Note: Your custom schedule will not be saved unless you create a new account or login to an existing account.
| Paper: | MLSP-P6.4 |
| Session: | Learning Theory and Models |
| Time: | Thursday, May 20, 15:30 - 17:30 |
| Presentation: |
Poster |
| Topic: |
Machine Learning for Signal Processing: Learning Theory and Modeling |
| Title: |
A NEW WAY OF PCA: INTEGRATED-SQUARED-ERROR AND EM ALGORITHMS |
| Authors: |
Jong-Hoon Ahn; POSTECH | | |
| | Seungjin Choi; POSTECH | | |
| | Jong-Hoon Oh; POSTECH | | |
| Abstract: |
Minimization of reconstruction error (squared-error) leads to principal subspace analysis (PSA) which estimates scaled and rotated principal axes of a set of observed data. In this paper we introduce a new alternative error, so called, integratedsquared-error, the minimization of which determines the exact principal axes (without rotational ambiguity) of a set of observed data. We consider the properties of the integrated-squared-error in the framework of coupled generative model, giving efficient EM algorithms for integrated-squared-error minimization. We revisit the generalized Hebbian algorithm (GHA) and show that it emerges from integrated-squared-error minimization in a single-layer linear feedforward neural network. |
| |
| Back | |