Adaptation of quadratic trajectory segment models for small vocabulary speech recognition

In this paper, we report our recent work on applications of the MAP approach to estimating the time-varying quadratic polynomial Gaussian mean functions in the nonstationary-state or trended HMM. Assuming uncorrelatedness among the quadratic polynomial coefficients in the trended HMM, we have obtained analytical results for the MAP estimates of the time-varying mean and precision parameters. We have implemented a speech recognizer based on these results in speaker adaptation experiments using TI46 corpora. Experimental results show that the quadratic trended HMM always outperforms the standard, stationary-state HMM and that adaptation of quadratic polynomial coefficients only is better than adapting both polynomial coefficients and precision matrices when fewer than four adaptation tokens are used