Machine Learning under Changing Environments

不断变化的环境下的机器学习

• 批准号: 20680007
• 负责人: SUGIYAMA Masashi
• 金额: $ 16.31万
• 依托单位: Tokyo Institute of Technology
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Young Scientists (A)
• 财政年份: 2008
• 资助国家: 日本
• 起止时间: 2008 至 2011
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-20680007/
• 关键词: 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; 非定常環境適応; 共変量シフト; 重点サンプリング; 重要度重み; ロボット制御; 強化学習; 政策反復法; 政策探索法; 非定常環境; 密度比推定

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.

现有的机器学习研究已经基于数据生成环境不随时间变化的假设开发了理论和算法。然而，最近的应用并不满足这种平稳性假设。因此，在这个项目中，我们开发了应对不断变化的环境的基础理论和实用算法。我们进一步将开发的算法应用于各种实际应用，包括机器人、图像识别、大脑信号分析、语音识别和自然语言处理。

项目成果

Reward-Weighted Regression with Sample Reuse for Direct Policy Search in Reinforcement Learning

• DOI: 10.1162/neco_a_00199
• 发表时间: 2011-11
• 期刊: Neural Computation
• 影响因子: 2.9
• 作者: Hirotaka Hachiya;Jan Peters;Masashi Sugiyama

A Framework of Adaptive Brain Computer Interfaces

• DOI: 10.1109/bmei.2009.5305646
• 发表时间: 2009-10
• 期刊: 2009 2nd International Conference on Biomedical Engineering and Informatics
• 作者: Yan Li;Y. Koike;Masashi Sugiyama

Dataset Shift in Machine Learning

• DOI: 10.7551/mitpress/9780262170055.001.0001
• 发表时间: 2009-02
• 期刊: Sustainability
• 影响因子: 3.9
• 作者: Joaquin Quionero-Candela;Masashi Sugiyama;A. Schwaighofer;Neil D. Lawrence

Semi-supervised speaker identification under covariate shift

• DOI: 10.1016/j.sigpro.2009.06.001
• 发表时间: 2010-08
• 期刊: Signal Process.
• 作者: M. Yamada;Masashi Sugiyama;T. Matsui

Lighting condition adaptation for perceived age estimation

用于感知年龄估计的照明条件适应

• 发表时间: 2011
• 期刊: Transactions on Information and Systems
• 作者: Ueki;K.;Sugiyama;M.;& Ihara;Y.
• 通讯作者: Y.