2024/08/05 更新

写真a

オオキタ ツヨシ
大北 剛
OKITA Tsuyoshi
Scopus 論文情報  
総論文数: 0  総Citation: 0  h-index: 10

Citation Countは当該年に発表した論文の被引用数

所属
大学院情報工学研究院 知能情報工学研究系
職名
准教授
メールアドレス
メールアドレス
研究室住所
福岡県飯塚市川津680-4
研究室電話
0948-29-7674
研究室FAX
0948-29-7674
外部リンク

研究キーワード

  • 人工知能

  • 生成AI

  • 大規模言語モデル

  • 深層/機械学習

研究分野

  • 情報通信 / ソフトコンピューティング

出身大学院

  • 2012年03月   ダブリン市大学(Dublin City University)   コンピュータ学科   博士課程・博士後期課程   修了   アイルランド

取得学位

  • ダブリン市大学(Dublin City University)  -  計算機科学(Computer Science)   2012年03月

学内職務経歴

  • 2022年04月 - 現在   九州工業大学   大学院情報工学研究院   知能情報工学研究系     准教授

  • 2019年04月 - 2022年03月   九州工業大学   大学院情報工学研究院   知能情報工学研究系     特任准教授

  • 2018年04月 - 2019年03月   九州工業大学   大学院工学研究院   基礎科学研究系     特任講師

  • 2017年01月 - 2018年03月   九州工業大学   大学院工学研究院   基礎科学研究系     研究職員

論文

  • 11th International Workshop on Human Activity Sensing Corpus and Applications (HASCA) 査読有り 国際誌

    Murao K., Enokibori Y., Gjoreski H., Lago P., Okita T., Siirtola P., Hiroi K., Scholl P.M., Ciliberto M., Urano K.

    UbiComp/ISWC 2023 Adjunct - Adjunct Proceedings of the 2023 ACM International Joint Conference on Pervasive and Ubiquitous Computing and the 2023 ACM International Symposium on Wearable Computing   773 - 776   2023年10月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    The recognition of complex and subtle human behaviors from wearable sensors will enable next-generation human-oriented computing in scenarios of high societal value (e.g., dementia care). This will require a large-scale human activity corpus and much-improved methods to recognize activities and the context in which they occur. This workshop deals with the challenges of designing reproducible experimental setups, running large-scale dataset collection campaigns, designing activity and context recognition methods that are robust and adaptive, and evaluating systems in the real world. We wish to reflect on future methods, such as lifelong learning approaches that allow open-ended activity recognition. This year HASCA will welcome papers from participants to the Fifth Sussex-Huawei Locomotion and Transportation Recognition Challenge in a special session.

    DOI: 10.1145/3594739.3605106

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85175476451&origin=inward

  • Towards LLMs for Sensor Data: Multi-Task Self-Supervised Learning 査読有り 国際誌

    Okita T., Ukita K., Matsuishi K., Kagiyama M., Hirata K., Miyazaki A.

    UbiComp/ISWC 2023 Adjunct - Adjunct Proceedings of the 2023 ACM International Joint Conference on Pervasive and Ubiquitous Computing and the 2023 ACM International Symposium on Wearable Computing   499 - 504   2023年10月

     詳細を見る

    担当区分:筆頭著者, 責任著者   記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    LLMs for vision and NLP domain has been popular by the widespread use of ChatGPT and GPT-4. This paper tackles to build LLMs for sensor domain of one-dimensional signals whose downstream task is activity recognition and emotion detection. We propose a new architecture of Transformer-based self-supervised learner which we name SENvT. This SENvT builds the LLMs for sensor data using 7 pretext objectives in multi-task learning together with contrastive learning. Experimental results show these three. First, we obtained better results for contrastive learning and the masked token task but not for other pretext tasks. Second, the masked token task was better in 60% rather than in 10%. Third, the RGW worked best in accuracy while the masked token task worked best in F1.

    DOI: 10.1145/3594739.3610745

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85175472929&origin=inward

  • Summary of SHL Challenge 2023: Recognizing Locomotion and Transportation Mode from GPS and Motion Sensors 査読有り 国際誌

    Wang L., Gjoreski H., Ciliberto M., Lago P., Murao K., Okita T., Roggen D.

    UbiComp/ISWC 2023 Adjunct - Adjunct Proceedings of the 2023 ACM International Joint Conference on Pervasive and Ubiquitous Computing and the 2023 ACM International Symposium on Wearable Computing   575 - 585   2023年10月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    In this paper we summarize the contributions of participants to the fifth Sussex-Huawei Locomotion-Transportation (SHL) Recognition Challenge organized at the HASCA Workshop of UbiComp/ISWC 2023. The goal of this machine learning/data science challenge is to recognize eight locomotion and transportation activities (Still, Walk, Run, Bike, Bus, Car, Train, Subway) from the motion (accelerometer, gyroscope, magnetometer) and GPS (GPS location, GPS reception) sensor data of a smartphone in a user-independent manner. The training data of a "train"user is available from smartphones placed at four body positions (Hand, Torso, Bag and Hips). The testing data originates from "test"users with a smartphone placed at one, but unknown, body position. We introduce the dataset used in the challenge and the protocol of the competition. We present a meta-analysis of the contributions from 15 submissions, their approaches, the software tools used, computational cost and the achieved results. The challenge evaluates the recognition performance by comparing predicted to ground-truth labels at every 10 milliseconds, but puts no constraints on the maximum decision window length. Overall, five submissions achieved F1 scores above 90%, three between 80% and 90%, two between 70% and 80%, three between 50% and 70%, and two below 50%. While the task this year is facing the technical challenges of sensor unavailability, irregular sampling, and sensor diversity, the overall performance based on GPS and motion sensors is better than previous years (e.g. the best performance reported in SHL 2020, 2021 and 2023 are 88.5%, 75.4% and 96.0%, respectively). This is possibly due to the complementary between the GPS and motion sensors and also the removal of constraints on the decision window length. Finally, we present a baseline implementation to help understand the contribution of each sensor modality to the recognition task.

    DOI: 10.1145/3594739.3610758

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85175460992&origin=inward

  • Characteristics of Li-Ion Battery at Accelerated C-Rate with Deep Learning Method 査読有り 国際誌

    Hoque M.A., Hassan M.K., Hajjo A., Okita T.

    Arabian Journal for Science and Engineering   2023年03月

     詳細を見る

    担当区分:最終著者   記述言語:英語   掲載種別:研究論文(学術雑誌)

    In this research, Lithium-ion (Li-ion) batteries were tested at four different charge rates (DCR): 0.2C, 0.5C, 1C, and 1.5C, and four different discharge rates (DDR): 0.5C, 0.9C, 1.3C, 1.6C. This paper proposes a capacity fade model for charging and discharging at accelerated current-rate (C-rate), to interpret the vulnerabilities of Li-ion batteries in energy storage system, because Lithium-ion (Li-ion) batteries are prone to ageing at the fluctuation of the loads in micro-grids. The characteristics of Li-ion batteries both at accelerated DCR and DDR are thoroughly investigated. It is discovered that charging and discharging Li-ion batteries outside of the standard C-rate accelerates their ageing. In addition, the degree of capacity fade is assessed at an accelerated C-rate to develop an ideal charge and discharge model for the micro-grids. Furthermore, the battery capacity fade model is then investigated with deep learning algorithm-based feed-forward neural network (FNN), and recurrent neural network with long-short term memory layer (RNN-LSTM). A comparison of the developed capacity fade models is performed, and it is discovered that the LSTM-RNN battery ageing model outperforms the conventional FNN network at accelerated C-rate. Nevertheless, the error metrics performance of both FNN and LSTM-RNN are less than 0.1%.

    DOI: 10.1007/s13369-023-08034-x

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85162231470&origin=inward

  • 2Dアニメーションのフレーム補間に関するサーベイ

    河津水紀, 大北 剛

    第47回IBISML研究会   2022年09月

     詳細を見る

    担当区分:最終著者   記述言語:日本語   掲載種別:研究論文(研究会,シンポジウム資料等)

  • 人物実写画を特定の漫画の画風に沿うように変換し画像を出力する技術のサーベイ

    中島崇晴,大北 剛

    第47回IBISML研究会   2022年09月

     詳細を見る

    担当区分:最終著者   記述言語:日本語   掲載種別:研究論文(研究会,シンポジウム資料等)

  • 10th International Workshop on Human Activity Sensing Corpus and Applications (HASCA) 査読有り

    Murao K., Enokibori Y., Gjoreski H., Lago P., Okita T., Siirtola P., Hiroi K., Scholl P.M., Ciliberto M., Urano K.

    UbiComp/ISWC 2022 Adjunct - Proceedings of the 2022 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2022 ACM International Symposium on Wearable Computers   321 - 323   2022年09月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    The recognition of complex and subtle human behaviors from wearable sensors will enable next-generation human-oriented computing in scenarios of high societal value (e.g., dementia care). This will require large-scale human activity corpus and much improved methods to recognize activities and the context in which they occur. This workshop deals with the challenges of designing reproducible experimental setups, running large-scale dataset collection campaigns, designing activity and context recognition methods that are robust and adaptive, and evaluating systems in the real world. We wish to reflect on future methods, such as lifelong learning approaches that allow open-ended activity recognition.

