カワマタ　タイスケ
川又　泰介助教
Taisuke Kawamata

■研究者基本情報

組織

  • 情報戦略機構

研究分野

  • 人文・社会, 教育工学, オンライン受験者認証
  • 情報通信, 知能情報学, ニューラルネットワーク
  • 情報通信, ウェブ情報学、サービス情報学, ビデオオンデマンド

研究キーワード

  • 画像処理
  • 顔認証
  • 筆記認証
  • Webテスト
  • 教育工学
  • eラーニング
  • テキストマイニング
  • バイオメトリクス
  • 機械学習

自身の研究に関連するSDGs

学位

  • 2020年03月　博士（工学）（東京理科大学）
  • 2017年03月　修士（工学）（東京理科大学）

学歴

  • 2017年04月 - 2020年03月, 東京理科大学, 工学研究科, 経営工学専攻
  • 2015年04月 - 2017年03月, 東京理科大学, 工学研究科, 経営工学専攻
  • 2010年04月 - 2015年03月, 東京理科大学, 工学部第二部, 経営工学科
  • 2006年04月 - 2009年03月, 茨城県立鉾田第一高等学校, 普通科

経歴

  • 2023年04月 - 現在, 茨城大学, 情報戦略機構, 助教, データ戦略部門
  • 2022年09月 - 2023年03月, 東京工芸大学, 工学部, 非常勤講師
  • 2022年06月 - 2023年03月, 成蹊大学, 理工学部 理工学科, 助教, データ数理専攻
  • 2022年04月 - 2023年03月, 東京理科大学, 工学研究科, 客員研究員
  • 2020年04月 - 2022年06月, 成蹊大学, 理工学部 情報科学科, 助教

委員歴

  • 2022年06月 - 2026年06月, 専門委員, 電子情報通信学会教育工学研究会 専門委員会
  • 2023年08月 - 2024年03月, 情報・システムソサイエティ学術奨励賞選定委員会 投票委員, 電子情報通信学会
  • 2022年08月 - 2023年03月, 情報・システムソサイエティ学術奨励賞選定委員会 投票委員, 電子情報通信学会

■研究活動情報

受賞

  • 2024年05月, GeoSciAI2024 最優秀賞（宇宙天気課題）, 太陽観測データからの太陽フレア発生予測, 日本地球惑星科学連合＆人工知能学会
    榎本 理恵;川又 泰介;北島 良三;坂本 真仁;橋本 俊之
    国内学会・会議・シンポジウム等の賞
  • 2023年10月, Excellent Poster Awards Outstanding Prize, Writer Alternation Detection by Sequential Hotelling's Theory for e-Testing, IEEE GCCE
    Taisuke Kawamata;Takako Akakura
    国際学会・会議・シンポジウム等の賞
  • 2016年09月, FIT研究奨励賞, e-Learningにおける顔類似度の時系列情報から確率的に本人判定を行う認証法, 情報科学技術フォーラム
    川又 泰介
    国内学会・会議・シンポジウム等の賞
  • 2015年11月, 研究奨励賞, e-Learning受講者の登録顔情報による認証のための各種逐次更新法の検討, 電子情報通信学会教育工学研究会
    川又 泰介
    国内学会・会議・シンポジウム等の賞

