Confidence Measures for Tandem Connectionist Feature Extraction
نویسندگان
چکیده
This report proposes and compares a number of tandem-like feature extraction schemes. The proposed schemes use relative phone posteriors as confidence measures estimated from the MLP outputs directly or using Gamma function. The analysis of variances shows that the proposed tandem-like features discriminate better between phone classes than the conventional tandem features. But these capabilities are lost when the complexity of the model (number of gaussians) increases. Evaluations are conducted on TIMIT database.
منابع مشابه
Tandem connectionist feature extraction for conventional HMM systems
Hidden Markov model speech recognition systems typically use Gaussian mixture models to estimate the distributions of decorrelated acoustic feature vectors that correspond to individual subword units. By contrast, hybrid connectionist-HMM systems use discriminatively-trained neural networks to estimate the probability distribution among subword units given the acoustic observations. In this wor...
متن کاملA Geometric View of Similarity Measures in Data Mining
The main objective of data mining is to acquire information from a set of data for prospect applications using a measure. The concerning issue is that one often has to deal with large scale data. Several dimensionality reduction techniques like various feature extraction methods have been developed to resolve the issue. However, the geometric view of the applied measure, as an additional consid...
متن کاملInvestigations Into Tandem Features
This report proposes and evaluates a number of tandem feature extraction schemes. The proposed schemes use confidence measures estimated from the MLP outputs to derive tandem-like features. The analysis of variance shows that the proposed features discriminate better between phone classes than conventional tandem features. But they become less discriminant as the HMM model become more complex i...
متن کاملConnectionist speech recognition of Broadcast News
This paper describes connectionist techniques for recognition of Broadcast News. The fundamental difference between connectionist systems and more conventional mixture-of-Gaussian systems is that connectionist models directly estimate posterior probabilities as opposed to likelihoods. Access to posterior probabilities has enabled us to develop a number of novel approaches to confidence estimati...
متن کاملA Generic Connectionist-Based Method for On-Line Feature Selection and Modelling with a Case Study of Gene Expression Data Analysis
The paper presents a novel generic method for on-line feature extraction from an incrementally trained connectionist system. The method is applied on a case study problem of identifying genes related to classes of diseases, in particular 14 types of cancer. The method is based on the evolving connectionist systems ECOS paradigm. The analysis of the discovered features through the application of...
متن کامل