    DOI: 10.1145/3544793.3560377

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85158983039&origin=inward

  • 脳血腫マーカーの画像パッチのマルチラベル学習

    加藤 舜斗, 河津 水紀, 中島 崇晴, 有村 公一, 飯原 弘二, 大北 剛

    DICOMOシンポジウム   2022年07月

     詳細を見る

    担当区分:最終著者   記述言語:日本語   掲載種別:研究論文(研究会,シンポジウム資料等)

  • ハミルトニアンニューラルネットワークの人間行動認識への応用

    豊坂 祐樹, 大北 剛

    DICOMOシンポジウム   2022年07月

     詳細を見る

    担当区分:最終著者   記述言語:日本語   掲載種別:研究論文(研究会,シンポジウム資料等)

  • 動きからの構造: 2D画像からの3Dテクスチャの再構築

    大野 友暉, 大北 剛

    DICOMOシンポジウム   2022年07月

     詳細を見る

    担当区分:最終著者   記述言語:日本語   掲載種別:研究論文(研究会,シンポジウム資料等)

  • 脳血腫の急成長の予測

    大北剛,中山俊太朗,山本周平,森山幹太,平野北斗,有村公一,飯原弘二

    第12回AIM 合同研究会   2022年03月

     詳細を見る

    担当区分:筆頭著者   記述言語:日本語   掲載種別:研究論文(研究会,シンポジウム資料等)

  • A Compression-Based Multiple Subword Segmentation for Neural Machine Translation 査読有り 国際誌

    Keita Nonaka , Kazutaka Yamanouchi , Tomohiro I , Tsuyoshi Okita , Kazutaka Shimada, Hiroshi Sakamoto

    Electronics ( MDPI )   11 ( 7 )   2022年03月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(学術雑誌)

    In this study, we propose a simple and effective preprocessing method for subword segmentation based on a data compression algorithm. Compression-based subword segmentation has recently attracted significant attention as a preprocessing method for training data in neural machine translation. Among them, BPE/BPE-dropout is one of the fastest and most effective methods compared to conventional approaches; however, compression-based approaches have a drawback in that generating multiple segmentations is difficult due to the determinism. To overcome this difficulty, we focus on a stochastic string algorithm, called locally consistent parsing (LCP), that has been applied to achieve optimum compression. Employing the stochastic parsing mechanism of LCP, we propose LCP-dropout for multiple subword segmentation that improves BPE/BPE-dropout, and we show that it outperforms various baselines in learning from especially small training data.

    DOI: 10.3390/electronics11071014

    DOI: 10.3390/electronics11071014

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85126910361&origin=inward

  • 距離画像推定情報を用いた複数人の行動認識

    豊坂祐樹,大北剛

    第72回UBI研究会   2021年11月

     詳細を見る

    担当区分:最終著者   記述言語:日本語   掲載種別:研究論文(研究会,シンポジウム資料等)

  • 弱教師ありセマンティックセグメンテーション

    大北剛,平野北斗,森山幹太,有村公一,飯原弘二

    第23回日本知能情報ファジィ学会九州支部学術講演会(ソフト九州)   2021年11月

     詳細を見る

    担当区分:筆頭著者   記述言語:日本語   掲載種別:研究論文(研究会,シンポジウム資料等)

  • 脳血腫マーカーの分類: 認識、物体検知、敵対的生成法、セマンティックセグメンテーション

    大北 剛, 平野 北斗, 有村 公一, 飯原 弘二

    人工知能学会第二種研究会資料 ( 一般社団法人 人工知能学会 )   2021 ( AIMED-011 )   05   2021年11月

     詳細を見る

    担当区分:筆頭著者, 責任著者   記述言語:日本語   掲載種別:研究論文(学術雑誌)

    DOI: 10.11517/jsaisigtwo.2021.aimed-011_05

    CiNii Research

  • Locomotion and Transportation Mode Recognition from GPS and Radio Signals: Summary of SHL Challenge 2021 査読有り

    Wang L., Ciliberto M., Gjoreski H., Lago P., Murao K., Okita T., Roggen D.

    UbiComp/ISWC 2021 - Adjunct Proceedings of the 2021 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2021 ACM International Symposium on Wearable Computers   412 - 422   2021年09月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    In this paper we summarize the contributions of participants to the fourth Sussex-Huawei Locomotion-Transportation (SHL) Recognition Challenge organized at the HASCA Workshop of UbiComp/ISWC 2021. The goal of this machine learning/data science challenge is to recognize eight locomotion and transportation activities (Still, Walk, Run, Bike, Bus, Car, Train, Subway) from the radio sensor data (GPS location, GPS reception, WiFi reception and Cell reception) of a smartphone in a user-independent manner. The training and testing data are collected by different users with a smartphone placed at the Hips position. We introduce the dataset used in the challenge and the protocol of the competition. We present a meta-Analysis of the contributions from 15 submissions, their approaches, the software tools used, computational cost and the achieved results. The challenge evaluates the recognition performance by comparing predicted to ground-Truth labels at every second, but puts no constraints on the maximum decision window length. Overall, two submissions achieved F1 scores between 70% and 80%, one between 60% and 70%, five between 50% and 60%, and seven below 50%. Due to the technical challenges of data synchronization, sensor unavailability and sensor diversity, the overall performance based on GPS and radio sensors is lower than the performance achieved by motion sensors in previous challenges (SHL 2018-2020). Finally, we present a baseline implementation to help understand the contribution of each sensor modality to the recognition task.

    DOI: 10.1145/3460418.3479373

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85115916446&origin=inward

  • 9th International Workshop on Human Activity Sensing Corpus and Applications (HASCA) 査読有り

    Murao K., Enokibori Y., Gjoreski H., Lago P., Okita T., Siirtola P., Hiroi K., Scholl P.M., Ciliberto M., Urano K.

    UbiComp/ISWC 2021 - Adjunct Proceedings of the 2021 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2021 ACM International Symposium on Wearable Computers   281 - 284   2021年09月

     詳細を見る

    担当区分:責任著者   記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    The recognition of complex and subtle human behaviors from wearable sensors will enable next-generation human-oriented computing in scenarios of high societal value (e.g., dementia care). This will require large-scale human activity corpus and much improved methods to recognize activities and the context in which they occur. This workshop deals with the challenges of designing reproducible experimental setups, running large-scale dataset collection campaigns, designing activity and context recognition methods that are robust and adaptive, and evaluating systems in the real world. We wish to reflect on future methods, such as lifelong learning approaches that allow open-ended activity recognition. This year HASCA will welcome papers from participants to the Fourth Sussex-Huawei Locomotion and Transportation Recognition Challenge and the Third Nursing Activity Recognition Challenge in special sessions.

    DOI: 10.1145/3460418.3479266

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85115942696&origin=inward

  • Activity Knowledge Graph Recognition by Eye Gaze: Identification of Distant Object in Eye Sight for Watch Activity 査読有り

    Toyosaka Y., Okita T.

    UbiComp/ISWC 2021 - Adjunct Proceedings of the 2021 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2021 ACM International Symposium on Wearable Computers   334 - 339   2021年09月

     詳細を見る

    担当区分:責任著者   記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    From the side of upper-level applications which require planning the actions in robot or those which need to search the whole log of activities in smart home, the action predicate expressions in the form of knowledge graphs may play an important role. The sequence of activities alone, which can be supplied by the conventional activity recognition systems, may not be sufficient for those applications. The subject of the particular activity is crucial information in most of the cases, and the object of the particular activity is often necessary to identify the characteristics. From this perspective, we have investigated the activities recognized by activity recognition systems, trying to identify their hidden elements which play the role of the subject and the object of the activities, i.e. activity knowledge graph. If we focus on these hidden elements, they are categorized in two: (1) person (subject)-person (object) interaction, and (2) person (subject)-object (object) interactions. Depending on the class of activities, these two are sometimes faced great difficulties: The hidden elements for walk, pick-up, open, and drink are quite easy but those for look-At, see, watch, and throw are difficult. The source of difficulties arises from the fact that the object (object) is not contacted from the person (subject). In this paper we have developed a method which identifies non-contacted object by the direction of the eye gaze of the person (subject) in the category of watch (activity). Using "Watching TV"data by Stair lab, the proposed system achieved 85% in accuracy.

    DOI: 10.1145/3460418.3479351

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85115984916&origin=inward

  • Activity Simulation from Signals 査読有り

    Okita T.

    UbiComp/ISWC 2021 - Adjunct Proceedings of the 2021 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2021 ACM International Symposium on Wearable Computers   59 - 60   2021年09月

     詳細を見る

    担当区分:筆頭著者   記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    Sensor-based human activity recognition technology has been used for estimating human action based on the sensor data. In this paper, we propose a new paradigm to render the human activity on a screen instead of classifying the activity among the activity labels. We could built this mockup of a simulator, combining our previous translation tool between signals [2] with the motion rendering systems [3]. We faced two problems which decrease the simulation ability a lot. We proposed two algorithms to increase the performance of this simulator in this preliminary work.

    DOI: 10.1145/3460418.3479275

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85115964745&origin=inward

  • Three-year Review of the 2018-2020 SHL Challenge on Transportation and Locomotion Mode Recognition from Mobile Sensors 査読有り 国際誌

    Lin Wang, Hristijan Gjoreski, Mathias Ciliberto, Paula Lago, Kazuya Murao, Tsuyoshi Okita and Daniel Roggen

    Frontiers   3   2021年09月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(学術雑誌)

    The Sussex-Huawei Locomotion-Transportation (SHL) Recognition Challenges aim to advance and capture the state-of-the-art in locomotion and transportation mode recognition from smartphone motion (inertial) sensors. The goal of this series of machine learning and data science challenges was to recognize eight locomotion and transportation activities (Still, Walk, Run, Bus, Car, Train, Subway). The three challenges focused on time-independent (SHL 2018), position-independent (SHL 2019) and user-independent (SHL 2020) evaluations, respectively. Overall, we received 48 submissions (out of 93 teams who registered interest) involving 201 scientists over the three years. The survey captures the state-of-the-art through a meta-analysis of the contributions to the three challenges, including approaches, recognition performance, computational requirements, software tools and frameworks used. It was shown that state-of-the-art methods can distinguish with relative ease most modes of transportation, although the differentiating between subtly distinct activities, such as rail transport (Train and Subway) and road transport (Bus and Car) still remains challenging. We summarize insightful methods from participants that could be employed to address practical challenges of transportation mode recognition, for instance, to tackle over-fitting, to employ robust representations, to exploit data augmentation, and to exploit smart post-processing techniques to improve performance. Finally, we present baseline results to compare the three challenges with a unified recognition pipeline and decision window length.