論文

  • 教師組織化写像を用いた日本語感情分析
    川又 泰介; ほか, 筆頭著者, 近年，大規模言語モデルを用いた高精度な感情推定手法が提案されているが，その推論過程や内部表現を人間が解釈することは依然として困難である．本研究では，感情分類モデルの内部構造を可視化し，そのクラス構造を解釈可能とすることを目的として，RBFネットワークにプーリング層を組み合わせ，誤差逆伝搬により隠れ層の組織化を行う教師組織化写像を提案する．日本語感情分析データセットWRIMEを用いた評価実験により，提案法はU-matrixによる可視化を通じて，隠れ層上に形成される感情クラスの構造を二次元平面として解釈可能であることを示した．, 情報処理学会
    情報処理学会論文誌, 2026年08月, ［査読有り］
  • 〔主要な業績〕無線LAN認証ログを用いた学生の活動傾向の分析
    川又 泰介; 山本 一幸; 大瀧 保広, 筆頭著者, 従来，学生の活動状況は教務情報に基づいて把握されてきたが，教務情報は修学に付随して発生することから，課外活動やキャンパス内での日常的な行動を捉えるには不十分であるという課題がある．そこで本研究では，大学の情報基盤に記録される無線LANアクセスポイント(AP)の認証ログを用いることで，学生の活動傾向を補足することが可能か検証を行う．分析のために，APの認証成功ログに対してWord2Vecによる分散表現を適用し，APを実数空間に埋め込むことで，キャンパス内の構造に対応した位置関係を構成した．さらに，この埋め込み空間を用いて学生の移動履歴を時系列的に分析し，逸脱度の推移に基づいてクラスタリングを行った．その結果，特定の認証情報と学生の属性・行動傾向との関連が確認された．, 大学ICT推進協議会
    学術情報処理研究, 2025年11月, ［査読有り］
  • A Teacher Organization Map with Pooling SOM for Writer Verification in Online Examination
    Taisuke Kawamata; Takako Akakura, 筆頭著者, To develop explainable artificial intelligence, we developed a SOM-like layer using a convolutional neural network architecture. This pseudo-SOM is integrated into the MLP and named Teacher Organization Map (TOM). TOM applied for writer identification to prevent proxy test-taking in e-Testing. Our results show that the verification accuracy is nearly 90% while maintaining explainability in the hidden layer., Institute of Electrical and Electronics Engineers Inc.
    Proceedings of 2025 IEEE Global Conference on Consumer Electronics (GCCE2025), 2025年09月, ［査読有り］
  • An Emotion Classification System Considering Analytic Hierarchy Process
    Yurika Nagamori; Kiichi Honsho; Hiroaki Narisawa; Shinji Sakamoto; Taisuke Kawamata, ラスト(シニア)オーサー, Emotion estimation systems have been investigated and developed for a long time due to their usefulness in various industrial applications. However, emotions estimation is a complex system. In this reason, we begin with a simple model that classifies a sentence into eight emotions based on Plutchik’s circle. This paper proposes an Analytic Hierarchy Process (AHP) based emotions estimation system. We evaluate the proposed system by comparing the conventional system, which is the Word2Vec-based emotions estimation system. The evaluation results show that the order of emotions in both systems are not big difference in the considered sentences. After getting a vector of words, AHP has explainable feature, also setting the weight coefficient to emotions. Moreover, AHP is suitable for managing individual emotions. This makes the proposed system valuable and easy to use in other industrial applications., Springer
    Advances in Internet, Data and Web Technologies. EIDWT 2025. Lecture Notes on Data Engineering and Communications Technologies, 2025年04月, ［査読有り］
  • Evaluation of the variance in an eye-tracking system considering the number of cameras
    Kiichi Honsho; Yurika Nagamori; Shinji Sakamoto; Taisuke Kawamata, ラスト(シニア)オーサー, Eye-tracking technology is becoming more important and widely used in various fields, such as computer-human interaction, medical rehabilitation, user behaviour analysis, marketing, automotive, training, and entertainment. In this work, we would like to improve the estimation accuracy easily and at low cost. So, we use multiple web cameras to estimate the gaze. This paper clarifies the estimation variance in a gaze estimation system by considering the number of cameras as a preliminary experiment. Our experimental results showed that the estimation variance did not decrease as the number of cameras increased. However, we found that the camera position affects the accuracy and variance of gaze estimation also in multiple camera system., Springer
