• Search Research Projects
  • Search Researchers
  • How to Use
  1. Back to previous page

Data augmentation and domain adaptation using the latent space of the deep generative model

Research Project

Project/Area Number 19K12164
Research Category

Grant-in-Aid for Scientific Research (C)

Allocation TypeMulti-year Fund
Section一般
Review Section Basic Section 61040:Soft computing-related
Research InstitutionRitsumeikan University

Principal Investigator

Nishikawa Ikuko  立命館大学, 情報理工学部, 教授 (90212117)

Project Period (FY) 2019-04-01 – 2023-03-31
Project Status Completed (Fiscal Year 2022)
Budget Amount *help
¥4,420,000 (Direct Cost: ¥3,400,000、Indirect Cost: ¥1,020,000)
Fiscal Year 2021: ¥910,000 (Direct Cost: ¥700,000、Indirect Cost: ¥210,000)
Fiscal Year 2020: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Fiscal Year 2019: ¥2,340,000 (Direct Cost: ¥1,800,000、Indirect Cost: ¥540,000)
Keywords深層生成モデル / 潜在空間 / 異常検知 / ドメイン適応 / データ拡張 / エントロピー / 敵対的学習 / 機械学習 / データ修正
Outline of Research at the Start

深層ネットワークによる自己符号化器やデータ生成器など学習器を組合せて、有効なデータ多様体を獲得し活用する手法を開発する。データ認識や生成の学習で獲得される特徴量の潜在空間は、データを効率的に表現する低次元多様体である。本研究では、以下の意味で有効なデータ多様体の獲得手法を設計し、得られた多様体を用いて、①不良データの自動修正や異常検知、②学習データを増強するデータ拡張、③類似の多様体構造を持つ異なるドメインへの転移学習やドメイン適応の各手法を開発する。人工物データや生体データなどの実データを用いて、各領域での有効性の定量的検証とともに、データ駆動の機械学習による新たなデータ利活用法を構築する。

Outline of Final Research Achievements

The conducted research on the effective utilization of the latent space obtained by deep generative model has mainly focused on the universal domain adaptation and unsupervised anomaly detection. The methods of acquire the latent space include not only the simple simultaneous acquisition of the encoder and the generator by auto-encoder or variational auto-encoder, but also originally proposed networks: a combination of an generator obtained by the adversarial training followed by the encoder to its input space for the unsupervised anomaly detection, and a combination of multiple encoders, decoders, and a classifier to obtain a pair of task specific latent space and domain specific latent space for the universal domain adaptation. Then, a novel method using class-wise discriminators is proposed for universal domain adaptation for the classification task. Moreover, unsupervised anomaly detection is applied to some biological dynamical data as an approach of inverse genomics.

Academic Significance and Societal Importance of the Research Achievements

生成モデルが、画像や音声、言語など多様なデータ形式に対して、精度の高いデータを出力できるのは、データをその実世界での姿に依存しない形で抽出し表現できているからです。その表現を、実世界での姿に依らないデータの本質だと期待して、潜在表現と呼んでいます。対象とするデータに対して良い潜在表現を獲得する方法を提案し、潜在空間上でのデータ間の相互関係やデータ分布を使うことで、あるデータセットでの学習結果を他のデータセットに転用して利用できるドメイン適応や、異常データ検知器の構築に正常データだけを用いる学習などを、高精度で実現しました。

Report

(5 results)
  • 2022 Annual Research Report   Final Research Report ( PDF )
  • 2021 Research-status Report
  • 2020 Research-status Report
  • 2019 Research-status Report
  • Research Products

    (18 results)

All 2023 2022 2021 2020 2019

All Journal Article (2 results) (of which Peer Reviewed: 2 results) Presentation (16 results) (of which Int'l Joint Research: 9 results)

  • [Journal Article] 階層的探索による二次元配置配線最適化2023

    • Author(s)
      谷村亮介、東優貴、高野諒、西川郁子
    • Journal Title

      システム制御情報学会論文誌

      Volume: 36-7

    • Related Report
      2022 Annual Research Report
    • Peer Reviewed
  • [Journal Article] 敵対的学習によるクラス推定を利用した部分ドメイン適応2022

    • Author(s)
      甲野晴太、植田考哉、高野諒、西川郁子
    • Journal Title

      システム制御情報学会論文誌

      Volume: 66 Pages: 101-108

    • Related Report
      2021 Research-status Report
    • Peer Reviewed
  • [Presentation] Phenotype anomaly detection for biological dynamics data using a deep generative model2022

    • Author(s)
      Eisuke Ito, Takaya Ueda, Ryo Takano, Yukako Tohsato, Koji Kyoda, Shuichi Onami, and Ikuko Nishikawa
    • Organizer
      31st International Conference on Artificial Neural Networks (ICANN2022)
    • Related Report
      2022 Annual Research Report
    • Int'l Joint Research
  • [Presentation] 深層生成モデルの潜在空間を用いた生物動態データの異常検知2022

    • Author(s)
      伊藤瑛甫, 植田考哉, 高野諒, 遠里由佳子, 京田耕司, 大浪修一, 西川郁子
    • Organizer
      第66回システム制御情報学会 研究発表講演会(SCI'22)
    • Related Report
      2022 Annual Research Report
  • [Presentation] アイウェア型赤外線距離センサアレイを用いた視線ジェスチャ認識2022