    DOI: 10.3389/fcomp.2021.713719

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85115915802&origin=inward

  • Summary of the Third Nurse Care Activity Recognition Challenge-Can We Do from the Field Data? 査読有り

    Alia S.S., Adachi K., Hossain T., Le N.T., Kaneko H., Lago P., Okita T., Inoue S.

    UbiComp/ISWC 2021 - Adjunct Proceedings of the 2021 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2021 ACM International Symposium on Wearable Computers   428 - 433   2021年09月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    DOI: 10.1145/3460418.3479391

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85115908587&origin=inward

  • 非接触性の人と物体における相互作用検出方法の提案

    豊坂祐樹, 大北剛

    DICOMOシンポジウム   2021年06月

     詳細を見る

    担当区分:最終著者   記述言語:日本語   掲載種別:研究論文(研究会,シンポジウム資料等)

  • 脳血腫の分類: セグメンテーションと分類のジョイント学習

    平野北斗, 大北剛

    arxiv   2021年03月

     詳細を見る

    担当区分:責任著者   記述言語:日本語   掲載種別:研究論文(学術雑誌)

    arXiv

  • Improvement of Human Action Recognition Using 3D Pose Estimation 査読有り

    Adachi K., Lago P., Okita T., Inoue S.

    Smart Innovation, Systems and Technologies   204   21 - 37   2021年01月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    While Human Action Recognition (HAR) using motion capture can perform well with high accuracy, it requires a high computational cost for recording and post-processing. To avoid this, we build a HAR system using 3D pose estimation from the single-camera video instead of motion capture. One drawback in this approach is that the performance is considerably dependent on the camera position. This paper investigates how we can use the pose estimate constantly without the effect of camera position even when the camera position in the test data is changed. We augment the data by rotating around the 3D pose estimate to improve the accuracy when using different camera positions in the test data and in the training data. The strategy of augmenting training data shows improvements up to 55.7% in accuracy, compared with the case of 2D pose with no augmentation.

    DOI: 10.1007/978-981-15-8944-7_2

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85107353674&origin=inward

  • 会議報告:NeurIPS 2020(The 34th Conference on Neural Information Processing Systems) 査読有り

    高瀬 翔, 坂本 陸, 大北 剛

    人工知能 ( 一般社団法人 人工知能学会 )   36 ( 4 )   527 - 531   2021年01月

     詳細を見る

    記述言語:英語   掲載種別:記事・総説・解説・論説等(学術雑誌)

    DOI: 10.11517/jjsai.36.4_527

    CiNii Article

    その他リンク: https://ci.nii.ac.jp/naid/130008060910

  • Summary of the sussex-huawei locomotion-transportation recognition challenge 2020 査読有り

    Wang L., Gjoreski H., Ciliberto M., Lago P., Murao K., Okita T., Roggen D.

    UbiComp/ISWC 2020 Adjunct - Proceedings of the 2020 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2020 ACM International Symposium on Wearable Computers   351 - 358   2020年09月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    In this paper we summarize the contributions of participants to the third Sussex-Huawei Locomotion-Transportation (SHL) Recognition Challenge organized at the HASCA Workshop of UbiComp/ISWC 2020. The goal of this machine learning/data science challenge is to recognize eight locomotion and transportation activities (Still, Walk, Run, Bike, Bus, Car, Train, Subway) from the inertial sensor data of a smartphone in a user-independent manner with an unknown target phone position. The training data of a "train"user is available from smartphones placed at four body positions (Hand, Torso, Bag and Hips). The testing data originates from "test"users with a smartphone placed at one, but unknown, body position. We introduce the dataset used in the challenge and the protocol of the competition. We present a meta-analysis of the contributions from 15 submissions, their approaches, the software tools used, computational cost and the achieved results. Overall, one submission achieved F1 scores above 80%, three with F1 scores between 70% and 80%, seven between 50% and 70%, and four below 50%, with a latency of maximum of 5 seconds.

    DOI: 10.1145/3410530.3414341

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85091885062&origin=inward

  • 8th international workshop on human activity sensing corpus and applications (HASCA) 査読有り

    Murao K., Enokibori Y., Gjoreski H., Lago P., Okita T., Siirtola P., Hiroi K., Scholl P.M., Ciliberto M.

    UbiComp/ISWC 2020 Adjunct - Proceedings of the 2020 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2020 ACM International Symposium on Wearable Computers   228 - 231   2020年09月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    The recognition of complex and subtle human behaviors from wearable sensors will enable next-generation human-oriented computing in scenarios of high societal value (e.g., dementia care). This will require large-scale human activity corpus and much improved methods to recognize activities and the context in which they occur. This workshop deals with the challenges of designing reproducible experimental setups, running large-scale dataset collection campaigns, designing activity and context recognition methods that are robust and adaptive, and evaluating systems in the real world. We wish to reflect on future methods, such as lifelong learning approaches that allow open-ended activity recognition. This year HASCA will welcome papers from participants to the Third Sussex-Huawei Locomotion and Transportation Recognition Challenge and Second Nursing Activity Recognition Challenge in special sessions.

    DOI: 10.1145/3410530.3414612

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85091833400&origin=inward

  • Perception of interaction between hand and object 査読有り

    Toyosaka Y., Okita T.

    UbiComp/ISWC 2020 Adjunct - Proceedings of the 2020 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2020 ACM International Symposium on Wearable Computers   290 - 295   2020年09月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    Action knowledge graphs can play a central role in smart cities, smart homes, robot planning, and so on. This is since both of the subject and the object of the actions can add more meaningful information for the higher-level application than the action alone as a predicate. We built a system that generates the action knowledge graphs from video using deep learning. Especially, we propose an algorithm which perceives the interaction between hand and object by measuring the proximity between them with considering the direction of fingers. We showed that this approach achieves the performance of 83% in accuracy using the Stair Lab data.

    DOI: 10.1145/3410530.3414363

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85091839586&origin=inward

  • Summary of the 2nd nurse care activity recognition challenge using lab and field data 査読有り

    Alia S.S., Lago P., Adachi K., Hossain T., Goto H., Okita T., Inoue S.

    UbiComp/ISWC 2020 Adjunct - Proceedings of the 2020 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2020 ACM International Symposium on Wearable Computers   378 - 383   2020年09月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    2nd Nurse Care Activity Recognition Challenge using Lab and Field Data is organized as a part of HASCA workshop and is continuation of Nurse Care Activity Recognition Challenge [7]. We give the description of the dataset and summarize the approaches used by the teams in this Challenge. In this challenge, data collected in both lab and real-world setting is provided to the challenge participants with an aim to bridge the gap between lab and practical field to reduce the workload of the nurses. The challenge was started on May 1, 2020 and continued until July 9, 2020. Accuracy is used as performance metric to evaluate the submissions. The winning team used k-NN classifier and achieved about 22.35% accuracy.

    DOI: 10.1145/3410530.3414611

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85091886199&origin=inward

  • Translation Between Waves, wave2wave 国際誌

    Tsuyoshi Okita, Hirotaka Hachiya, Sozo Inoue, Naonori Ueda

    arxiv   2020年07月

     詳細を見る

    担当区分:筆頭著者   記述言語:英語   掲載種別:研究論文(学術雑誌)

  • 手と物体の相互作用の認識とその知識グラフの生成

    豊坂祐樹, 大北剛

    DICOMOシンポジウム   361 - 367   2020年06月

     詳細を見る

    記述言語:日本語   掲載種別:研究論文(研究会,シンポジウム資料等)

    Virtual   2020年06月30日  -  2020年07月02日

  • U-netを用いた異常検知による肺炎の検知

    長村 徹, 徳永 旭将, 大北 剛

    DICOMOシンポジウムプロシーディングス   526 - 533   2020年06月

     詳細を見る

    記述言語:日本語   掲載種別:研究論文(研究会,シンポジウム資料等)

  • SHAP値や重要度を用いたモデル解釈性: 包除積分ネットワークとXGBoostの比較

    板橋将之, 本田あおい, 大北剛

    火の国シンポジウムプロシーディングス   2020年03月

     詳細を見る

    担当区分:責任著者   記述言語:日本語   掲載種別:研究論文(研究会,シンポジウム資料等)

  • アマゾン商品レビューの感情分析と評価スコアの相関性の検証

    高山なつき, Mario Koppen, 大北剛

    火の国シンポジウムプロシーディングス   2020年03月

     詳細を見る

    担当区分:責任著者   記述言語:日本語   掲載種別:研究論文(研究会,シンポジウム資料等)

  • 会議報告:The 33rd Conference on Neural Information Processing Systems(NeurIPS 2019) 査読有り

    大北 剛

    人工知能 ( 一般社団法人 人工知能学会 )   35 ( 3 )   465 - 466   2020年01月

     詳細を見る

    記述言語:英語   掲載種別:記事・総説・解説・論説等(学術雑誌)

    DOI: 10.11517/jjsai.35.3_465

    CiNii Article

    その他リンク: https://ci.nii.ac.jp/naid/130007917827

  • ビデオからの3次元姿勢推定と機械学習を用いた行動認識の試み

    安達康平,大北剛,井上創造

    ソフト九州プロシーディングス   2019年11月

     詳細を見る

    記述言語:日本語   掲載種別:研究論文(研究会,シンポジウム資料等)

  • 不変リスク最小化の考察

    大北剛

    ソフト九州プロシーディングス   2019年11月

     詳細を見る

    担当区分:筆頭著者   記述言語:日本語   掲載種別:研究論文(研究会,シンポジウム資料等)

  • 混合正規分布のデータセットにおけるAntlion clusteringと既存のクラスタリング手法との比較

    豊坂祐樹,福田亮治,大北剛,宮野英次

    ソフト九州プロシーディングス   2019年11月

     詳細を見る

    記述言語:日本語   掲載種別:研究論文(研究会,シンポジウム資料等)

  • Summary of the Sussex-Huawei locomotion-transportation recognition challenge 2019 査読有り

    Wang L., Gjoreski H., Ciliberto M., Lago P., Murao K., Okita T., Roggen D.