    Advances in Internet, Data and Web Technologies. EIDWT 2025. Lecture Notes on Data Engineering and Communications Technologies, 2025年04月, ［査読有り］
  • 〔主要な業績〕問合わせメールの件名を用いた分類
    川又 泰介; 山本 一幸; 大瀧 保広, 筆頭著者, 情報系センターにおける業務の一つである問合わせ対応の効率化を目的に，問合わせデータを機械的にカテゴライズする方法を検討した．問合わせはメールで行われることが多いが，メールの本文には問合わせと無関係なノイズが多く含まれ，それらのノイズを前処理で削除するための工程が，自動的なカテゴライズを困難なものにしている．そこで本研究では，メールの件名が本文の要約であり，また短文であることからノイズが少ないという点に着目し，件名の活用方法について検討を行う．短文に含まれる単語情報を効果的に活用するため，文書の文脈を考慮することが可能な深層言語モデルBERTを用いて特徴抽出を行った．実際の問い合わせデータを対象に，抽出した特徴をクラスタリングし，人間のカテゴライズと比較した．その結果，件名を用いた分類は本文を用いた分類よりも意味としてまとまったクラスタを生成することが明らかになった．これにより，件名を活用することで，メール分析における前処理の課題を解決できる可能性が示唆された．, 大学ICT推進協議会
    学術情報処理研究, 2024年11月, ［査読有り］
  • Writer Verification Using Cluster Search Trees for e-Testing
    Taisuke Kawamata; Takako Akakura, 筆頭著者, This study addresses the risk of proxy test-taking in e-testing by developing a tablet-based written response system to prevent substitute answers. We examined how the method employs cluster search, which is equivalent to binary search using hierarchical clustering, to efficiently authenticate examinees based on stroke patterns. The study explores stroke authentication using minimum distance, focusing on efficiently locating reference strokes in datasets with varying stroke counts. Experimental results across Kanji and numeric datasets demonstrate that while cluster search showed slightly lower accuracy than linear search, it offered reduced computational complexity., Institute of Electrical and Electronics Engineers Inc.
    Proceedings of 2024 IEEE Global Conference on Consumer Electronics (GCCE2024), 2024年11月, ［査読有り］
  • 〔主要な業績〕研究業績情報を用いた大学教員の階層的分類
    川又 泰介, 筆頭著者, 大学の所有する研究情報を可視化するために，教員の研究業績に付与されたタイトルから抽出した言語特徴に階層的クラスタリングを適用し，得られた構造を二分探索することによる教員のクラスタ検索機能を実装した．researchmapから収集した茨城大学所属教員の業績タイトルに対してSentenceBERTによる特徴抽出とWard法による分類を実行した結果，大学が定めた学部・学科所属とは異なる分類結果が得られたが，実際の研究内容を踏まえると妥当なクラスタリング結果が得られた．また，業績タイトルによる教員検索精度を評価した結果，筆頭著者以外の研究業績と，共同研究者の存在が，検索の精度に影響を及ぼすことが明らかになった．, 情報処理学会
    情報処理学会論文誌デジタルプラクティス, 2024年10月, ［査読有り］
  • 〔主要な業績〕多肢選択式Webテストにおける話者照合を用いた受験者認証
    川又 泰介; 服部 翼; 加納 徹; 赤倉 貴子, 筆頭著者, 多肢選択式のWebテストにおける替玉受験の防止を目的として，汎用マイクからの音声回答に話者照合法を適用する認証システムを開発した．実験の結果より提案手法は，筆記解答認証を用いた先行研究よりも認証精度が高く，筆記解答認証を用いた先行研究よりも心的負担が低いという結果が明らかになった．, 電子情報通信学会
    電子情報通信学会論文誌D, 2024年03月, ［査読有り］
  • Writer Alternation Detection by Sequential Hotelling's Theory for e-Testing
    Taisuke Kawamata; Takako Akakura, 筆頭著者, To prevent proxy test-taking during unsynchronized e-Testing, we examined a method for examinee alternation detection based on online handwriting information, incremental statistics, and Hotelling's theory. In this study, Hotelling's T2and Hotelling's transform (PCA) were compared, and the findings indicated that T2 had a lower calculation complexity and higher accuracy than PCA. In addition, the results showed that PCA reduced the rejection rate of earnest students., Institute of Electrical and Electronics Engineers Inc.
    Proceedings of 2023 IEEE Global Conference on Consumer Electronics (GCCE2023), 2023年10月, ［査読有り］
  • Proposal for Writing Authentication Method Using Sequential Characters in e-Testing
    Riku Tatsukawa; Toru Kano; Taisuke Kawamata; Takako Akakura, This study focuses on countermeasures against spoofing. Previous studies have confirmed that personal authentication, using handwriting information, is possible in e-Testing. However, they could only be applied to e-Testing in which the answer is given in one digit (0-9) because the authentication was performed using one digit. This study created an e-Testing system that answers 2-6 digits and proposed a writing authentication method using sequential characters in e-Testing with help of a pen-tablet PC. The experiment yielded that when authenticating with two characters (bigrams) instead of one character (unigrams), the accuracy was higher. Furthermore, it was confirmed that personal authentication is possible even in e-Testing that answers two digits or more., Institute of Electrical and Electronics Engineers