    • Author(s)
      田渕裕貴,双見京介,高野諒,西川郁子
    • Organizer
      第66回システム制御情報学会 研究発表講演会(SCI'22)
    • Related Report
      2022 Annual Research Report
  • [Presentation] 線形モデルと非線形モデルを複合したモデルによる風向風速予測2022

    • Author(s)
      荒木章伍,宮野尚哉,西川郁子
    • Organizer
      第66回システム制御情報学会 研究発表講演会(SCI'22)
    • Related Report
      2022 Annual Research Report
  • [Presentation] 有望な部分解空間の絞り込みによる二段階最適化を用いた二次元配置最適化2022

    • Author(s)
      谷村亮介,東優貴, 高野諒, 西川郁子
    • Organizer
      第66回システム制御情報学会 研究発表講演会(SCI'22)
    • Related Report
      2022 Annual Research Report
  • [Presentation] Phenotype Anomaly Detection in Early C. elegans Embryos by Variational Auto-Encoder2021

    • Author(s)
      Takumi Oibayashi, Takaya Ueda, Yuki Kimura, Yukako Tohsato, and Ikuko Nishikawa
    • Organizer
      2021 IEEE 9th International Conference on Bioinformatics and Computational Biology (ICBCB)
    • Related Report
      2021 Research-status Report
    • Int'l Joint Research
  • [Presentation] 敵対的学習による分布間差異の推定を利用した部分ドメイン適応2021

    • Author(s)
      甲野晴太,植田考哉,西川郁子
    • Organizer
      第65回システム制御情報学会研究発表講演会(SCI’21)
    • Related Report
      2021 Research-status Report
  • [Presentation] 画像の構造的類似度を用いた自己符号化器による異常領域検出2021

    • Author(s)
      追林拓光,木村勇貴,植田考哉,西川郁子
    • Organizer
      第65回システム制御情報学会研究発表講演会(SCI’21)
    • Related Report
      2021 Research-status Report
  • [Presentation] 二階層のメタヒューリスティクスによる電子機器と配線経路の二次元最適化2021

    • Author(s)
      東優貴,利根大輝,山根悠,高野諒,西川郁子
    • Organizer
      第65回システム制御情報学会研究発表講演会(SCI’21)
    • Related Report
      2021 Research-status Report
  • [Presentation] Wasserstein Distance-Based Domain Adaptation and Its Application to Road Segmentation2021

    • Author(s)
      Seita Kono, Takaya Ueda, Enrique Arriaga-Varela and Ikuko Nishikawa
    • Organizer
      International Joint Conference on Neural Networks 2021, IJCNN2021
    • Related Report
      2020 Research-status Report
    • Int'l Joint Research
  • [Presentation] Temporal Anomaly Detection by Deep Generative Models with Applications to Biological Data2020

    • Author(s)
      Takaya Ueda, Yukako Tohsato and Ikuko Nishikawa
    • Organizer
      29th International Conference on Artificial Neural Networks, ICANN2020
    • Related Report
      2020 Research-status Report
    • Int'l Joint Research
  • [Presentation] Latent space decomposition into task-specific and domain-specific subspaces for domain adaptation2020

    • Author(s)
      Takaya Ueda and Ikuko Nishikawa
    • Organizer
      IEEE World Congress on Computational Intelligence 2020 (IEEE WCCI 2020)
    • Related Report
      2019 Research-status Report
    • Int'l Joint Research
  • [Presentation] Phenotype Anomaly Detection in Early C. elegans Embryos by Variational Auto-Encoder2020

    • Author(s)
      Takumi Oibayashi, Takaya Ueda, Yuki Kimura, Yukako Tohsato, and Ikuko Nishikawa
    • Organizer
      2020 IEEE 8th International Conference on Bioinformatics and Computational Biology (ICBCB 2020)
    • Related Report
      2019 Research-status Report
    • Int'l Joint Research
  • [Presentation] Analysis of Time Series Anomalies Using Causal InfoGAN and Its Application to Biological Data2019

    • Author(s)
      Takaya Ueda, Masataka Seo, Yukako Tohsato, Ikuko Nishikawa
    • Organizer
      The International Conference on Natural Computation, Fuzzy Systems and Knowledge Discovery (ICNC-FSKD) 2019
    • Related Report
      2019 Research-status Report
    • Int'l Joint Research
  • [Presentation] Characterizing Phenotype Abnormality by Variational Auto Encoder2019

    • Author(s)
      Yuki Kimura, Takaya Ueda, Seo Masataka, Yukako Tohsato, Ikuko Nishikawa
    • Organizer
      The International Conference on Natural Computation, Fuzzy Systems and Knowledge Discovery (ICNC-FSKD) 2019
    • Related Report
      2019 Research-status Report
    • Int'l Joint Research
  • [Presentation] Diversity Preservation in Genetic Algorithm by Lifespan Control2019

    • Author(s)
      Yu Yamane, Masataka Seo, Ikuko Nishikawa
    • Organizer
      The International Conference on Natural Computation, Fuzzy Systems and Knowledge Discovery (ICNC-FSKD) 2019
    • Related Report
      2019 Research-status Report
    • Int'l Joint Research

URL: 

Published: 2019-04-18   Modified: 2024-01-30  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi