バイアスを考慮した逐次型能動学習 Incremental Active Learning in Consideration of Bias

    • 杉山 将 Sugiyama Masashi
    • 東京工業大学大学院情報理工学研究科計算工学専攻 Department of Computer Science, Tokyo Institute of Technology
    • 小川 英光 Ogawa Hidemitsu
    • 東京工業大学大学院情報理工学研究科計算工学専攻 Department of Computer Science, Tokyo Institute of Technology

Abstract

教師付き学習における能動学習とは,高い汎化能力が獲得できるように入力信号を最適化する問題である.これまでに提案されてきた多くの能動学習法では,学習結果のバリアンスが最小になるように入力信号を設計している.即ち,学習結果のバイアスはゼロか無視できるくらい小さいと仮定している.本論文では,バイアスを考慮した能動学習法を与え,その有効性を計算機実験で示す.

The problem of designing input signals for optimal generalization in supervised learning is called active learning. In many active learning methods devised so far, the sampling location minimizing the variance of the learning results is selected. This implies that the bias of the learning results is assumed to be zero or small enough to be neglected. In this paper, we propose an active learning method with the bias reduction. The effectiveness of the proposed method is demonstrated through computer simulations.

Journal

IEICE technical report. Neurocomputing   [List of Volumes]

IEICE technical report. Neurocomputing 99(473), 15-22, 1999-11-26  [Table of Contents]

The Institute of Electronics, Information and Communication Engineers

References:  25

You must have a user ID to see the references.If you already have a user ID, please click "Login" to access the info.New users can click "Sign Up" to register for an user ID.

Preview

Preview

Codes

  • NII Article ID (NAID) :
    110003234376
  • NII NACSIS-CAT ID (NCID) :
    AN10091178
  • Text Lang :
    ENG
  • Article Type :
    ART
  • Databases :
    CJP  NII-ELS 

Share