時間依存型密度比推定による高次元変化検知

• Project/Area Number: 13J03189
• Research Category: Grant-in-Aid for JSPS Fellows
• Allocation Type: Single-year Grants
• Section: 国内
• Research Field: Statistical science
• Research Institution: Tokyo Institute of Technology
• Principal Investigator: 柳 松 東京工業大学, 情報理工学研究科, 特別研究員(PD)
• Project Period (FY): 2013-04-01 – 2015-03-31
• Project Status: Completed (Fiscal Year 2014)
• Budget Amount: ¥2,300,000 (Direct Cost: ¥2,000,000、Indirect Cost: ¥300,000); Fiscal Year 2014: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000); Fiscal Year 2013: ¥1,000,000 (Direct Cost: ¥1,000,000)
• Keywords: MACHINE LEARNING / DENSITY RATIO ESTIMATION / CHANGE DETECTION / GRAPHICAL MODEL / Change Detection / Markov Network / Density Ratio Estimation / Machine Learning

Outline of Annual Research Achievements

In this research, we have mainly focused on the problem of change detection in high-dimensional time-series, and more generally, on dependent dataset.
We have spent most our time in the first year developing a methodology for change detection in Graphical Models, where we input two sets of data drawn from two different distributions with different interactions among random variables. In such methodology, we assumed that the changes between two stages are subtle and most of the interactions remain unchanged. As a consequence of this assumption, the sparsity is assumed in our statistical model and via Density Ratio Estimation method, the sparse changes between two Graphical Models are learned.
In this year, we conducted a theoretical study for such methodology and give statistical guarantees of the superiority of the proposed change detection method. Specifically, we give the sufficient conditions that our change detection method works, in terms of sample complexity against the increasing number of changed edges and dimensions.
Moreover, the above methodology itself is for learning changes from two sets of data. However, It is nature to ask that is it possible to apply such powerful method to the learning of the Graphical Model structure itself? Our new idea is simply learning the difference between the join distribution and the product of marginal distributions.
To sum up, we have not only finished the research promised in the proposal, but also investigated a new (and important) application of the proposed method and theory.

Research Progress Status

26年度が最終年度であるため、記入しない。

Strategy for Future Research Activity

26年度が最終年度であるため、記入しない。

Research Products

• [Journal Article] Direct Learning of Sparse Changes in Markov Networks by Density Ratio Estimation (2014)
  • Author(s): Liu, S., Quinn, J. A., Gutmann, M. U., Sugiyama, M.
  • Journal Title: Journal Neural Computation Volume: (印刷中)
  • Peer Reviewed
• [Journal Article] Direct Divergence Approximation between Probability Distributions and its Applications in Machine Learning. (2013)
  • Author(s): Sugiyama, M., Liu, S., du Plessis, M. C., Yamanaka, M., Yamad a, M., Suzuki, T., & Kanamori, T.
  • Journal Title: Journal of Computing Science and Engineering Volume: Vol.7no.2 Issue: 2 Pages: 99-111
  • DOI: 10.5626/jcse.2013.7.2.99
  • Peer Reviewed
• [Presentation] Support consistency of direct sparse-change learning in Markov Support consistency of direct sparse-change learning in Markov networksSupport consistency of direct sparse-change learning in Markov networks (2015)
  • Author(s): Liu, S. and Suzuki, T. and Sugiyama, M.
  • Organizer: Twenty-Ninth AAAI Conference on Artificial Intelligence (AAAI2015)
  • Place of Presentation: Hyatt Regency Austin, 208 Barton Springs Road, Austin, US,
  • Year and Date: 2015-01-26 – 2015-01-30
• [Presentation] Support consistency of direct sparse-change learning in Markov Support consistency of direct sparse-change learning in Markov networksSupport consistency of direct sparse-change learning in Markov networks (2014)
  • Author(s): Liu, S. and Suzuki, T. and Sugiyama, M.
  • Organizer: 第17回情報論的学習理論ワークショップ(IBIS2014)
  • Place of Presentation: Nagoya University, Nagoya, Japan
  • Year and Date: 2014-11-17
• [Presentation] Bias Reduction and Metric Learng for Nearest-neighbor Estimation of Kullback-Leibler Divergence. (2014)
  • Author(s): Yung-Kyun Noh
  • Organizer: Seventeenth International Conference on Artificial Intelligence and Statistics (AISTATS2014)
  • Place of Presentation: Grand Hotel Reykjavik(発表確定)
  • Year and Date: 2014-04-24
• [Presentation] Direct Learning of Sparse Changes in Markov Networks by Density Ratio Estimation (2013)
  • Author(s): Song Liu
  • Organizer: European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases
  • Place of Presentation: Congress Centre U HájkÜ, Prague, Czech.
  • Year and Date: 2013-09-27
• [Presentation] Direct Learning of Sparse Changes in Markov Networks by Density Ratio Estimation (2013)
  • Author(s): Song Liu
  • Organizer: 情報論的学習理論と機械学習研究会(IBISML)
  • Place of Presentation: 早稲田大学 西早稲田キャンパス
  • Year and Date: 2013-07-18
• [Remarks]
  • URL: http://sugiyama-www.cs.titech.ac.jp/~song/