an Explorer-like graphical user interface. The preliminary experimental results promise that the method works not only for obtaining the relationship between topics of conferences, but also for discovering the relationship between any information entities users are interested in.

DOI： 10.1109/DEXA.2002.1045899

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：Japanese  

Language：English  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：Japanese   Publishing type：Research paper (scientific journal)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (other academic)  

We are aiming to develop a personal assistant system which handles its user’s files stored in his/her computers and his/her interesting information that can be accessed on the Web. The system categorizes the files and Web pages gathered and extracts information specified by him/her and stores it into a structured database for use as the knowledge of its dialogue module. This paper presents a prototype of the system that handles information about conferences of interest to the user. The system extracts conference names, submission deadlines, dates, URIs and locations from e-mail messages the user has received, and web pages gathered by both crawling and meta-search. Extracted information is stored into a database so that the user can interactively search conference information via a user interface with natural language queries.

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：Japanese  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English  

The main thrust of the KODAMA research project is an investigation into novel ways of agentifying the Web based on the pre-existing hyper-link structure, and how this community of agents can automatically achieve and update its interpretation policies and cooperate with other agents to retrieve online distributed relevant information on the Web. The paper focuses on the use of the user interface agents in the KODAMA system for personalized information filtering and adapting, and discusses the method to update a user profile and interpretation policies adaptively for the user. This is an ideal and challenging environment for interface agents. The proposed idea is to employ an adaptive autonomous user interface agent that works for satisfying a user's information needs cooperatively with other agents in the KODAMA system, Web page agents and server agents.

DOI： 10.1109/DEXA.2000.875014

Language：English   Publishing type：Research paper (other academic)  

Agent technology is becoming more prevalent as the availability of network access, and the demand for the end uses of agents becomes greater. Intelligent agents for information filtering and retrieval applications and tools are being employed in a variety of ways on the Web. A centralized agent for information discovery has usually limited capabilities for finding diverse and distributed information online. The main thrust of the paper is to present a framework that allows distributed adaptive information KODAMA agents to work together to browse and retrieve distributed information based on the preexisting hyperlink structure on the Web and how this community of agents can automatically extract meta-information and cooperate to retrieve online distributed relevant information. We have developed the software architecture, and a working prototype showing the benefits to the Web of interactive architectures based on the coordination of the hyperlink structure on the network. The authors summarize the current results of the project, and discuss some ideas on future work.

DOI： 10.1109/PADSW.2000.884668

Language：English   Publishing type：Research paper (other academic)  

With the explosion of the Internet, we are evolving a worldwide network computing environment. At this point, the surged challenge is the next evolutionary technology for the Internet-oriented applications. A KODAMA (Kyushu university Open and Distributed Autonomous Multi-Agent) system is deployed for this demand. KODAMA system can interconnect disparate and distributed agents together to solve some problem whose solution is typically beyond any singular agent's capabilities. The key aspects of KODAMA agents are their autonomy, collaboration, flexibility, as well as stability. To obtain all these features, an appropriate scheme of collaboration among agents is clearly critical. Our new approach focuses on both raising the level of abstraction at which agents cooperate with each other, and hiding the implementation details of communications from agents.

DOI： 10.1109/PADSW.2000.884664

Language：English   Publishing type：Research paper (other academic)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：Japanese  

Language：Japanese  

Language：Japanese  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (international conference proceedings)  

Language：English   Publishing type：Research paper (scientific journal)  

Language：English   Publishing type：Research paper (other academic)  

Mixture-Preference Bayesian Matrix Factorization for implicit feedback datasets

Event date： 2025.3

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：長崎市   Country：Japan  

Event date： 2025.2

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：名古屋工業大学   Country：Japan  

Event date： 2025.2

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：名古屋工業大学   Country：Japan  

Event date： 2025.2

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：名古屋工業大学   Country：Japan  

Event date： 2024.12

Language：English   Presentation type：Oral presentation (general)  

Venue：Okayama University   Country：Japan  

Event date： 2024.12

Language：English   Presentation type：Oral presentation (general)  

Venue：Okayama University   Country：Japan  

Event date： 2024.12

Language：English   Presentation type：Oral presentation (general)  

Venue：Okayama University   Country：Japan  

Event date： 2024.12

Language：English   Presentation type：Oral presentation (general)  

Venue：Okayama University   Country：Japan  

Event date： 2024.12

Language：English  

Venue：Bangkok, Thailand   Country：Thailand  

Event date： 2024.9

Language：Japanese  

Venue：北海道札幌市   Country：Japan  

Event date： 2024.7

Language：English   Presentation type：Oral presentation (general)  

