Speaker Adaptation with Transformation Matrix Linear Interpolation

A transformation matrix linear interpolation (TMLI) approach for speaker adaptation is proposed. TMLI uses the transformation matrixes produced by MLLR from selected training speakers and the testing speaker. With only 3 adaptation sentences, the performance shows a 12.12% word error rate reduction. As the number of adaptation sentences increases, the performance saturates quickly. To improve the behavior of TMLI for large amounts of adaptation data, the TMLI+MAP method which combines TMLI with MAP technique is proposed. Experimental results show TMLI+MAP achieved better recognition accuracy than MAP and MLLR+MAP for both small and large amounts of adaptation data. Key words speech recognition - speaker adaptation - MLLR - MAP - maximum likelihood model interpolation (MLMI) CLC number TN 912. 34 Foundation item: Supported by the Science and Technology Committee of Shanghai (01JC14033)Biography: XU Xiang-hua (1977-), female, Ph. D. candidate, research direction: large vocabulary continuous Mandarin speech recognition and speaker adaptation

A Combined Speaker Adaptation Method for Mandarin Speech Recognition

A speaker adaptation method that combines transformation matrix linear interpolation with maximum a posteriori (MAP) was proposed. Firstly this method can keep the asymptotical characteristic of MAP. Secondly, as the method uses linear interpolation with several speaker-dependent (SD) transformation matrixes, it can fully use the prior knowledge and keep fast adaptation. The experimental results show that the combined method achieves an 8.24% word error rate reduction with only one adaptation utterance, and keeps asymptotic to the performance of SD model for large amounts of adaptation data.

Maximum Likelihood Estimator for the Proportional Hazards Model with Incomplete Information

Right randomly censored data with incomplete infor-mation are frequently met in practice.Although much study about right randomly censored data has been seen in the proportional hazards model,relatively little is known about the inference of regression parameters for right randomly censored data with in-complete information in such model.In particular,theoretical properties of the maximum likelihood estimator of the regression parameters have not been proven yet in that model.In this paper,we show the consistency and asymptotic normality of the maxi-mum likelihood estimator of unknown regression parameters.

Empirical Likelihood-based Inferences in Varying Coefficient Models with Missing Data

In this paper, we consider the empirical likelihood-based inferences for varying coefficient models Y = X^τα（U） ＋ ε when X are subject to missing at random. Based on the inverse probability-weighted idea, a class of empirical log-likelihood ratios, as well as two maximum empirical likelihood estimators, are developed for α（u）. The resulting statistics are shown to have standard chi-squared or normal distributions asymptotically.Simulation studies are also constructed to illustrate the finite sample properties of the proposed statistics.