MIYAMOTO Hiroyuki

写真a

Title

Associate Professor

Laboratory

2-4 Hibikino, Wakamatsu-ku, Kitakyushu-shi, Fukuoka

Research Fields, Keywords

Neural Network, Robot

E-mail

E-mail address

Phone

+81-93-695-6126

Fax

+81-93-695-6126

Undergraduate Education 【 display / non-display

  • 1985.03   Osaka University   Faculty of Engineering Science   Graduated   JAPAN

Post Graduate Education 【 display / non-display

  • 1994.03  Osaka University  Graduate School, Division of Engineering Science  Doctoral Program  Accomplished Credits for Doctoral Program  JAPAN

Degree 【 display / non-display

  • Osaka University -  Doctor of Engineering  1999.01

Biography in Kyutech 【 display / non-display

  • 2014.04
    -
    Now

    Kyushu Institute of TechnologyGraduate School of Life Science and Systems Engineering   Department of Human Intelligence Systems   Associate Professor  

  • 2007.04
    -
    2014.03

    Kyushu Institute of TechnologyGraduate School of Life Science and Systems Engineering   Department of Brain Science and Engineering   Associate Professor  

Biography before Kyutech 【 display / non-display

  • 1996.10
    -
    2001.03

    Kawato Dynamic Brain Project JST   Researcher   JAPAN

  • 1994.04
    -
    1996.09

    ATR Human Information Processing Research Lab   Researcher   JAPAN

  • 1987.04
    -
    1989.01

    Sumitomo Electric Industries   Researcher   JAPAN

Academic Society Memberships 【 display / non-display

  • 2003.04
    -
    Now
     

    Japanese Neural Network Society  JAPAN

  • 2003.04
    -
    Now
     

    The institue of electronics, information and communication engineers (IEICE)  JAPAN

 

Publications (Article) 【 display / non-display

  • A source domain extension method for inductive transfer learning based on flipping output

    Koishi Y., Ishida S., Tabaru T., Miyamoto H.

    Algorithms    12 ( 5 )   2019.05  [Refereed]

     View Summary

    © 2019 by the authors. Transfer learning aims for high accuracy by applying knowledge of source domains for which data collection is easy in order to target domains where data collection is difficult, and has attracted attention in recent years because of its significant potential to enable the application of machine learning to a wide range of real-world problems. However, since the technique is user-dependent, with data prepared as a source domain which in turn becomes a knowledge source for transfer learning, it often involves the adoption of inappropriate data. In such cases, the accuracy may be reduced due to "negative transfer." Thus, in this paper, we propose a novel transfer learning method that utilizes the flipping output technique to provide multiple labels in the source domain. The accuracy of the proposed method is statistically demonstrated to be significantly better than that of the conventional transfer learning method, and its effect size is as high as 0.9, showing high performance.

    DOI Scopus

  • Improvement of Sampling Inspection for Wire Bonding Using Thin AE Sensor and Transfer Learning

    Koishi Yasutake, Ishida Shuichi, Tabaru Tatsuo, Iwasaki Wataru, Miyamoto Hiroyuki

    Journal of The Japan Institute of Electronics Packaging  ( The Japan Institute of Electronics Packaging )  22 ( 3 ) 209 - 217   2019.01  [Refereed]

     View Summary

    <p>Wire bonding is an important process in semiconductor manufacturing. The quality of wire bonding is guaranteed by destructive sampling inspection. However, sampling inspection is not able to detect defective products that occur non-systematically. Therefore, it is desirable to shift to inspection of each part. In this paper, we propose an inspection method based on the wire bonding state using a Thin-film AE sensor and transfer learning. We place the thin film AE sensor in the immediate vicinity of the bonding junction and measure the applied AE wave. We improve the quality determination performance in the sampling inspection by determining the quality from the measured AE waves using the transfer learning method we developed. We applied the proposed method to experiment using actual bonding samples and confirmed that it showed a 10.7% higher quality determination ability than sampling inspection.</p>

    DOI Scopus CiNii

  • Basic Study on Viewpoints Classification Method using Car Distribution on the Road

        2018.11  [Refereed]

     View Summary

    © 2018 IEEE. In order to provide a notification system for the blind awaiting a bus, a classification of viewpoints is necessary. In a previous study, we classified the viewpoints in non-congested traffic using road features extraction. However, the viewpoints in congested traffic have not yet been discussed; hence this paper proposes a viewpoints classification using car distribution on the road in congested traffic. The study aimed to classify the viewpoints to two classes as "Good Viewpoints" and "Bad Viewpoints" with 100 images for each class. The experimental results showed a 76% accuracy of the proposed method.