Venue：Takamatsu, Kagawa   Country：Japan  

Event date： 2024.7

Language：English   Presentation type：Oral presentation (general)  

Venue：Takamatsu, Kagawa   Country：Japan  

Event date： 2024.7

Language：English   Presentation type：Oral presentation (general)  

Venue：Osaka   Country：Japan  

Other Link： https://ieeecompsac.computer.org/2024/

Event date： 2024.6

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：岩手県花巻市   Country：Japan  

Event date： 2024.6

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：岩手県花巻市   Country：Japan  

Event date： 2024.6

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：岩手県花巻市   Country：Japan  

Event date： 2024.3

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：大阪   Country：Japan  

Event date： 2024.3

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：兵庫県三宮   Country：Japan  

Event date： 2024.3

Language：Japanese  

Venue：東京一橋会館   Country：Japan  

Event date： 2024.3

Language：Japanese  

Venue：東京一橋会館   Country：Japan  

Event date： 2024.3

Language：Japanese  

Venue：東京一橋会館   Country：Japan  

Event date： 2024.3

Language：Japanese  

Venue：東京一橋会館   Country：Japan  

Event date： 2024.3

Language：Japanese  

Venue：長崎大学   Country：Japan  

Event date： 2023.12

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：富山県富山市 富山国際会議場   Country：Japan  

Event date： 2023.12

Language：English   Presentation type：Oral presentation (general)  

Venue：Singapore   Country：Singapore  

Event date： 2023.12

Language：English   Presentation type：Oral presentation (general)  

Venue：HongKong   Country：China  

Other Link： https://paclic2023.github.io/

Event date： 2023.11 - 2023.12

Language：English  

Venue：京都市   Country：Japan  

Event date： 2023.11

Language：English  

Venue：Fuzhou   Country：Japan  

Event date： 2023.11

Language：English   Presentation type：Oral presentation (general)  

Venue：Jakarta   Country：Indonesia  

Other Link： https://www.pricai.org/2023/

Event date： 2023.11

Language：English   Presentation type：Oral presentation (general)  

Venue：Jakarta   Country：Indonesia  

Other Link： https://www.pricai.org/2023/

Event date： 2023.11 - 2024.11

Language：Japanese  

Venue：福井県大野市 大野商工会議所 第1・2研修室   Country：Japan  

Event date： 2023.10

Language：English   Presentation type：Oral presentation (general)  

Venue：Venice   Country：Italy  

Other Link： https://www.wi-iat.com/wi-iat2023/index.html

Event date： 2023.10

Language：Japanese   Presentation type：Oral presentation (general)  

Country：Japan  

Event date： 2023.10

Language：English   Presentation type：Oral presentation (general)  

Venue：University of Birmingham and Eastside Rooms, UK   Country：United Kingdom  

Other Link： https://uobevents.eventsair.com/cikm2023//

Event date： 2023.10

Language：English   Presentation type：Oral presentation (general)  

Venue：Honolulu, Hawaii   Country：United States  

Other Link： https://www.google.com/url?q=https://ieeesmc2023.org/&sa=D&source=calendar&ust=1691719640241730&usg=AOvVaw3Ly5NqqzaHjLgpj6be-BCX

Event date： 2023.7

Language：English   Presentation type：Oral presentation (general)  

Venue：Hitotsubashi, Tokyo   Country：Japan  

Event date： 2023.6

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：国立情報学研究所 19階 　1902/03会議室   Country：Japan  

Event date： 2023.5

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：沖縄   Country：Japan  

Other Link： https://ipsj.ixsq.nii.ac.jp/ej/index.php?action=pages_view_main&active_action=repository_view_main_item_snippet&index_id=6201&pn=1&count=20&order=7&lang=japanese&page_id=13&block_id=8

Event date： 2023.2

Language：Japanese   Presentation type：Symposium, workshop panel (public)  

Venue：福岡システムLSI総合開発センター2階　会議室A・B室 福岡市早良区百道浜3-8-33   Country：Japan  

Event date： 2022.12

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：九州大学西新プラザ   Country：Japan  

Event date： 2022.12

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：九州大学西新プラザ   Country：Japan  

Event date： 2022.12

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：九州大学西新プラザ   Country：Japan  

Event date： 2022.12

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：九州大学西新プラザ   Country：Japan  

Event date： 2022.12

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：九州大学西新プラザ   Country：Japan  

Event date： 2022.12

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：九州大学西新プラザ   Country：Japan  

Event date： 2022.12

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：柏の葉カンファレンスセンター   Country：Japan  

