Page 75 |
Save page Remove page | Previous | 75 of 196 | Next |
|
small (250x250 max)
medium (500x500 max)
Large (1000x1000 max)
Extra Large
large ( > 500x500)
Full Resolution
All (PDF)
|
This page
All
|
ˆ FX(x) = ˆ PN X ((−1, x]) 18. In the k-phase of this algorithm the criterion checks every atom A of the current partition Qk by evaluating the empirical MI gain obtained by par-titioning A with a product structure adaptively generated with the marginal distribution of the training points in A, denoted by Q(A)19. If this gain is above a critical threshold the algorithm splits the atom to upgrade the partition by Qk+1 = (Qk\{A})[Q(A). and continues in this region applying recursively the aforementioned splitting criterion. But in the negative case, the algorithm stops the refinement of this region under the assump-tion that condition to the event X 2 A, X and Y can be considered almost independent, i.e., ˆIQ2(A) N (X; Y |X 2 A) < ) I(X; Y |X 2 A) 0. Furthermore to control estima-tion error, we introduce a threshold in the splitting rule to control the minimum number of training points associated with A, for having a good representation of the joint distri-bution between X and Y in this target region. The pseudo-code is presented in Fig. 3.3, which considers the following set of parameters: • (s, r) 2 N2, s > r: used for generating product refinements, see Fig. 3.3 for details, • > 0: threshold for the MI gain, • Nc 2 N: minimum number of points for probability estimation. Finally, in our target problem we have X1, X2, X3 and Y and we need to estimate = I(X1,X2; Y |X3), with the the empirical data {(xi 1, xi 2, xi 3, yi) : i = 1, ..,N}. Then the non-parametric estimation goes as follows: 1) Use Darbellay-Vajda algorithm to construct partition QN1 ,2 for (X1,X2) using {xi 1, xi 2, yi : i = 1, ..,N}. 18We consider X as a scalar random variable, however the construction extents naturally for the finite dimensional scenario. 19The marginal MI gain can be expressed by ˆ PN X (A) · ˆIQ2(A) N (X; Y |X 2 A). 62
Object Description
Title | On optimal signal representation for statistical learning and pattern recognition |
Author | Silva, Jorge |
Author email | jorgesil@usc.edu; josilva@ing.uchile.cl |
Degree | Doctor of Philosophy |
Document type | Dissertation |
Degree program | Electrical Engineering |
School | Viterbi School of Engineering |
Date defended/completed | 2008-06-23 |
Date submitted | 2008 |
Restricted until | Unrestricted |
Date published | 2008-10-21 |
Advisor (committee chair) | Narayanan, Shrikanth S. |
Advisor (committee member) |
Kuo, C.-C. Jay Ordóñez, Fernando I. |
Abstract | This work presents contributions on two important aspects of the role of signal representation in statistical learning problems, in the context of deriving new methodologies and representations for speech recognition and the estimation of information theoretic quantities.; The first topic focuses on the problem of optimal filter bank selection using Wavelet Packets (WPs) for speech recognition applications. We propose new results to show an estimation-approximation error tradeoff across sequence of embedded representations. These results were used to formulate the minimum probability of error signal representation (MPE-SR) problem as a complexity regularization criterion. Restricting this criterion to the filter bank selection, algorithmic solutions are provided by exploring the dyadic tree-structure of WPs. These solutions are stipulated in terms of a set of conditional independent assumptions for the acoustic observation process, in particular, a Markov tree property across the indexed structure of WPs. In the technical side, this work presents contributions on the extension of minimum cost tree pruning algorithms and their properties to affine tree functionals. For the experimental validation, a phone classification task ratifies the goodness of Wavelet Packets as an analysis scheme for non-stationary time-series processes, and the effectiveness of the MPE-SR to provide cost effective discriminative filter bank solution for pattern recognition.; The second topic addresses the problem of data-dependent partitions for the estimation of mutual information and Kullback-Leibler divergence (KLD). This work proposes general histogram-based estimates considering non-product data-driven partition schemes. The main contribution is the stipulation of sufficient conditions to make these histogram-based constructions strongly consistent for both problems. The sufficient conditions consider combinatorial complexity indicator for partition families and the use of large deviation type of inequalities (Vapnik-Chervonenkis inequalities). On the application side, two emblematic data-dependent constructions are derived from this result, one based on statistically equivalent blocks and the other, on a tree-structured vector quantization scheme. A range of design values was stipulated to guarantee strongly consistency estimates for both framework. Furthermore, experimental results under controlled settings demonstrate the superiority of these data-driven techniques in terms of a bias-variance analysis when compared to conventional product histogram-based and kernel plug-in estimates. |
Keyword | signal representation in statistical learning; Bayes decision theory; basis selection; tree-structured bases and Wavelet packet (WP); complexity regularization; minimum cost tree pruning; family pruning problem; mutual information estimation; divergence estimation; data-dependent partitions; statistical learning theory; concentration inequalities; tree-structured vector quantization. |
Language | English |
Part of collection | University of Southern California dissertations and theses |
Publisher (of the original version) | University of Southern California |
Place of publication (of the original version) | Los Angeles, California |
Publisher (of the digital version) | University of Southern California. Libraries |
Provenance | Electronically uploaded by the author |
Type | texts |
Legacy record ID | usctheses-m1684 |
Contributing entity | University of Southern California |
Rights | Silva, Jorge |
Repository name | Libraries, University of Southern California |
Repository address | Los Angeles, California |
Repository email | cisadmin@lib.usc.edu |
Filename | etd-Silva-2450 |
Archival file | uscthesesreloadpub_Volume32/etd-Silva-2450.pdf |
Description
Title | Page 75 |
Contributing entity | University of Southern California |
Repository email | cisadmin@lib.usc.edu |
Full text | ˆ FX(x) = ˆ PN X ((−1, x]) 18. In the k-phase of this algorithm the criterion checks every atom A of the current partition Qk by evaluating the empirical MI gain obtained by par-titioning A with a product structure adaptively generated with the marginal distribution of the training points in A, denoted by Q(A)19. If this gain is above a critical threshold the algorithm splits the atom to upgrade the partition by Qk+1 = (Qk\{A})[Q(A). and continues in this region applying recursively the aforementioned splitting criterion. But in the negative case, the algorithm stops the refinement of this region under the assump-tion that condition to the event X 2 A, X and Y can be considered almost independent, i.e., ˆIQ2(A) N (X; Y |X 2 A) < ) I(X; Y |X 2 A) 0. Furthermore to control estima-tion error, we introduce a threshold in the splitting rule to control the minimum number of training points associated with A, for having a good representation of the joint distri-bution between X and Y in this target region. The pseudo-code is presented in Fig. 3.3, which considers the following set of parameters: • (s, r) 2 N2, s > r: used for generating product refinements, see Fig. 3.3 for details, • > 0: threshold for the MI gain, • Nc 2 N: minimum number of points for probability estimation. Finally, in our target problem we have X1, X2, X3 and Y and we need to estimate = I(X1,X2; Y |X3), with the the empirical data {(xi 1, xi 2, xi 3, yi) : i = 1, ..,N}. Then the non-parametric estimation goes as follows: 1) Use Darbellay-Vajda algorithm to construct partition QN1 ,2 for (X1,X2) using {xi 1, xi 2, yi : i = 1, ..,N}. 18We consider X as a scalar random variable, however the construction extents naturally for the finite dimensional scenario. 19The marginal MI gain can be expressed by ˆ PN X (A) · ˆIQ2(A) N (X; Y |X 2 A). 62 |