Machine Learning under Changing Environments

2011 Fiscal Year Final Research Report

• Project/Area Number: 20680007
• Research Category: Grant-in-Aid for Young Scientists (A)
• Allocation Type: Single-year Grants
• Research Field: Intelligent informatics
• Research Institution: Tokyo Institute of Technology
• Principal Investigator: SUGIYAMA Masashi 東京工業大学, 大学院・情報理工学研究科, 准教授 (90334515)
• Project Period (FY): 2008 – 2011
• Keywords: 学習と知識獲得 / 知能情報処理 / 機械学習 / データマイニング

Research Abstract

Existing machine learning research has developed theories and algorithms based on the assumption that data generating environments do not change over time. However, recent applications do not satisfy such a stationarity assumption. In this project, we therefore developed fundamental theories and practical algorithms for coping with changing environments. We further applied the developed algorithms to various real-world applications including robotics, image recognition, brain signal analysis, speech recognition, and natural language processing.

Research Products

• [Journal Article] Lighting condition adaptation for perceived age estimation (2011)
  • Author(s): Ueki, K., Sugiyama, M., & Ihara, Y.
  • Journal Title: Transactions on Information and Systems Volume: vol.E94-D, no.2 Pages: 392-395
  • Peer Reviewed
• [Journal Article] Reward weighted regression with sample reuse for direct policy search in reinforcement learning (2011)
  • Author(s): Hachiya, H., Peters, J., & Sugiyama M.
  • Journal Title: Neural Computation Volume: vol.23, no.11 Pages: 2798-2832
  • Peer Reviewed
• [Journal Article] Dimensionality reduction for density ratio estimation in high-dimensional spaces (2010)
  • Author(s): Sugiyama, M., Kawanabe, M., & Chui, P. L.
  • Journal Title: Neural Networks Volume: vol.23, no.1 Pages: 44-59
  • Peer Reviewed
• [Journal Article] Semi-supervised speaker identification under covariate shift (2010)
  • Author(s): Yamada, M., Sugiyama, M., & Matsui, T.
  • Journal Title: Signal Processing Volume: vol.90, no.8 Pages: 2353-2361
  • Peer Reviewed
• [Journal Article] Efficient exploration through active learning for value function approximation in reinforcement learning (2010)
  • Author(s): Akiyama, T., Hachiya, H., & Sugiyama, M.
  • Journal Title: Neural Networks Volume: vol.23, no.5 Pages: 639-648
  • Peer Reviewed
• [Journal Article] Application of covariate shift adaptation techniques in brain computer interfaces (2010)
  • Author(s): Li, Y., Kambara, H., Koike, Y., & Sugiyama, M.
  • Journal Title: IEEE Transactions on Biomedical Engineering Volume: vol.57, no.6 Pages: 1318-1324
  • Peer Reviewed
• [Journal Article] Direct density ratio estimation for large-scale covariate shift adaptation (2009)
  • Author(s): Tsuboi, Y., Kashima, H., Hido, S., Bickel, S., & Sugiyama, M.
  • Journal Title: Journal of Information Processing Volume: vol.17 Pages: 138-155
• [Journal Article] Pool-based active learning in approximate linear regression (2009)
  • Author(s): ugiyama, M. & Nakajima, S
  • Journal Title: Machine Learning Volume: vol.75, no.3 Pages: 249-274
  • Peer Reviewed
• [Journal Article] A least-squares approach to direct importance estimation (2009)
  • Author(s): Kanamori, T., Hido, S., & Sugiyama, M.
  • Journal Title: Journal of Machine Learning Research Volume: vol.10 Pages: 1391-1445
  • Peer Reviewed
• [Journal Article] A density-ratio framework for statistical data processing (2009)
  • Author(s): Sugiyama, M., Kanamori, T., Suzuki, T., Hido, S., Sese, J., Takeuchi, I., & Wang, L.
  • Journal Title: IPSJ Transactions on Computer Vision and Applications Volume: vol.1 Pages: 183-208
  • Peer Reviewed
• [Journal Article] Direct importance estimation with Gaussian mixture models (2009)
  • Author(s): Yamada, M. & Sugiyama, M
  • Journal Title: IEICE Transactions on Information and Systems Volume: vol.E92-D, no.10 Pages: 2159-2162
  • Peer Reviewed
• [Journal Article] Adaptive importance sampling for value function approximation in off-policy reinforcement learning (2009)
  • Author(s): Hachiya, H., Akiyama, T., Sugiyama, M., & Peters, J.
  • Journal Title: Neural Networks Volume: vol.22, no.10 Pages: 1399-1410
  • Peer Reviewed
• [Journal Article] Direct importance estimation for covariate shift adaptation (2008)
  • Author(s): Sugiyama, M., Suzuki, T., Nakajima, S., Kashima, H., von Bunau, P., & Kawanabe, M.
  • Journal Title: Annals of the Institute of Statistical Mathematics Volume: vol.60, no.4 Pages: 699-746
  • Peer Reviewed
