Showing 3 results for Transformation
S. Sharifian and S. M. Ahadi,
Volume 23, Issue 2 (1-2005)
Abstract
A variety of methods are used for speaker adaptation in speech recognition. In some techniques, such as MAP estimation, only the models with available training data are updated. Hence, large amounts of training data are required in order to have significant recognition improvements. In some others, such as MLLR, where several general transformations are applied to model clusters, the results are desirable for small training data, but with increasing training data, the performance improvement reaches the saturation lvel. In this paper, a new approach is introduced that makes use of the advantages of both mentioned techniques to improve the recognition rate. Here, the models with available training data are trained using MAP while
for those with insufficient training data, appropriate prior parameters for MAP estimation are found using MLLR. This technique has yielded better performance in comparison to either MAP or MLLR, in a system based on FARSDAT speech corpus.
N. Setoudeh, A. Saidi, A. Shafyei and N. J. Welham,
Volume 25, Issue 1 (7-2006)
Abstract
Anatase-to-rutile phase transformation was studied in milled and unmilled samples. Ball milling was carried out in two types of ball mills, planetary and tumbler, with a ball-to-powder ratio of 40:1 over 2-48 hours. First, the unmilled samples
were heated in the furnace at various temperatures for different periods of time. The results revealed that the anatase-to-rutile transformation completed at 980 after 48 hours. The rate of transformation in milled samples was greatly higher than that of unmilled ones. Activation energy in unmilled samples was about 440 kj/mol. The rate of transformation in the planetary ball mill was higher than that in tumbler mill. In the former, transformation almost finished after 16 hours of milling while in the lattar, it did not finish even after 48 hours. XRD results revealed that the transformation proceeds through an intermediate srilankite phase in all milled samples. However, srilankite was not observed in the unmilled samples.
M. Bagheri, B. Keshtegar,
Volume 37, Issue 1 (9-2018)
Abstract
In this paper, a new method is proposed for fuzzy structural reliability analysis; it considers epistemic uncertainty arising from the statistical ambiguity of random variables. The proposed method, namely, fuzzy dynamic-directional stability transformation method, includes two iterative loops. An internal algorithm performs the reliability analysis using the dynamic-directional stability transformation method and an external algorithm performs the fuzzy analysis by applying the alpha-cut level optimization method based on the genetic algorithm. Implementation of the proposed method, which solves some nonlinear performance functions, indicates the efficiency and robustness of the dynamic-directional stability transformation method, as compared to other first order reliability methods.