    UbiComp/ISWC 2019- - Adjunct Proceedings of the 2019 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2019 ACM International Symposium on Wearable Computers   849 - 856   2019年09月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    In this paper we summarize the contributions of participants to the Sussex-Huawei Transportation-Locomotion (SHL) Recognition Challenge organized at the HASCA Workshop of UbiComp 2019. The goal of this machine learning/data science challenge is to recognize eight locomotion and transportation activities (Still, Walk, Run, Bike, Bus, Car, Train, Subway) from the inertial sensor data of a smartphone in a placement independent manner. The training data is collected with smartphones placed at three body positions (Torso, Bag and Hips), while the testing data is collected with a smartphone placed at another body position (Hand). We introduce the dataset used in the challenge and the protocol for the competition. We present a meta-analysis of the contributions from 14 submissions, their approaches, the software tools used, computational cost and the achieved results. Overall, three submissions achieved F1 scores between 70% and 80%, five with F1 scores between 60% and 70%, five between between 50% and 60%, and one below 50%, with a latency of a maximum of 5 seconds.

    DOI: 10.1145/3341162.3344872

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85072892002&origin=inward

  • 7th international workshop on human Activity Sensing Corpus and Applications (HASCA) 査読有り

    Murao K., Enokibori Y., Gjoreski H., Lago P., Okita T., Siirtola P., Hiroi K., Scholl P.M., Ciliberto M.

    UbiComp/ISWC 2019- - Adjunct Proceedings of the 2019 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2019 ACM International Symposium on Wearable Computers   671 - 673   2019年09月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    The recognition of complex and subtle human behaviors from wearable sensors will enable next-generation human-oriented computing in scenarios of high societal value (e.g., dementia care). This will require large-scale human activity corpuses and much improved methods to recognize activities and the context in which they occur. This workshop deals with the challenges of designing reproducible experimental setups, running large-scale dataset collection campaigns, designing activity and context recognition methods that are robust and adaptive, and evaluating systems in the real world. We wish to reflect on future methods, such as lifelong learning approaches that allow open-ended activity recognition. This year HASCA will welcome papers from participants to the Second Sussex-Huawei Locomotion and Transportation Recognition Competition and Open Lab Nursing Activity Recognition Challenge in special sessions.

    DOI: 10.1145/3341162.3347765

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85072892160&origin=inward

  • Nurse care activity recognition challenge: Summary and results 査読有り

    Lago P., Alia S.S., Takeda S., Mairittha T., Mairittha N., Faiz F., Nishimura Y., Adachi K., Okita T., Charpillet F., Inoue S.

    UbiComp/ISWC 2019- - Adjunct Proceedings of the 2019 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2019 ACM International Symposium on Wearable Computers   746 - 751   2019年09月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    Although activity recognition has been studied for a long time now, research and applications have focused on physical activity recognition. Even if many application domains require the recognition of more complex activities, research on such activities has attracted less attention. One reason for this gap is the lack of datasets to evaluate and compare different methods. To promote research in such scenarios, we organized the Open Lab Nursing Activity Recognition Challenge focusing on the recognition of complex activities related to the nursing domain. Nursing domain is one of the domains that can benefit enormously from activity recognition but has not been researched due to lack of datasets. The competition used the CARECOM Nurse Care Activity Dataset, featuring 7 activities performed by 8 subjects in a controlled environment with accelerometer sensors, motion capture and indoor location sensor. In this paper, we summarize the results of the competition.

    DOI: 10.1145/3341162.3345577

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85072882944&origin=inward

  • Reduction of marker-body matching work in activity recognition using motion capture 査読有り

    Takeda S., Lago P., Okita T., Inoue S.

    UbiComp/ISWC 2019- - Adjunct Proceedings of the 2019 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2019 ACM International Symposium on Wearable Computers   835 - 842   2019年09月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    In this paper, activity recognition is performed using an optical motion capture system that can measure three-dimensional position information of reflective markers attached to the body. The individual markers detected by motion capture are automatically associated with which part of the body they are attached to. However, due to the overlapping of obstacles and other body parts and misplacement of the markers, these may be hidden from the camera and enter a blind spot, which may frequently cause a marker to be associated to different body parts erroneously. Usually, these errors need to be corrected manually after measurement, but this work is very time consuming, cumbersome and requires some skill. In this research, it is thought that there is no problem in recognizing the activity even if the process of spending the effort of correcting the correspondence between the marker after measurement and the body is omitted in the activity recognition using the motion capture. Because feature quantities are extracted from activity data when performing action recognition, even if an error occurs in part of the marker data, the effect is small because the correct feature quantities are selected and other marker data can compensate for an error. In addition, in this paper, we proposed a method to recognize the activity using the data when the human body template preparation required before Mocap data measurement is omitted, which is one of marker body matching work. The verification showed that even if the marker body matching operation was omitted, it was possible to recognize the action with high accuracy.

    DOI: 10.1145/3341162.3345591

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85072894653&origin=inward

  • MEASURed: Evaluating Sensor-Based Activity Recognition Scenarios 査読有り 国際誌

    Paula Lago, Shingo Takeda, Tsuyoshi Okita, Sozo Inoue

    Human Activity Sensing: Corpus and Applications ( Springer Nature )   2019年09月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(学術雑誌)

    Human Activity Recognition from accelerometer sensors is key to enable applications such as fitness tracking or health status monitoring at home. However, evaluating the performance of activity recognition systems in real-life deployments is challenging to the multiple differences in sensor number, placement and orientation that may arise in real settings. Considering such differences requires a large amount of labeled data. To overcome the challenges and costs associated to the collection of a wide range of heterogeneous data, we propose a simulator, called MEASURed, which uses motion capture to simulate accelerometer data on different settings. Then, using the simulated data to estimate the performance of activity recognition models under different scenarios. In this chapter, we describe MEASURed and evaluate its performance in estimating the accuracy of activity recognition models. Our results show that MEASURed can estimate the average accuracy of an activity recognition model using real accelerometer magnitude data. By using motion capture to simulate accelerometer data, the sensor research community can profit from visual datasets that have been collected by other communities to evaluate performance of activity recognition in a wide range of activities. MEASURed can be used to evaluate activity recognition classifiers in settings with different number, placement, and sampling rate of accelerometer sensors. The evaluation on a broad spectrum of scenarios gives a more general view of models and their limitations.

  • Evaluation of transfer learning for human activity recognition among different datasets 査読有り

    Islam M.S., Okita T., Inoue S.

    Proceedings - IEEE 17th International Conference on Dependable, Autonomic and Secure Computing, IEEE 17th International Conference on Pervasive Intelligence and Computing, IEEE 5th International Conference on Cloud and Big Data Computing, 4th Cyber Science and Technology Congress, DASC-PiCom-CBDCom-CyberSciTech 2019   854 - 859   2019年08月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    Human activity recognition is a potential area of research. For better performance, it requires significant amount of labelled data. Collecting labeled activity data is expensive and time-consuming. To solve this problem, transfer learning has been demonstrated very effective as it gathers knowledge from labeled train data of source domain and transfers that knowledge to target domain, which has little or no labeled data. In this paper, we propose unsupervised transfer learning from source dataset to target dataset, which are completely different in terms of number of users and samples. We have used Maximum Mean Discrepancy (MMD) based transfer learning model and compared with base Convolutional Neural Network (CNN) model. We have used 4 datasets for experiment. We have trained the model on a source dataset and then transferred the model to a target dataset, which has no labels to classify activities. We have found that transfer learning model has achieved better performance compared to the base model.

    DOI: 10.1109/DASC/PiCom/CBDCom/CyberSciTech.2019.00155

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85075116858&origin=inward

  • モーションキャプチャを用いた行動認識におけるマーカー身体対応付け作業の削減

    Shingo Takeda, Paula Lago, Tsuyoshi Okita, Sozo Inoue, Yoshinori Ideno

    DICOMOシンポジウムプロシーディングス   2019年07月

     詳細を見る

    記述言語:日本語   掲載種別:研究論文(研究会,シンポジウム資料等)

  • センサ行動認識における機械学習のための実験室高精度データと現場長時間データの比較

    Hiroki Goto, Shingo Takeda, Paula Lago, Tsuyoshi Okita, Sozo Inoue

    DICOMOシンポジウムプロシーディングス   2019年07月

     詳細を見る

    記述言語:日本語   掲載種別:研究論文(研究会,シンポジウム資料等)

  • 医療相談のための対話システムにおける要求分析

    二宮仁志, Tittaya Mairittha, 大北剛, 井上創造

    ソフト九州プロシーディングス   2018年12月

     詳細を見る

    記述言語:日本語   掲載種別:研究論文(研究会,シンポジウム資料等)

  • センサ行動認識のための機械学習を用いた加速度データシミュレーション

    武田 紳吾, Paula Lago, 大北剛, 井上創造

    ソフト九州プロシーディングス   2018年12月

     詳細を見る

    記述言語:日本語   掲載種別:研究論文(研究会,シンポジウム資料等)

  • 種々の敵対的生成ネットワークに対するGeometry Scoreによる評価

    福島康太, 大北剛, 井上創造

    ソフト九州プロシーディングス   2018年12月

     詳細を見る

    記述言語:日本語   掲載種別:研究論文(研究会,シンポジウム資料等)

  • Summary of the Sussex-Huawei locomotion-transportation recognition challenge 査読有り

    Wang L., Murao K., Gjoreski H., Okita T., Roggen D.

