On-line learning theory of soft committee machines with correlated hidden units - Steepest gradient descent and natural gradient descent

Masato Inoue, Hyeyoung Park, Masato Okada

研究成果: Article査読

24 被引用数 (Scopus)

抄録

The permutation symmetry of the hidden units in multilayer perceptrons causes the saddle structure and plateaus of the learning dynamics in gradient learning methods. The correlation of the weight vectors of hidden units in a teacher network is thought to affect this saddle structure, resulting in a prolonged learning time, but this mechanism is still unclear. In this paper, we discuss it with regard to soft committee machines and on-line learning using statistical mechanics. Conventional gradient descent needs more time to break the symmetry as the correlation of the teacher weight vectors rises. On the other hand, no plateaus occur with natural gradient descent regardless of the correlation for the limit of a low learning rate. Analytical results support these dynamics around the saddle point.

本文言語English
ページ(範囲)805-810
ページ数6
ジャーナルjournal of the physical society of japan
72
4
DOI
出版ステータスPublished - 2003 4
外部発表はい

ASJC Scopus subject areas

  • 物理学および天文学(全般)

フィンガープリント

「On-line learning theory of soft committee machines with correlated hidden units - Steepest gradient descent and natural gradient descent」の研究トピックを掘り下げます。これらがまとまってユニークなフィンガープリントを構成します。

引用スタイル