    DOI Scopus

  • Label Estimation Method with Modifications for Unreliable Examples in Taming

    Koishi Yasutake, Ishida Shuichi, Tabaru Tatsuo, Miyamoto Hiroyuki

    IJNC    8 ( 2 ) 153 - 165   2018.07

     View Summary

    Methods for improving learning accuracy by utilizing a plurality of data sets with different reliabilities have been studied extensively. Unreliable data sets often include data with incorrect labels, and the accuracy of learning from such data sets is thus affected. Here, we focused on a learning problem, Taming, which deals with two kinds of data sets with different reliabilities. We propose a label estimation method for use in data sets that include data with incorrect labels. The proposed method is an extension of BaggTaming, which has been proposed as a solution to Taming. We conducted experiments to verify the effectiveness of the proposed method by using a benchmark data set in which the labels were intentionally changed to make them incorrect. We confirmed that learning accuracy could be improved by using the proposed method and data sets with modified labels.

    DOI CiNii

  • Anomaly Detection of Rotary Vacuum Pump Using Thin AE Sensor and Reconstruction Error of Autoencoder

    UCHIDA Masato, ISHIDA Shuichi, TABARU Tatsuo, MIYAMOTO Hiroyuki

    TSICE    54 ( 7 ) 599 - 605   2018.07

     View Summary

    <p>A vacuum pump is always using in semiconductor manufacturing equipment. Anomaly of a vacuum pump leads to troubles such as stoppage of equipment. Therefore, anomaly detection of a vacuum pump is very important. We considered a vacuum pump state estimation using thin AE (Acoustic Emission) sensor and machine learning. However, anomaly data is not always available in a production line. Anomaly detector is necessary to learn from only normal data. There is an anomaly detection using autoencoder in recent years. In autoencoder, normal data is reconstructed and anomaly data is not reconstructed. As a result, anomaly detection can using reconstruction error. In this paper, we propose anomaly detection for rotary vacuum pump by thin AE sensor and reconstruction error of autoencoder. Performance of the proposed method is evaluated by detect exhaust anomaly of a vacuum pump.</p>

    DOI CiNii

display all >>

Publications (Books) 【 display / non-display

  • ロボット情報学ハンドブック、計算論に基づく模倣:ヒューマノイドロボットDB、項目名「見まねによる運動学習」

    宮本弘之、琴坂信哉 ( Joint Work )

    ナノオプトニクスエナジー  2010.03

  • Computing the optimal trajectory of arm movement: the TOPS (Task Optimization in the Presence of Signal-dependent noise) model

    ( Joint Work )

    2002.01

Conference Prsentations (Oral, Poster) 【 display / non-display

  • 薄型AEセンサとオートエンコーダを用いた回転式真空ポンプの排気不良の検出

    内田 雅人

    計測自動制御学会 産業応用部門2018年度大会  (国立研究開発法人産業技術総合研究所 臨海副都心センター 別館)  2018.11  -  2018.11  計測自動制御学会

  • 薄型AEセンサと転移学習を用いたワイヤボンディング抜取検査の性能向上

    小石泰毅

    エレクトロニクス実装学会 官能検査システム化技術 セッション  (東京理科大学 野田キャンパス(千葉県野田市山崎2641))  2018.03  -  2018.03  エレクトロニクス実装学会

  • 飼いならし学習を用いたAE源の位置標定

    内田雅人

    第36回計測自動制御学会九州支部学術講演会  (鹿児島大学郡元キャンパス(工学部))  2017.11  -  2017.11  計測自動制御学会

  • 多様性を考慮した飼いならし学習の精度向上に関する研究

    小石泰穀

    第36回計測自動制御学会九州支部学術講演会  (鹿児島大学郡元キャンパス(工学部))  2017.11  -  2017.11  計測自動制御学会

  • Development of User Interface that enables operation by physical action(poster)

    International Symposium on Applied Engineering and Sciences (SAES 2017)  2017.11  -  2017.11 

display all >>

 

Activities of Academic societies and Committees 【 display / non-display

  • 2000.04
    -
    Now

    The institue of electronics, information and communication engineers (IEICE)