    UbiComp/ISWC 2018 - Adjunct Proceedings of the 2018 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2018 ACM International Symposium on Wearable Computers   1521 - 1530   2018年10月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    In this paper we summarize the contributions of participants to the Sussex-Huawei Transportation-Locomotion (SHL) Recognition Challenge organized at the HASCA Workshop of UbiComp 2018. The SHL challenge is a machine learning and data science competition, which aims to recognize eight transportation activities (Still, Walk, Run, Bike, Bus, Car, Train, Subway) from the inertial and pressure sensor data of a smartphone. We introduce the dataset used in the challenge and the protocol for the competition. We present a meta-analysis of the contributions from 19 submissions, their approaches, the software tools used, computational cost and the achieved results. Overall, two entries achieved F1 scores above 90%, eight with F1 scores between 80% and 90%, and nine between 50% and 80%.

    DOI: 10.1145/3267305.3267519

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85058337765&origin=inward

  • 6th international workshop on human Activity Sensing Corpus and Applications (HASCA) 査読有り

    Murao K., Enokibori Y., Gjoreski H., Lago P., Okita T., Siirtola P., Hiroi K., Scholl P.M.

    UbiComp/ISWC 2018 - Adjunct Proceedings of the 2018 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2018 ACM International Symposium on Wearable Computers   1392 - 1395   2018年10月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    The recognition of complex and subtle human behaviors from wearable sensors will enable next-generation human-oriented computing in scenarios of high societal value (e.g., dementia care). This will require large-scale human activity corpuses and much improved methods to recognize activities and the context in which they occur. This workshop deals with the challenges of designing reproducible experimental setups, running large-scale dataset collection campaigns, designing activity and context recognition methods that are robust and adaptive, and evaluating systems in the real world. We wish to reflect on future methods, such as lifelong learning approaches that allow open-ended activity recognition. Unique this year, HASCA will welcome papers from participants to the Sussex-Huawei Locomotion and Transportation Recognition Competition in a special session.

    DOI: 10.1145/3267305.3274145

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85058343540&origin=inward

  • A multi-sensor setting activity recognition simulation tool 査読有り

    Takeda S., Okita T., Lago P., Inoue S.

    UbiComp/ISWC 2018 - Adjunct Proceedings of the 2018 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2018 ACM International Symposium on Wearable Computers   1444 - 1448   2018年10月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    Motion capture generates data which are often more accurate than those captured by multiple of accelerometer sensors by their physical specification. Based on the observation that accelerometer data can be obtained by the second derivation of position data from motion capture, we propose a simulator, called MEASURed, for activity recognition classifiers. MEASURed can accommodate any number of virtual accelerometer sensors on the body based on some given motion capture data. Therefore, MEASURed can evaluate activity recognition classifiers in settings with different number, placement, and sampling rate of accelerometer sensors. Our results show that the F1-Score estimated by MEASURed is close to that obtained with the real accelerometer data.

    DOI: 10.1145/3267305.3267509

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85058289678&origin=inward

  • Activity recognition: Translation across sensor modalities using deep learning 査読有り

    Okita T., Inoue S.

    UbiComp/ISWC 2018 - Adjunct Proceedings of the 2018 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2018 ACM International Symposium on Wearable Computers   1462 - 1471   2018年10月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    We propose a method to translate between multi-modalities using an RNN encoder-decoder model. Based on such a model allowing to translate between modalities, we built an activity recognition system. The idea of equivalence of modality was investigated by Banos et al. This paper replaces this with deep learning. We compare the performance of translation with/without clustering and sliding window. We show the preliminary performance of activity recognition attained the F1 score of 0.78.

    DOI: 10.1145/3267305.3267512

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85058312644&origin=inward

  • Poster: Improving sensor-based activity recognition using motion capture as additional information 査読有り

    Lago P., Okita T., Takeda S., Inoue S.

    UbiComp/ISWC 2018 - Adjunct Proceedings of the 2018 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2018 ACM International Symposium on Wearable Computers   118 - 121   2018年10月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    We propose a new method for human activity recognition using a single accelerometer sensor and additional sensors for training. The performance of inertial sensors for complex activities drops considerably compared with simple activities due to inter-class similarities. In such cases deploying more sensors may improve the performance. But such strategy is often not feasible in reality due to costs or privacy concerns among others. In this context, we propose a new method to use additional sensors only in training phase. We introduce the idea of mapping the test data to a codebook created from the additional sensor information. Using the Berkeley MHAD dataset our preliminary results show this worked positively; improving in 10.0% both the average F1-score and the average accuracy. Notably, the improvement for the stand, sit and sit to stand activities was higher, typical activities for which the inertial sensor is less informative when using the wrist-worn accelerometer.

    DOI: 10.1145/3267305.3267596

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85058340955&origin=inward

  • Pre-consulting dialogue systems for telemedicine: Yes/no intent classification 査読有り

    Mairittha T., Okita T., Inoue S.

    UbiComp/ISWC 2018 - Adjunct Proceedings of the 2018 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2018 ACM International Symposium on Wearable Computers   742 - 745   2018年10月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    Telemedicine is an emerging challenge for the shortage of qualified professionals, particularly in under-resourced regions. Physical assessment by a non-medical doctor is a practice in telemedicine which discovers essential symptom of a patient who needs to consult a doctor. We aim at facilitating this stage with a conversational chatbot which identifies the patient by conversation. Adopting the procedures of physical assessment one critical types of conversation involves in the self-diagnosis. Further, it turned out that useful kinds of questions in chatbot at this stage are related to Yes/No questions. We discovered that particular difficulties lie in the ambiguous replies by the patients: a patient modifies a question which makes them answer yes or no, a response does not the corresponding reply to the question, a reply involves some part yes and some part no, and so on. Focusing on this particular type of question we introduce a text classifier using Long Short-Term Memory (LSTM) and build a corpus using Twitter.

    DOI: 10.1145/3267305.3267704

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85058328754&origin=inward

  • Dialogue breakdown detection with long short term memory 査読有り

    Mairittha T., Okita T., Inoue S.

    Lecture Notes of the Institute for Computer Sciences, Social-Informatics and Telecommunications Engineering, LNICST   240   245 - 250   2018年01月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    This paper aims to detect the utterance which can be categorized as the breakdown of the dialogue flow. We propose a logistic regression-based and a Long Short-Term Memory (LSTM)-based methods. Using the input with utterance-response pairs, the performance of the LSTM-based method is superior to that of the logistic regression-based method in 36% measured with F-measure. We also measured the performance using the performance with utterance-response pairs: the performance with the input only with responses is unexpectedly inferior to those with responses in 6% to 23% measured with F-measure.

    DOI: 10.1007/978-3-319-90740-6_18

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85046867198&origin=inward

  • ダンスの上手い人のマイニング的な分析

    大北 剛, 井上 創造

    人工知能学会第二種研究会資料 ( 一般社団法人 人工知能学会 )   2018 ( 18 )   03   2018年01月

     詳細を見る

    担当区分:筆頭著者   記述言語:英語   掲載種別:研究論文(研究会,シンポジウム資料等)

    <p>IoT において,「歩く」,「立ち上がる」などの言語による行動のラベルを目的としたセンサからの行動認識を 可視化する技術は,「歩く」,「立ち上がる」という言語による行動のラベルを目的とした映像からの行動認識との 技術の融合を意味する.これは一転して,「歩く」,「立ち上がる」などの言語による行動のラベルのバイアスを排 除する新たな行動認識の形を提案し,新たなマイニングのモデルを提案する. ダンスの上手い人と下手な人のどこ が具体的に異なるかをセンサと映像からのマルチモダルな行動認識から探るプラットフォームの構築を報告する.</p>

    DOI: 10.11517/jsaisigtwo.2018.AM-18_03

    CiNii Article

    CiNii Research

    その他リンク: https://ci.nii.ac.jp/naid/130008080212

  • Parallelization of Neural Network Training for NLP with Hogwild! 査読有り 国際誌

    Valentin Deyringer and Alexander Fraser and Helmut Schmid and Tsuyoshi Okita

    The Prague Bulletin of Mathematical Linguistics ( Charles University, Czech Republic )   109   29 - 38   2017年10月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(学術雑誌)

    DOI: 10.1515/pralin-2017-0036

  • Recognition of multiple overlapping activities using compositional cnn-lstm model 査読有り

    Okita T., Inoue S.

    UbiComp/ISWC 2017 - Adjunct Proceedings of the 2017 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2017 ACM International Symposium on Wearable Computers   165 - 168   2017年09月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    This paper introduces a new task, a recognition of multiple overlapping activities in the context of activity recognition. We propose a compositional CNN+LSTM algorithm. The experimental results show on the artificial dataset that it improved the accuracy from 27% to 43%.

    DOI: 10.1145/3123024.3123095

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85030836480&origin=inward

  • The DCU discourse parser: A sense classification task 査読有り 国際誌

    Okita T., Wang L., Liu Q.