Event date： 2022.11

Language：English   Presentation type：Oral presentation (general)  

Venue：Niagara Falls   Country：Canada  

Other Link： https://www.wi-iat.com/wi-iat2022/index.html

Event date： 2022.11

Language：English   Presentation type：Oral presentation (general)  

Venue：オンライン Zoom   Country：Indonesia  

Other Link： http://sitsefmipa.unimed.ac.id/

Event date： 2022.10

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：皆生温泉「皆生グランドホテル 天水」   Country：Japan  

Event date： 2022.10

Language：English   Presentation type：Oral presentation (general)  

Venue：Atlanta, Georgia   Country：United States  

Other Link： https://www.cikm2022.org/

Event date： 2022.10

Language：English   Presentation type：Oral presentation (general)  

Venue：Clarion Congress Hotel Prague, Czech Republic   Country：Czech Republic  

Other Link： https://ieeesmc2022.org/

Event date： 2022.9

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：浜松   Country：Japan  

Event date： 2022.9

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：浜松   Country：Japan  

Event date： 2022.7 - 2021.7

Language：English   Presentation type：Oral presentation (general)  

Venue：Durham   Country：United Kingdom  

Other Link： https://aied2022.webspace.durham.ac.uk/

Event date： 2022.7

Language：English   Presentation type：Oral presentation (general)  

Venue：Madrid   Country：Spain  

Event date： 2022.7

Language：English   Presentation type：Oral presentation (general)  

Venue：Kanazawa   Country：Japan  

Other Link： https://iaiai.org/conference/aai2022/conferences/eskm-2022/

Event date： 2022.7

Language：English   Presentation type：Oral presentation (general)  

Venue：Kanazawa   Country：Japan  

Other Link： https://iaiai.org/conference/aai2022/conferences/ltle-2022/

Event date： 2022.7

Language：English   Presentation type：Oral presentation (general)  

Venue：Kanazawa   Country：Japan  

Other Link： https://iaiai.org/conference/aai2022/conferences/eskm-2022/

Event date： 2022.7

Language：English   Presentation type：Oral presentation (general)  

Venue：Kanazawa   Country：Japan  

Other Link： https://iaiai.org/conference/aai2022/conferences/eskm-2022/

Event date： 2022.5

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：沖縄　北谷町   Country：Japan  

Event date： 2022.5

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：沖縄　北谷町   Country：Japan  

Event date： 2022.5

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：沖縄科学技術大学院大学（OIST） and オンライン OIST会場: シドニー・ブレナー レクチャーシアター（セミナールームB250） 〒904-0412 沖縄県国頭郡恩納村谷茶１９１９−１   Country：Japan  

Event date： 2022.4

Language：English   Presentation type：Oral presentation (general)  

Venue：Online   Country：Other  

Event date： 2022.3

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：オンライン Zoom および　愛媛大学   Country：Japan  

Event date： 2022.2

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：オンライン Zoom および　石川県金沢市 金沢勤労者プラザ 204研修室   Country：Japan  

Event date： 2022.2

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：オンライン Zoom および　石川県金沢市 金沢勤労者プラザ 204研修室   Country：Japan  

Event date： 2022.2

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：オンライン Zoom および　石川県金沢市 金沢勤労者プラザ 204研修室   Country：Japan  

Event date： 2021.12

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：zoomによるオンライン開催  

Event date： 2021.12

Language：English   Presentation type：Oral presentation (general)  

Venue：Online   Country：Australia  

Event date： 2021.12

Language：English   Presentation type：Oral presentation (general)  

Venue：Online   Country：Japan  

Event date： 2021.12

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：オンライン   Country：Japan  

Event date： 2021.12

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：オンライン   Country：Japan  

Event date： 2021.12

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：オンライン   Country：Japan  

Event date： 2021.11

Language：English   Presentation type：Oral presentation (general)  

Venue：Online   Country：Thailand  

Event date： 2021.11

Language：English   Presentation type：Oral presentation (general)  

Venue：Online   Country：Thailand  

Event date： 2021.11

Language：English   Presentation type：Oral presentation (general)  

Venue：Tokyo   Country：Japan  

Event date： 2021.10

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：霧島温泉，鹿児島県   Country：Japan  

Event date： 2021.10

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：霧島温泉，鹿児島県   Country：Japan  

Event date： 2021.9

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：オンライン Zoom   Country：Japan  

Event date： 2021.9

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：オンライン Zoom   Country：Japan  

Event date： 2021.6 - 2021.7

Language：English   Presentation type：Oral presentation (general)  

Venue：Paris   Country：France