Search

14 Sep 2010: All three test sets, A, B and C,were used for evaluating the schemes. ... This work used three of the test sets containing digit. sequences (phone numbers).

20 Oct 2010: 106. 7.6 Variation of posterior log probability of the correct alignment and AER ontraining and test data sets as training progresses, for Arabic to English. ... 107. 7.7 Variation of posterior log probability of the correct alignment and AER ontraining

25 Oct 2010: The mixture model makes predictionsfor arbitrary new test points and typically has a relativelysmall number of parameters. ... where NG is the number ofground truth images in the test data set).

19 Nov 2010: The two forms of modified tail distribution modelling were inves-tigated on the 1998 Hub4 partitioned evaluation test set. ... The test was performed on a subset of the 1997 par-titioned evaluation that was used for development [4].

8 Sep 2010: Test set A was used as the development set for tuningparameters for all systems, such as the penalty factor Cfor SVMs. ... The integratedtraining yielded a 9% reduction in error rate. The full results for all three test sets are shown in TableII.

6 Oct 2010: Lattice Rescoring Methods forStatistical Machine Translation. Graeme Blackwood. Cambridge University Engineering Departmentand. Clare College. Dissertation submitted to the University of Cambridgefor the degree of Doctor of Philosophy.

14 Sep 2010: In addition to background additive noise convolutional distortion wasadded to test set C. ... Table 3VTS (γ = 1) and SVM rescoring performance WER (%) tests sets A, B, C (ϵ = 2)using HTK features, SVMs trained on test set A N2-N4 10-20dB SNR.

5 Jan 2010: 22] J. Li, M. Siniscalchi, and C-H. Lee, “Approximate test risk minimization theough soft margin training,” in ICASSP, 2007.

25 Feb 2010: 76. 4.6 Score-normalisation. 784.6.1 Zero normalisation. 784.6.2 Test normalisation. 794.6.3 Adaptive T-norm. ... optimised on either the trainingset (LR-trn) or test set (LR).

10 Jun 2010: 60. 3.5 The recognition setup for test data using the ML-SAT system. ... estimated for the test data, others are estimated during the training of the.