    CoNLL 2015 - 19th Conference on Computational Natural Language Learning, Proceedings of the Shared Task   71 - 77   2015年01月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    This paper describes the discourse parsing system developed at Dublin City University for participation in the CoNLL 2015 shared task. We participated in two tasks: a connective and argument identification task and a sense classification task. This paper focuses on the latter task and especially the sense classification for implicit connectives.

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85072752049&origin=inward

  • The DCU discourse parser for connective, argument identification and explicit sense classification 査読有り 国際誌

    Wang L., Hokamp C., Okita T., Zhang X., Liu Q.

    CoNLL 2015 - 19th Conference on Computational Natural Language Learning, Proceedings of the Shared Task   89 - 94   2015年01月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    This paper describes our submission to the CoNLL-2015 shared task on discourse parsing. We factor the pipeline into sub-components which are then used to form the final sequential architecture. Focusing on achieving good performance when inferring explicit discourse relations, we apply maximum entropy and recurrent neural networks to different sub-tasks such as connective identification, argument extraction, and sense classification. The our final system achieves 16.51%, 12.73% and 11.15% overall F1 scores on the dev, WSJ and blind test sets, respectively.

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84973379014&origin=inward

  • The DCU Terminology Translation System for the Medical Query Subtask at WMT14 査読有り 国際誌

    Xiaofeng Wu, Rejwanul Haque, Tsuyoshi Okita, Piyush Arora, Andy Way, Qun Liu

    Proceedings of the 9th Workshop on Statistical Machine Translation   2014年06月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

  • Active Learning-based Local Graph Matching for Textual Entailment 査読有り 国際誌

    Tsuyoshi Okita

    Proceedings of the 10th International Symposium on Natural Language Processing   2013年10月

     詳細を見る

    担当区分:筆頭著者   記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

  • Shallow Semantically-Informed PBSMT and HPBSMT 査読有り 国際誌

    Tsuyoshi Okita, Qun Liu, Josef van Genabith

    Proceedings of the 8th Workshop on Statistical Machine Translation   2013年08月

     詳細を見る

    担当区分:筆頭著者   記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

  • Joint Space Neural Probabilistic Language Model for Statistical Machine Translation 国際誌

    Tsuyoshi Okita

    arxiv   2013年01月

     詳細を見る

    担当区分:筆頭著者   記述言語:英語   掲載種別:研究論文(大学,研究機関等紀要)

    arXiv

  • Results from the ML4HMT-12 Shared Task on Applying Machine Learning Techniques to Optimise the Division of Labour in Hybrid Machine Translation 査読有り 国際誌

    Christian Federmann, Tsuyoshi Okita, Maite Melero, Marta R. Costa-Jussa, Toni Badia and Josef van Genabith

    Proceedings of ML4HMT Workshop   2012年12月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

  • Neural Probabilistic Language Model for System Combination 査読有り 国際誌

    Tsuyoshi Okita

    Proceedings of ML4HMT Workshop   2012年12月

     詳細を見る

    担当区分:筆頭著者   記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

  • Topic Modeling-based Domain Adaptation for System Combination 査読有り 国際誌

    Tsuyoshi Okita and Antonio Toral and Josef van Genabith

    Proceedings of ML4HMT Workshop   2012年12月

     詳細を見る

    担当区分:筆頭著者   記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

  • Sentence-level Quality Estimation for MT System Combination 査読有り 国際誌

    Tsuyoshi Okita and Raphael Rubino and Josef van Genabith

    Proceedings of ML4HMT Workshop   2012年12月

     詳細を見る

    担当区分:筆頭著者   記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

  • System Combination with Extra Alignment Information 査読有り 国際誌

    Xiaofeng Wu, Tsuyoshi Okita, Josef van Genabith, and Qun Liu

    Proceedings of ML4HMT Workshop   2012年12月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

  • Minimum Bayes risk decoding with enlarged hypothesis space in system combination 査読有り

    Okita T., Van Genabith J.

    Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)   7182 LNCS ( PART 2 )   40 - 51   2012年03月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    This paper describes a new system combination strategy in Statistical Machine Translation. Tromble et al. (2008) introduced the evidence space into Minimum Bayes Risk decoding in order to quantify the relative performance within lattice or n-best output with regard to the 1-best output. In contrast, our approach is to enlarge the hypothesis space in order to incorporate the combinatorial nature of MBR decoding. In this setting, we perform experiments on three language pairs ES-EN, FR-EN and JP-EN. For ES-EN JRC-Acquis our approach shows 0.50 BLEU points absolute and 1.9% relative improvement obver the standard confusion network-based system combination without hypothesis expansion, and 2.16 BLEU points absolute and 9.2% relative improvement compared to the single best system. For JP-EN NTCIR-8 the improvement is 0.94 points absolute and 3.4% relative, and for FR-EN WMT09 0.30 points absolute and 1.3% relative compared to the single best system, respectively. © 2012 Springer-Verlag.

    DOI: 10.1007/978-3-642-28601-8_4

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84858307159&origin=inward

  • Annotated corpora for word alignment between Japanese and english and its evaluation with MAP-based word aligner 査読有り

    Okita T.

    Proceedings of the 8th International Conference on Language Resources and Evaluation, LREC 2012   3241 - 3248   2012年01月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    This paper presents two annotated corpora for word alignment between Japanese and English. We annotated on top of the IWSLT-2006 and the NTCIR-8 corpora. The IWSLT-2006 corpus is in the domain of travel conversation while the NTCIR-8 corpus is in the domain of patent. We annotated the first 500 sentence pairs from the IWSLT-2006 corpus and the first 100 sentence pairs from the NTCIR-8 corpus. After mentioned the annotation guideline, we present two evaluation algorithms how to use such hand-annotated corpora: although one is a well-known algorithm for word alignment researchers, one is novel which intends to evaluate a MAP-based word aligner of Okita et al. (2010b).

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85037351924&origin=inward

  • Statistical machine translation with factored translation model: MWEs, separation of affixes, and others 査読有り

    Okita T., Ceausu A., Way A.

    Proceedings of the 24th International Florida Artificial Intelligence Research Society, FLAIRS - 24   353 - 354   2011年09月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    This paper discusses Statistical Machine Translation when the target side is morphologically richer language. This paper intends to discuss the issues which are not covered by a factored translation model of Moses especially targetting EN-JP translation: the effect of Multi-Word Expressions, the separation of affixes, and other monolingual morphological issues. We intend to discuss these over a factored translation model. Copyright © 2011, Association for the Advancement of Artificial Intelligence. All rights reserved.

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=80052409677&origin=inward

  • Given bilingual terminology in statistical machine translation: MWE-sensitive word alignment and hierarchical Pitman-Yor process-based translation model smoothing 査読有り

    Okita T., Way A.

    Proceedings of the 24th International Florida Artificial Intelligence Research Society, FLAIRS - 24   269 - 274   2011年09月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    This paper considers a scenario when we are given almost perfect knowledge about bilingual terminology in terms of a test corpus in Statistical Machine Translation (SMT). When the given terminology is part of a training corpus, one natural strategy in SMT is to use the trained translation model ignoring the given terminology. Then, two questions arises here. 1) Can a word aligner capture the given terminology? This is since even if the terminology is in a training corpus, it is often the case that a resulted translation model may not include these terminology. 2) Are probabilities in a translation model correctly calculated? In order to answer these questions, we did experiment introducing a Multi-Word Expression-sensitive (MWE-sensitive) word aligner and a hierarchical Pitman-Yor process-based translation model smoothing. Using 200k JP-EN NTCIR corpus, our experimental results show that if we introduce an MWE-sensitive word aligner and a new translation model smoothing, the overall improvement was 1.35 BLEU point absolute and 6.0% relative compared to the case we do not introduce these two. Copyright © 2011, Association for the Advancement of Artificial Intelligence. All rights reserved.

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=80052406399&origin=inward

  • Pitman-Yor Process-based Language Mode 査読有り 国際誌

    Tsuyoshi Okita and Andy Way

    International Journal of Asian Language Processing   21 ( 2 )   55 - 70   2011年04月

     詳細を見る

    担当区分:筆頭著者   記述言語:英語   掲載種別:研究論文(学術雑誌)

  • Pitman-Yor Process-based Language Model 査読有り 国際誌

    Tsuyoshi Okita, Andy Way

    International Journal of Asian Language Processing   21 ( 2 )   57 - 70   2011年03月

     詳細を見る

    担当区分:筆頭著者   記述言語:英語   掲載種別:研究論文(学術雑誌)

  • MWE-sensitive Word Aligner in Factored Translation Model 国際誌

    Tsuyoshi Okita, Andy Way

    Proceedings of Machine Translation and Morphologically-rich Languages   2011年01月

     詳細を見る

    担当区分:筆頭著者   記述言語:英語   掲載種別:研究論文(研究会,シンポジウム資料等)

    Israel   Haifa   2011年01月23日  -  2011年01月27日

  • Hierarchical Pitman-Yor language model for machine translation 査読有り

    Okita T., Way A.

    Proceedings - 2010 International Conference on Asian Language Processing, IALP 2010   245 - 248   2010年12月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    The hierarchical Pitman-Yor process-based smoothing method applied to language model was proposed by Goldwater and by Teh; the performance of this smoothing method is shown comparable with the modified Kneser-Ney method in terms of perplexity. Although this method was presented four years ago, there has been no paper which reports that this language model indeed improves translation quality in the context of Machine Translation (MT). This is important for the MT community since an improvement in perplexity does not always lead to an improvement in BLEU score; for example, the success of word alignment measured by Alignment Error Rate (AER) does not often lead to an improvement in BLEU. This paper reports in the context of MT that an improvement in perplexity really leads to an improvement in BLEU score. It turned out that an application of the Hierarchical Pitman-Yor Language Model (HPYLM) requires a minor change in the conventional decoding process. Additionally to this, we propose a new Pitman-Yor process-based statistical smoothing method similar to the Good-Turing method although the performance of this is inferior to HPYLM. We conducted experiments; HPYLM improved by 1.03 BLEU points absolute and 6% relative for 50k EN-JP, which was statistically significant. © 2010 IEEE.

