| 講演抄録／キーワード |
| 講演名 |
2020-11-13 13:30
The Optimum Prediction Theory of the Matrix Input Operator by Additive Operator Filter bank
Takuro Kida（Titech）・○Yuichi Kida（Ohu Univ.） |
| 抄録 |
（和） |
(事前公開アブストラクト) With respect to a matrix-filter bank that the matrix analysis-filter bank ${bf H}$ and the matrix sampling-filter bank ${bf S}$ are given, it is accomplished to present the optimum matrix synthesis-filter bank ${bf Z}$ that minimizes all the worst-case measures of matrix-error-signals ${bf E}(omega)={bf F}(omega)-{bf Y}(omega)$ between the input matrix-signals ${bf F}(omega)$ and the output matrix-signals ${bf Y(omega)}$ of the matrix-filter bank, at the same time. As the direct conclusion of this result, it is shown that there exists the optimum and linear approximation system based on the scanned data of given multi-dimensional past knowledge ${bf f}({bf x}(t))$ if a given nonlinear deep learning system uses the equivalent data as these data finally. |
| （英） |
With respect to a matrix-filter bank that the matrix analysis-filter bank ${bf H}$ and the matrix sampling-filter bank ${bf S}$ are given, it is accomplished to present the optimum matrix synthesis-filter bank ${bf Z}$ that minimizes all the worst-case measures of matrix-error-signals ${bf E}(omega)={bf F}(omega)-{bf Y}(omega)$ between the input matrix-signals ${bf F}(omega)$ and the output matrix-signals ${bf Y(omega)}$ of the matrix-filter bank, at the same time. As the direct conclusion of this result, it is shown that there exists the optimum and linear approximation system based on the scanned data of given multi-dimensional past knowledge ${bf f}({bf x}(t))$ if a given nonlinear deep learning system uses the equivalent data as these data finally. |
| キーワード |
（和） |
/ / / / / / / |
| （英） |
signal approximation / / / / / / / |
| 文献情報 |
映情学技報, vol. 44, no. 28, BCT2020-67, pp. 41-46, 2020年11月. |
| 資料番号 |
BCT2020-67 |
| 発行日 |
2020-11-05 (BCT) |
| ISSN |
Print edition: ISSN 1342-6893 Online edition: ISSN 2424-1970 |
| PDFダウンロード |
|