    Proceedings of 2023 IEEE Global Conference on Consumer Electronics (GCCE2023), 2023年10月, ［査読有り］
  • 〔主要な業績〕オンライン教育環境における顔画像と筆記情報を用いたマルチモーダル疲労測定手法の提案
    川又 泰介; 赤倉 貴子, 筆頭著者, 自学自習が主となるeラーニングは，以前より孤独感や作業環境に起因する疲労が発生しやすいという課題が挙げられている．疲労による学習効率低下を防止するため，著者らは学習者の疲労状態測定システムの開発を行う．本稿ではそのために，普及の進むタブレットPCで取得できる情報として，Webカメラから取得した顔情報とデジタルペンから取得した筆記特徴を用いたマルチモーダル疲労測定手法を提案し，実験により性能を評価した結果，顔と筆記情報の結合により精度向上の可能性が示唆された．, 電子情報通信学会
    電子情報通信学会論文誌D, 2023年02月, ［査読有り］
  • Writer Alternation Detection for Online Exam by Exponential Moving PCA
    Taisuke Kawamata; Takako Akakura, 筆頭著者, Many online handwriting authentications have been developed to prevent proxy-test taking in unsynchronized online exams. Here, we propose an unsupervised anomaly detection method based on principal component analysis calculated from incremental statistics to develop an examinee alternation detection system without template information or pre-training. The evaluation results show that the accuracy of this system is over 90%., Institute of Electrical and Electronics Engineers
    Proceedings of 2022 IEEE Global Conference on Consumer Electronics (GCCE2022), 2022年10月, ［査読有り］
  • Automatic Evaluation of Learning Behaviors for Online Lectures by OpenPose
    Taisuke Kawamata; Takako Akakura, 筆頭著者, To prevent learners from doing something unrelated during online lectures, we propose an automatic evaluation method for learning behavior. This method extracts the pose of learners and their teacher by OpenPose, constructs a model that estimates the teacher's pose at j seconds in an online lecture from that of a learner, and evaluates them by measuring the difference between the estimated and actual values. The result of an experiment showed that this could detect unrelated actions in about 90% of instances., Institute of Electrical and Electronics Engineers
    Proceedings of IEEE Global Conference on Life Sciences and Technologies (LifeTech2022), 2022年03月, ［査読有り］
  • Analysis of Computer Science Textbooks by Topic Modeling and Dynamic Time Warping
    Taisuke Kawamata; Yoshitatsu Matsuda; Takayuki Sekiya; Kazunori Yamaguchi, 筆頭著者, The design of appropriate curricula in the rapidly expanding field of computer science (CS) is one of the most important issues in higher education, and there are many features to be considered. To provide useful information for developing appropriate curricula, we investigated the actual textbooks traditionally used in CS by applying the latent Dirichlet allocation to actual textbooks. In this paper, we focused on the topic progression of the textbooks and showed the results of clustering analysis with sequence clustering. We discovered that the structure of CS textbooks can be explained by five or six topics, and that dynamic time warping (DTW) can classify knowledge areas for CS more appropriately than the LDA-based clustering without DTW., Institute of Electrical and Electronics Engineers
    Proceedings of 2021 IEEE International Conference on Teaching, Assessment, and Learning for Engineering (TALE2021), 2021年12月05日, ［査読有り］
  • Face Authentication for E-testing Using Sequential Reference Update