• [Journal Article] A batch ensemble approach to active learning with model selection (2008)
  • Author(s): Sugiyama, M. & Rubens, N.
  • Journal Title: Neural Networks Volume: vol.21, no.9 Pages: 1278-1286
  • Peer Reviewed
• [Presentation] Perceived age estimation under lighting condition change by covariate shift adaptation (2010)
  • Author(s): Ueki, K., Sugiyama, M., & Ihara, Y.
  • Organizer: In Proceedings of 20th International Conference on Pattern Recognition(ICPR2010)
  • Place of Presentation: Istanbul, Turkey
  • Year and Date: 20100823-26
• [Presentation] Automatic audio tagging using covariate shift adaptation (2010)
  • Author(s): Wichern, G., Yamada, M., Thornburg, H., Sugiyama, M., & Spanias, A.
  • Organizer: In Proceedings of IEEE International Conference on Acoustics, Speech, and Signal Processing(ICASSP2010)
  • Place of Presentation: Dallas, Texas, USA
  • Year and Date: 20100314-19
• [Presentation] A framework of adaptive brain computer interfaces (2009)
  • Author(s): Li, Y., Koike, Y., & Sugiyama, M.
  • Organizer: In Proceedings of the 2nd International Conference on BioMedical Engineering and Informatics(BMEI09)
  • Place of Presentation: Tianjin, China
  • Year and Date: 20091017-19
• [Presentation] Active policy iteration : Efficient exploration through active learning for value function approximation in reinforcement learning (2009)
  • Author(s): Akiyama, T., Hachiya, H., & Sugiyama, M
  • Organizer: In Proceedings of the Twenty-First International Joint Conference on Artificial Intelligence(IJCAI2009)
  • Place of Presentation: Pasadena, California, USA
  • Year and Date: 20090711-17
• [Presentation] Covariate shift adaptation for semi-supervised speaker identification (2009)
  • Author(s): Yamada, M., Sugiyama, M., & Matsui, T.
  • Organizer: In Proceedings of IEEE International Conference on Acoustics, Speech, and Signal Processing(ICASSP2009)
  • Place of Presentation: Taipei, Taiwan
  • Year and Date: 20090419-24
• [Presentation] Efficient sample reuse in EM-based policy search (2009)
  • Author(s): Hachiya, H., Peters, J., & Sugiyama, M.
  • Organizer: Presented at the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases(ECML-PKDD2009)
  • Place of Presentation: Berlin, Springer, Bled, Slovenia
  • Year and Date: 20090000
• [Presentation] Pool-based agnostic experiment design in linear regression (2008)
  • Author(s): Sugiyama, M. & Nakajima, S
  • Organizer: Presented at the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases(ECML-PKDD2008)
  • Place of Presentation: Berlin, Springer, Antwerp, Belgium
  • Year and Date: 20080915-19
• [Presentation] Adaptive importance sampling with automatic model selection in value function approximation (2008)
  • Author(s): Hachiya, H., Akiyama, T., Sugiyama, M., & Peters, J
  • Organizer: Proceedings of the Twenty-Third AAAI Conference on Artificial Intelligence(AAAI2008)
  • Place of Presentation: Chicago, Illinois, USA
  • Year and Date: 20080713-17
• [Presentation] Active learning with model selection in linear regression (2008)
  • Author(s): Sugiyama, M. & Rubens, N
  • Organizer: Proceedings of the Eighth SIAM International Conference on Data Mining(SDM2008)
  • Place of Presentation: Atlanta, Georgia, USA
  • Year and Date: 20080424-26
• [Presentation] Direct density ratio estimation for large-scale covariate shift adaptation (2008)
  • Author(s): Tsuboi, Y., Kashima, H., Hido, S., Bickel, S., & Sugiyama, M.
  • Organizer: Proceedings of the Eighth SIAM International Conference on Data Mining(SDM2008)
  • Place of Presentation: Atlanta, Georgia, USA
  • Year and Date: 20080424-26
• [Presentation] Efficient direct density ratio estimation for non-stationarity adaptation and outlier detection (2008)
  • Author(s): Kanamori, T., Hido, S., & Sugiyama, M.
  • Organizer: In D. Koller, D. Schuurmans, Y. Bengio, and L. Botton(Eds.), Advances in Neural Information Processing Systems 21,(Presented at Neural Information Processing Systems(NIPS2008), Vancouver, British Columbia
  • Place of Presentation: Cambridge, MA, MIT Press, Canada, Dec
  • Year and Date: 2008-08-13
• [Book] Machine Learning in Non-Stationary Environments : Introduction to Covariate Shift Adaptation (2012)
  • Author(s): Sugiyama, M. & Kawanabe, M
  • Total Pages: 308
  • Publisher: MIT Press, Cambridge, MA, USA
• [Book] Dataset Shift in Machine Learning (2009)
  • Author(s): Quinonero-Candela, J., Sugiyama, M., Schwaighofer, A., & Lawrence, N. D.
  • Total Pages: 248
  • Publisher: MIT Press, Cambridge, MA, USA
• [Remarks]
  • URL: http://sugiyama-www.cs.titech.ac.jp/~sugi/