    DOI: 10.1109/IALP.2010.34

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=79551520331&origin=inward

  • Gap Between Theory and Practice: Noise Sensitive Word Alignment in Machine Translation 査読有り 国際誌

    Tsuyoshi Okita, Yvette Graham, Andy Way

    Journal of Machine Learning Research Workshop and Conference Proceedings   11   119 - 126   2010年09月

     詳細を見る

    担当区分:筆頭著者   記述言語:英語   掲載種別:研究論文(学術雑誌)

  • Multi-Word Expression-Sensitive Word Alignment 査読有り 国際誌

    Tsuyoshi Okita, Alfredo Maldonado Guerra, Yvette Graham, Andy Way

    In Proceedings of the Fourth International Workshop On Cross Lingual Information Access ( Coling 2010 Organizing Committee )   26 - 33   2010年08月

     詳細を見る

    担当区分:筆頭著者   記述言語:英語   掲載種別:研究論文(学術雑誌)

    China   Beijing   2010年08月28日  -  2010年08月28日

    This paper presents a new word align-
    ment method which incorporates knowl-
    edge about Bilingual Multi-Word Expres-
    sions (BMWEs). Our method of word
    alignment first extracts such BMWEs in
    a bidirectional way for a given corpus and
    then starts conventional word alignment,
    considering the properties of BMWEs in
    their grouping as well as their alignment
    links. We give partial annotation of align-
    ment links as prior knowledge to the word
    alignment process; by replacing the max-
    imum likelihood estimate in the M-step
    of the IBM Models with the Maximum A
    Posteriori (MAP) estimate, prior knowl-
    edge about BMWEs is embedded in the
    prior in this MAP estimate. In our exper-
    iments, we saw an improvement of 0.77
    Bleu points absolute in JP–EN. Except
    for one case, our method gave better re-
    sults than the method using only BMWEs
    grouping. Even though this paper does
    not directly address the issues in Cross-
    Lingual Information Retrieval (CLIR), it
    discusses an approach of direct relevance
    to the field. This approach could be
    viewed as the opposite of current trends
    in CLIR on semantic space that incorpo-
    rate a notion of order in the bag-of-words
    model (e.g. co-occurences).

    Kyutacar

  • Data cleaning for word alignment 査読有り

    Okita T.

    ACL-IJCNLP 2009 - Joint Conf. of the 47th Annual Meeting of the Association for Computational Linguistics and 4th Int. Joint Conf. on Natural Language Processing of the AFNLP, Proceedings of the Conf.   72 - 80   2009年01月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    Parallel corpora are made by human beings. However, as an MT system is an aggregation of state-of-the-art NLP technologies without any intervention of human beings, it is unavoidable that quite a few sentence pairs are beyond its analysis and that will therefore not contribute to the system. Furthermore, they in turn may act against our objectives to make the overall performance worse. Possible unfavorable items are n : m mapping objects, such as paraphrases, non-literal translations, and multiword expressions. This paper presents a pre-processing method which detects such unfavorable items before supplying them to the word aligner under the assumption that their frequency is low, such as below 5 percent. We show an improvement of Bleu score from 28.0 to 31.4 in English-Spanish and from 16.9 to 22.1 in German-English. © 2009 ACL and AFNLP.

    DOI: 10.3115/1667884.1667895

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=80052416271&origin=inward

  • Low-Resource Machine Translation Using MATREX: The DCU Machine Translation System for IWSLT 2009 査読有り

    Ma Y., Okita T., Çetinoǧlu Ö., Du J., Way A.

    2009 International Workshop on Spoken Language Translation, IWSLT 2009   29 - 36   2009年01月

     詳細を見る

    記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    In this paper, we give a description of the Machine Translation (MT) system developed at DCU that was used for our fourth participation in the evaluation campaign of the International Workshop on Spoken Language Translation (IWSLT 2009). Two techniques are deployed in our system in order to improve the translation quality in a low-resource scenario. The first technique is to use multiple segmentations in MT training and to utilise word lattices in decoding stage. The second technique is used to select the optimal training data that can be used to build MT systems. In this year's participation, we use three different prototype SMT systems, and the output from each system are combined using standard system combination method. Our system is the top system for Chinese-English CHALLENGE task in terms of BLEU score.

    Scopus

    その他リンク: https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85133165029&origin=inward

  • Load Balance Protocol of Cluster on Grid: Pervasive Maximum Algorithmic Parallelism 査読有り 国際誌

    Tsuyoshi Okita

    Procedings of the 6th International Conference on Principles of Distributed Systems (OPODIS 2002) ( Suger, Saint-Denis, rue Catulienne, France )   3   203 - 210   2002年12月

     詳細を見る

    担当区分:筆頭著者   記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    France   Reims   2002年12月11日  -  2002年12月13日

    Pervasive computing software help to manage information and reduce the complexity of available computing resources in a timely manner. On the other hand, a grid is a resource whose information is changing anytime and anywhere, where the availability of CPUs is only informed in run time when a cluster asks to the grid. On the other hand, a cluster is often implemented in a static way, which assumes some particular parallel architecture and the number of CPUs. Even when a cluster can consume maximum available CPU resources, if a cluster i implemented assuming the number of CPUs, a cluster could not run using more than this number of CPUs. Our mechanim of load balance protocol of cluster provides a dynamic way of implementing a cluster that can consume maximum available CPU resources on run time. In order to do so, we propoe two mechanisms: to provide yet another (graphical) parallel language to describe clusters and to provide the protocols between a cluster on a grid. While many paralell languages resolve parallel architecture dependencies in compile time, our parallel language resolve parallel architecture dependencies in run time. Our load balance protocol bases on this (graphical) parallel language and it provides implementation of them.

    Kyutacar

  • PRTccp: Priority-driven Real-Time Concurrent Constraint Programming 査読有り 国際誌

    Tsuyoshi Okita

    Proceedings of the 14th Nordic Workshop on Programming Theory   2002年11月

     詳細を見る

    担当区分:筆頭著者   記述言語:英語   掲載種別:研究論文(国際会議プロシーディングス)

    Estonia   Tallinn   2002年11月20日  -  2002年11月22日

    This paper presents real-time formal language PRTccp, which complements priority-driven concerns in Tccp. First, we showed the necessity of formal language for priority-driven system compared to reactive real-time system. Secondly, we showed the grammar of PRTccp. Thirdly, we showed a small example of scheduler.

    Kyutacar

▼全件表示

口頭発表・ポスター発表等

  • JPEGの画像表現を用いた画像生成の高速化

    大北剛, 管谷克彦, 坂本比呂志

    IBIS 2021 

     詳細を見る

    開催期間: 2021年11月10日 - 2021年11月12日   記述言語:英語  

  • グローバルな情報を加味するセマンティックセグメンテーションとラベルの重複を許す分類のジョイント学習

    平野北斗, 竹本和広, 大北剛

    第23回情報理論的学習理論ワークショップ(IBIS2020) 

     詳細を見る

    開催期間: 2020年11月23日 - 2020年11月26日   記述言語:英語   開催地:Virtual  

  • Spatio-temporal Model for Intracerebral Hemorrhage: Embedding Methods Solving Violating Assumptions on ML

    Tsuyoshi Okita

    LRML Learning Meaningful Representations of Life Workshop 

     詳細を見る

    開催期間: 2019年12月13日   記述言語:英語   開催地:Vancouver, Canada  

  • 欠損値問題や不均衡データを埋め込みとして用いた時空間機械学習モデル

    大北剛

    欠損値問題や不均衡データを埋め込みとして用いた時空間機械学習モデル  IBIS2019

     詳細を見る

    開催期間: 2019年11月20日 - 2019年11月23日   記述言語:日本語  

  • Translation of Signals: Wave2wave

    Tsuyoshi Okita, Hirotaka Hachiya, Sozo Inoue, Naonori Ueda

    Discovery Science 

     詳細を見る

    開催期間: 2019年10月28日 - 2019年10月30日   記述言語:英語  

  • メビウス型包除積分ニューラルネットワークによるデータ解析

    本田あおい, 大北剛

    実解析学シンポジウム 

     詳細を見る

    開催期間: 2019年10月25日 - 2019年10月27日   記述言語:日本語  

  • Cross Modal Translation of Signals and Its Application to Activity Recognition 招待有り

    Tsuyoshi Okita, Hirotaka Hachiya, Sozo Inoue, Naonori Ueda

    The First Japan-Israel Machine Learning Workshop 

     詳細を見る

    開催期間: 2018年11月19日 - 2018年11月20日   記述言語:英語   開催地:Tel Aviv, Israel  

  • Language Grounded Planning for Dialogue Systems

    Tsuyoshi Okita and Sozo Inoue

    NVIDIA's GPU Technology Conference (GTC)  

     詳細を見る

    開催期間: 2018年09月13日 - 2018年09月14日   記述言語:英語  

  • Language Grounded Activity Recognition and Planning

    Tsuyoshi Okita and Sozo Inoue

    Second International Workshop on Symbolic-Neural Learning (SNL-2018) 