    Taisuke Kawamata; Takako Akakura, 筆頭著者, In an e-testing system, spoofing may occur if the teacher is unable to see examinees during tests. We examined variations in facial images captured by a webcam during e-testing with the aim of detecting impersonation. In previous work, we proposed an authentication method based on updating an examinee’s facial images, which is limited to VOD-based e-learning. We thus optimized its application to e-testing in this paper. Our findings indicate that the proposed method improves authentication accuracy., Institute of Electrical and Electronics Engineers
    Proceedings of 2020 IEEE International Conference on Teaching, Assessment, and Learning for Engineering (TALE2020), 2020年12月, ［査読有り］
  • Sequential Authentication using Handwriting Biometrics for Free Response e-Testing
    Taisuke Kawamata; Takako Akakura, 筆頭著者, To prevent proxy-test taking among examinees in unsynchronized e-Testing, a previous work proposed an online handwriting authentication. That method was limited to applied for end of each answer. For free response tests that needed to authenticate throughout the answer, we used the Bayesian prior information to examine a sequential handwriting authentication procedure. The evaluation results indicate that the accuracy of this procedure is higher than the previous method in examinees authentication during mathematics exam with referring the Chinese character., 電子情報通信学会
    IEICE TRANSACTIONS on Information and Systems, 2020年04月, ［査読有り］
  • Hybrid handwriting authentication model for online testing using a Siamese neural network and MLP
    Daisuke Hayashi; Taisuke Kawamata; Toru Kano; Takako Akakura, We propose a hybrid authentication model for online testing that combines a handwriting model using a Siamese neural network and a dynamic information model using a multilayer perceptron. The resulting hybrid authentication model obtained better accuracy than either the handwriting model alone or the dynamic information model alone., Institute of Electrical and Electronics Engineers
    Proceedings of 2020 IEEE 2nd Global Conference on Life Sciences and Technologies (LifeTech), 2020年03月12日, ［査読有り］
  • 特許法の問題解決過程モデルに基づく論理式・論理回路組立型知識獲得支援システム
    赤倉 貴子; 川又 泰介; 加藤 浩一郎, In this study, we developed a system based on a problem-solving process model of the Japanese Patent Act to support knowledge acquisition among engineering students, who are better at interpreting the use of logical formulas and logic circuits than they are at reading legal texts. The knowledge acquisition support system helped students construct legal articles by themselves by using logical formulas and logic circuits. We found that this system was rated favorably by some students and that it was useful for their knowledge acquisition. Meanwhile, it is necessary to improve the system for those students who did not rate it favorably and to devise a support system for term exams. We speculate that such improvements will increase usage of the system, which in turn will lead to further support for knowledge acquisition., 工学教育
    工学教育, 2019年07月, ［査読有り］
  • RELATIONSHIP BETWEEN THE MULTIDIMENSIONALITY OF THE KNOWLEDGE ACQUISITION PROCESS FOR LAW AND THE USE OF LEARNING SUPPORT SYSTEMS THAT SIMULATE LEGAL TEXTS AS LOGIC CIRCUITS
    Takako Akakura; Taisuke Kawamata; Koichiro Kato, In a previous study, we developed systems with several functions, such as simulation of legal texts as logic circuits, and we revealed that engineering students studying intellectual property law have various preferences for the learning support systems. The study also showed that there were differences in individuals’ ability to train on each system.

