On-line Learning of Unlearnable True Teacher through Mobile Ensemble Teachers
-
- Hirama Takeshi
- Department of Basic Science, University of Tokyo
-
- Hukushima Koji
- Department of Basic Science, University of Tokyo
この論文をさがす
抄録
The on-line learning of a hierarchical learning model is studied by a method based on statistical mechanics. In our model, a student of a simple perceptron learns from not a true teacher directly, but ensemble teachers who learn from a true teacher with a perceptron learning rule. Since the true teacher and ensemble teachers are expressed as nonmonotonic and simple perceptrons, respectively, the ensemble teachers go around the unlearnable true teacher with the distance between them fixed in an asymptotic steady state. The generalization performance of the student is shown to exceed that of the ensemble teachers in a transient state, as was shown in similar ensemble-teachers models. Furthermore, it is found that moving the ensemble teachers even in the steady state, in contrast to the fixed ensemble teachers, is efficient for the performance of the student.
収録刊行物
-
- Journal of the Physical Society of Japan
-
Journal of the Physical Society of Japan 77 (9), 094801-094801, 2008
一般社団法人 日本物理学会
- Tweet
詳細情報 詳細情報について
-
- CRID
- 1390001204197392768
-
- NII論文ID
- 130005437088
- 110006933501
- 210000107437
-
- NII書誌ID
- AA00704814
-
- BIBCODE
- 2008JPSJ...77i4801H
-
- ISSN
- 13474073
- 00319015
-
- NDL書誌ID
- 9646772
-
- 本文言語コード
- en
-
- データソース種別
-
- JaLC
- NDL
- Crossref
- CiNii Articles
-
- 抄録ライセンスフラグ
- 使用不可