     詳細を見る

    開催期間: 2018年07月05日 - 2018年07月06日   記述言語:英語   開催地:Nagoya  

  • Brazilator: Machine Translation and Sentiment Analysis for World Cup 2014

    Santiago Cortes, Piyush Arora, Chris Hokamp, Federico Fancellu, Alex Killen, Lamia Tounsi, Antonio Toral, Ankit Srivastava, Maria Alecu, Iacer Calixto, Sheila Castilho, Keith Curtis, Federico Gaspari, Akira Hayakawa, Teresa Lynn, Peyman Passban, Eziz Tursun, Ali Hosseinzadeh Vahid, Xiaofeng Wu, Xiaojun Zhang, Debasis Ganguly, Louise Irwin, Anna Kostekidou, Liangyou Li, Tsuyoshi Okita, Ximo Planells, David Racca, Joris Vreeke, Jian Zhang, Andy Way, Will Lewis, Declan Groves, Federico Garcea, and Chris Wendt

    Association for Machine Translation in the Americas  AMTA

     詳細を見る

    開催期間: 2014年10月22日 - 2014年10月26日   記述言語:英語   開催地:Vancouver, Canada  

    デモ発表

▼全件表示

工業所有権

  • データ予測装置及びデータ予測方法

    大北 剛、井上 創造

     詳細を見る

    出願番号:特開2020-61145(P2020-61145A)  出願日:2019年10月04日

    ディープラーニング(深層学習)を用いて異なるモダリティ間でのデータの変換を行うことにより、各種のモダリティを有効活用して、その活用範囲を広げることができるデータ予測装置及びデータ予測方法を提供する。

  • 計表処理装置および計表処理方法、並びに伝送媒体

    大北 剛

     詳細を見る

    出願番号:特願平9−283310  出願日:1997年10月16日

    公開番号:特開平11−120257  公開日:1999年04月30日

    【課題】計表の行側の項目および列側の項目を、階層との整合を図りながら作成することができるようにする。
    【解決手段】階層V−2を3個の項目に分割したとき、その上位の階層V−1は変更しないが、下位の階層V−3は、階層V−2に対応して3個の項目に分割する。行の項目と列の項目で規定される文字を入力する領域も、項目に対応して分割する。

講演

  • 生成AIに関する研究

    明専会広島支部総会  2024年08月  明専会

     詳細を見る

    発表言語:日本語   講演種別:招待講演  

  • NeurIPS2020における深層学習の研究動向

    AIトレンドトップカンファレンス報告会  2021年03月  人工知能学会

     詳細を見る

    発表言語:日本語   講演種別:特別講演  

    NeurIPS2020における深層学習の研究動向をレポートした.

学術関係受賞

  • Outstanding Article Award --- 2021 Editor's Pick: Computer Science

    Frontiers   2022年01月28日

    Lin Wang, Hristijan Gjoreski, Mathias Ciliberto, Paula Lago, Kazuya Murao, Tsuyoshi Okita, Daniel Roggen

     詳細を見る

    受賞国:スイス連邦

  • Best Paper Award

    Activity and Behavior Computing国際会議   2020年08月29日

    Kohei Adachi, Paula Lago, Tsuyoshi Okita, Sozo Inoue

     詳細を見る

    受賞国:日本国

    Improvement of Human Action Recognition Using 3D Pose Estimation

科研費獲得実績

  • センサデータに基づく大規模基盤モデル・生成AIの開発

    研究課題番号:24K15107  2024年04月 - 2027年03月   基盤研究(C)

  • 計算モデルにガイドされた急成長を伴う時空間モデルの開発

    研究課題番号:20K12065  2020年04月 - 2024年03月   基盤研究(C)

  • Precision medicineの確立に資する統合医療データベースの利活用に関する研究

    研究課題番号:19AC1003  2019年04月 - 2021年03月   厚生労働科学研究費補助金 行政政策研究分野 政策科学総合研究(臨床研究等ICT基盤構築・人工知能実装研究)

その他競争的資金獲得実績

  • Precision medicineの確立に資する統合医療データベースの利活用に関する研究

    2019年04月 - 2022年03月

    厚生労働省 科学研究費補助金 政策科学総合研究事業  

担当授業科目(学内)

  • 2021年度   データサイエンス演習I

     詳細を見る

    科目区分:大学院専門科目

  • 2021年度   深層学習特論

     詳細を見る

    科目区分:大学院専門科目

  • 2020年度   深層学習特論

     詳細を見る

    科目区分:大学院専門科目

  • 2023年度   ビジョンと言語の深層学習特論MI

     詳細を見る

    科目区分:大学院専門科目

  • 2023年度   ビジョンと言語の深層学習特論AI

     詳細を見る

    科目区分:大学院専門科目

  • 2023年度   ビジョンと言語の深層学習特論DS

     詳細を見る

    科目区分:大学院専門科目

  • 2023年度   深層学習

     詳細を見る

    科目区分:学部専門科目

  • 2023年度   知能情報工学プロジェクト

     詳細を見る

    科目区分:学部専門科目

  • 2022年度   データサイエンス演習I

     詳細を見る

    科目区分:大学院専門科目

  • 2022年度   データサイエンス演習Ⅱ

     詳細を見る

    科目区分:大学院専門科目

  • 2022年度   深層学習特論

     詳細を見る

    科目区分:大学院専門科目

  • 2021年度   データサイエンス演習Ⅱ

     詳細を見る

    科目区分:大学院専門科目

  • 2021年度   知能情報工学プロジェクト

     詳細を見る

    科目区分:学部教養科目

  • 2020年度   知能情報工学プロジェクト

     詳細を見る

    科目区分:学部専門科目

  • 2020年度   データサイエンス演習Ⅱ

     詳細を見る

    科目区分:大学院専門科目

  • 2020年度   データサイエンス演習I

     詳細を見る

    科目区分:大学院専門科目

  • 2019年度   データサイエンス演習Ⅱ

     詳細を見る

    科目区分:大学院専門科目

  • 2019年度   データサイエンス演習I

     詳細を見る

    科目区分:大学院専門科目

▼全件表示

FD活動への参加

  • 2022年10月   ルーブリック評価入門

学会・委員会等活動

  • DICOMOシンポジウム   プログラム委員  

    2023年02月 - 2023年09月

社会貢献活動(講演会・出前講義等)

  • データサイエンスプロ短期コース/深層学習特化型公開講座

    役割:講師, 運営参加・支援

    2023年03月14日 - 2023年03月30日

     詳細を見る

    対象: 教育関係者, 研究者, 社会人・一般, 企業

    種別:その他

  • データサイエンスプロ短期コース/深層学習特化型公開講座

    役割:講師, 企画, 運営参加・支援

    九州工業大学  データサイエンスプロ短期コース/深層学習特化型公開講座   2022年02月22日 - 2022年03月31日

     詳細を見る

    対象: 研究者, 社会人・一般

    種別:その他

    全15コマの講義で, 参加者は33人. 165万円の収益を得た.

  • データサイエンスプロ短期コース/深層学習特化型公開講座

    役割:講師, 運営参加・支援

    九州工業大学  データサイエンスプロ短期コース/深層学習特化型公開講座  Virtual  2021年03月02日 - 2021年03月30日

     詳細を見る

    対象: 研究者, 社会人・一般

    種別:出前授業

    深層学習の講座全14コマ, 火木コースと土日コースの2回実施(累計28コマの講義)
    火木コース(3月2日,4日,9日,11日,16日,18日,23日,25日,30日)
    土日コース(3月6日,14日,21日,27日,28日)
    講師: 大北 剛, 井上 創造, 齊藤 剛史, 徳永 旭将, 竹本 和広

    18人が火曜木曜の平日コース, 5人が土曜日曜の週末コースであった.

    全体で23人X5万円=115万円を得た.

  • データサイエンスプロ短期コース/機械学習講座

    役割:講師, 企画, 運営参加・支援

    2019年04月01日 - 2020年03月31日

     詳細を見る

    対象: 研究者, 社会人・一般

    種別:出前授業

    機械学習講座をトヨタ自動車九州(宮田)にて行う. 3コマ全17回.

国際会議開催(学会主催除く)

  • HASCA 2023 Workshop

    Kazuya Murao,Yu Enokibori,Hristijan Gjoreski,Paula Lago,Tsuyoshi Okita,Pekka Siirtola,Kei Hiroi,Philipp M. Scholl,Mathias Ciliberto,Kenta Urano  Cancun, Mexico  2023年10月08日 - 2023年10月10日

  • HASCA 2022 workshop

    ACM UbiComp  2022年09月15日

  • HASCA 2021 Workshop

    Kazuya Murao, Yu Enokibori, Hristijan Gjoreski, Paula Lago, Tsuyoshi Okita, Pekka Siirtola, Kei Hiroi, Philipp M Scholl, Mathias Ciliberto  Virtual  2021年09月21日

  • HASCA 2020 Workshop

    Kazuya Murao, Yu Enokibori, Hristijan Gjoreski, Paula Lago, Tsuyoshi Okita, Pekka Siirtola, Kei Hiroi, Philipp M Scholl, Mathias Ciliberto  Virtual  2020年09月12日

  • HASCA 2019 Workshop

    K Murao, Y Enokibori, H Gjoreski, P Lago, T Okita, P Siirtola, K Hiroi  UK  2019年09月09日

  • HASCA 2018 Workshop

    K Murao, Y Enokibori, H Gjoreski, P Lago, T Okita, P Siirtola, K Hiroi  Singapore  2018年10月12日

  • Deep Learning for Machine Translation Winter School

    Qun Liu, Tsuyoshi Okita, Chris Hokamp, John Judge, Joachim Wagner  Ireland  2015年10月18日 - 2015年10月24日

  • COLING workshop ML4HMT

    Christian Federmann, Tsuyoshi Okita, Maite Melero, Marta R. Costa-Jussa, Toni Badia and Josef van Genabith  India  2012年12月08日 - 2012年12月09日

  • AMTA workshop on Monolingual Machine Translation

    Tsuyoshi Okita, Artem Sokolov, Taro Watanabe  US, San Diego  2012年11月01日

▼全件表示