    In this study, we classified students on several dimensions with respect to the knowledge acquisition process for intellectual property law and found that there are adaptive systems for each student. We consider the knowledge acquisition process to have four dimensions, as listed below:
    1. Learning style:
    (a) Acquiring knowledge by attempting to solve exercises.
    (b) Attempting to solve exercises only after thoroughly understanding the theory.
    2. Ability to read legal texts: (a) Good. (b) Poor.
    3. Subjective evaluation of reading legal texts: (a) Easy. (b) Difficult.
    4. Subjective evaluation of mathematical learning: (a) Easy. (b) Difficult.

    The students classified in these four dimensions constitute 16 types (24), and we report the results of our analysis based on the criteria listed below:
    1. The type of student and what kind of system they used.
    2. The type of student and how they simulated legal texts on the learning support system.
    3. The type of student and how they used the learning support system and acquired knowledge., International Academy of Technology, Education and Development
    Proceedings of 13th International Technology, Education and Development Conference（INTED2019), 2019年03月, ［査読有り］
  • RELEVANCE ANALYSIS OF CONTENTS TYPE AND FACE AUTHENTICATION FOR PREVENTING PLAGIARISM OF E-CONTENT
    Taisuke Kawamata; Takako Akakura, 筆頭著者, Many universities have introduced e-Learning systems to allow learners to take lecture courses remotely, which has led to thousands of lectures now being available online.On an e-Learning system, spoofing may occur if the teacher is unable to see the learners during lectures, which in turn creates certain problems. The lectures and associated materials are intellectual property that must be protected by preventing spoofing. Thus, it is necessary to authenticate learner identity during an e-Learning session.

    For preventing plagiarism of e-content and protecting intellectual property rights, we discuss a methodology of biometrics authentication using facial images. We implemented facial authentication to three types of e-Learning: text reading, video watching and exam type. The study shows the difference of learning behavior in each type such as text position on a display, occluding face during the exam and more. We analyzed the facial images on each behavior, and examined the method of control these behaviors to introduce the facial authentication., International Academy of Technology, Education and Development
    Proceedings of 13th International Technology, Education and Development Conference（INTED2019), 2019年03月, ［査読有り］
  • 知的財産法講義のための学習支援システムの利用状況と知識獲得状態の関係分析
    赤倉 貴子; 川又 泰介; 東本 崇仁; 加藤 浩一郎, We analyze the relationship between use of a learning support system for face-to-face lectures on intellectual property law and knowledge acquisition. We first describe a system that we are currently developing and operating. We next analyze the relationship between system use and knowledge acquisition over the past 5 years (FY 2013–2017) for engineering students participating in these lectures. The analysis results show that daytime students frequently used the system for mid-term and final exams, and that nighttime students tended to access the system throughout the term. Finally, we analyze relationships between login count, video system usage time, and exercise system usage times in FY 2017. The results suggest that the video system is useful for acquiring basic knowledge, and that the exercise system promotes acquisition of applied knowledge. The exercise system for studying legal texts, logical expressions, and logic circuits was well received by engineering students, suggesting that system usage facilitates understanding of legal texts. In future work, we will improve the exercise system for logical expressions and logic circuits., 工学教育
    工学教育, 2019年03月, ［査読有り］
  • e-TestingにおけるWebカメラとペンタブレットを用いた逐次受験者認証システムの開発
    川又 泰介; 赤倉 貴子, 筆頭著者, Web上で試験を行うe-Testingは時空間の制約なく受験が可能という利点がある一方で，試験中の不正行為が容易に発生し得るという問題がある．そこで，試験中のなりすまし受験を防止するために，バイオメトリクスを用いた受験者認証法が提案されているが，認証の要求による受験の阻害が課題として挙げられている．そこで本研究では，受験を阻害することなく継続的に受験者を認証するため，Webカメラとペンタブレットを用いて顔認証と筆記認証を組み合わせた受験者認証システムを開発した．システムを実環境に近い状況で使用した結果，顔・筆記共に実時間で認証処理を完了可能であり，被験者は認証処理を意識することなく受験することが可能であることが分かった．, 電子情報通信学会
    電子情報通信学会論文誌D, 2019年03月, ［査読有り］
  • さまざまな学習支援システムを利用した工学部における知的財産法教育の実践
    赤倉 貴子; 川又 泰介; 加藤浩一郎, Engineering students taking face-to-face classes on intellectual property law have indicated a desire to study a greater amount of applicable content. Accordingly, we have developed and are currently operating various learning support systems. The first system has video recordings of lectures as well as additional functions such as practice exercises. The second system is a self-learning support system that utilizes the fact that law statements can be represented by logical expressions. In the third system, the user interface of the second system is represented by logic elements and logic circuits. In this report, we explain these three systems, and we compare and analyze the use of each system by students and their evaluation of the systems. It was found that, overall, students preferred the first system. In addition, first system was evaluated highly by students who placed emphasis on acquiring knowledge, while the second and third systems were highly evaluated by those who prefer to do problem exercises., 工学教育
    工学教育, 2019年01月, ［査読有り］
  • Development and Use of a Video On Demand e-Learning System with Logic Circuit Exercises for Teaching Intellectual Property Law
    Takako Akakura; Taisuke Kawamata; Koichiro Kato, We are developing web-based e-learning systems for learning intellectual property law. One system records face-to-face lectures as its main function, with exercise and other functions added. In this study, we developed a system that can create exercises using logic circuits. Analysis of student usage results from FY 2013 to FY 2017 showed that the video system was widely used and that a system that teaches legal texts using logic circuits introduced in FY 2017 was positively evaluated., Institute of Electrical and Electronics Engineers
    Proceedings of 2018 IEEE International Conference on Teaching, Assessment, and Learning for Engineering (TALE), 2018年12月, ［査読有り］
  • Development of a Learning Support System that Automatically Generates Solutions by Converting Legal Articles into Logical Expressions
    Takako Akakura; Taisuke Kawamata; Koichiro Kato, We previously showed that legal articles can be represented as logical formulas, and that in legal education, computers can automatically compare solutions with answers input by students. Whereas the previously developed system advanced learning through exercises, in the present research we aimed to develop a system with functions for learning and exercises. We first analyzed how students use the exercise function, with the goal of making a system that is adaptive to students. As a result, we identified two student types who use the system differently: those who learn first and do exercises later, and those who learn through repeated exercises. When pages for learning were created, the new system was given high subjective ratings by the former student type., Institute of Electrical and Electronics Engineers
    Proceedings of 2018 IEEE 7th Global Conference on Consumer Electronics (GCCE), 2018年10月, ［査読有り］
  • e-Learningにおける入力顔情報を用いた参照情報の逐次更新による受講者認証
    川又 泰介; 石井 隆稔; 赤倉 貴子, 筆頭著者, e-Learningにおけるなりすまし防止手法として，顔認証における参照情報をe-Learning中に逐次更新する手法を開発した．評価実験より，入力情報と登録情報との類似度と，e-Learning中に発生するイベントによって更新の影響を調節することで，従来の方法よりも認証精度が向上する可能性が示唆された．, 電子情報通信学会
    電子情報通信学会論文誌D, 2018年04月, ［査読有り］
  • Development of A Learning Support System for Studying Intellectual Property Law Based on Representing Legal Statements as Logical Circuits
    Takako Akakura; Taisuke Kawamata; Koichiro Kato, We have previously developed a learning support system for intellectual property law that utilizes the fact that legal statements can be represented using propositional logic and first-order predicate logic. The developed system provides learners with feedback about their mistakes by finding the difference between the logical structure of the correct answer and the logical structure that the learner has entered. The system has been highly evaluated for facilitating learning by engineering students, who tend to be good at logical thinking. A logical structure is used for the internal control of the system and learners are aware of this as they enter their answers. However, the user interface was difficult to use as it involves using a dropdown list to enter logical operators such as ∧, ∨, ∀, or ∃, and we received many suggestions that this area be improved. Accordingly, in this study we have developed a learning support system where students create structures for legal statements by themselves, using circuit symbols such as AND, OR, NAND, NOR or NOT rather than symbols such as ∧, ∨, ∀, or ∃. For example, in order for something to be patentable under the Patent Act, it must satisfy conditions such as being an invention under the Patent Act, being novel, having an inventive step, and not violating the public interest. In determining whether an invention is novel, it must not violate the patentability conditions in clauses (i) to (iii) of Article 29(1) of the Patent Act. However, even if an invention fails (iii) as a result of being described in a distributed publication for example, there is still a high probability that it is patentable if it falls under the exceptions in Article 30(1) or (2). This kind of decision can be written as a logical circuit by using a 3-input NOR gate to join the clauses of Article 29(1) followed by a 3-input OR gate connecting the NOR output, Article 30(1), and Article 30(2), so that the final value becomes “1” (patentable) if either Article 30(1) or Article 30(2) is met, even if the output of the NOR gate is “0” (not patentable). This allows the system to have the learner construct the circuit and to simulate what the output will be for the logical circuit that the student has actually created, thereby giving feedback as to whether the answer was correct and, if not, which part of the circuit contains the error. We found that if students had already taken a course in logical circuits such a system was extremely easy for them to use, and that the user interface was also very highly evaluated., International Academy of Technology, Education and Development
    Proceedings of 12th International Technology, Education and Development Conference（INTED), 2018年03月, ［査読有り］
  • Development of a blended learning system for engineering students studying intellectual property law, and an analysis of the relationship between system usage and the knowledge acquisition process
    Takako Akakura; Taisuke Kawamata; Koichiro Kato, Institute of Electrical and Electronics Engineers
    Proceedings of 2017 IEEE International Conference on Teaching, Assessment and Learning for Engineering, TALE 2017, 2018年01月08日, ［査読有り］
  • Examination of sub-stroke and face authentication for e-Testing
    Taisuke Kawamata; Takako Akakura, 筆頭著者, Institute of Electrical and Electronics Engineers
    Proceedings of 2017 IEEE Region 10 Conference (TENCON), 2017年12月19日, ［査読有り］
  • Face Authentication for e-Learning Using Time Series Information
    Taisuke Kawamata; Takatoshi Ishii; Takako Akakura, 筆頭著者
    Proceedings of 2016 IEEE International Conference on Teaching, Assessment, and Learning for Engineering (TALE), 2016年12月, ［査読有り］
  • Student Authentication by Updated Facial Information with Weighting Coefficient in e-Learning
    Taisuke Kawamata; Takatoshi Ishii; Susumu Fujimori; Takako Akakura, 筆頭著者
    Proceedings of 2016 IEEE Region 10 Conference (TENCON), 2016年11月, ［査読有り］
  • Student Authentication Method by Sequential Update of Face Information Registered in e-Learning System
    Taisuke Kawamata; Susumu Fujimori; Takako Akakura, 筆頭著者
    Human Interface and the Management of Information: Applications and Services, Pt II, 2016年07月, ［査読有り］

MISC

講演・口頭発表等

担当経験のある科目（授業）

所属学協会

  • 2021年10月 - 現在, 情報処理学会
  • 2016年06月 - 現在, 日本教育工学会
  • 2015年01月 - 現在, 電子情報通信学会
  • 2018年04月 - 2025年04月, 知的資産経営学会
  • 2024年06月 - 2025年03月, 人工知能学会
  • 2018年11月 - 2024年09月, 工学教育協会

Works（作品等）

共同研究・競争的資金等の研究課題

メディア報道

学術貢献活動