Structural Maximum A POSTERIORI Linear Regression for Fast HMM Adaptation

Transformation-based model adaptation techniques like maximum likelihood linear regression (MLLR) rely on an accurate selection of the number of transformations for a given amount of adaptation data. If too many transformations are used, the transformation parameters may be poorly estimated, can overfit the adaptation data, and offer poor generalization. On the other hand, if the number of transformations is too small, the adapted models can only provide a moderate improvement over the baseline models. AN adaptation approach should therefore be flexible in order to estimate reliability a large number of transformations when the amount of adaptation data is large, and a small number of transformations when only a few adaptation utterance are available. In this work, we show taht a significant improvedment can be obtained over MLLR with dynamic regression classes, first by replacing the maximum likelihood estimation criterion by a maximum a posterior criterion then by introducing a tree-straucuire for a th prior densities of the transformations. The effectiveness of the proposed approach is ilklustrated ont he spoke3 test st of the WSJ taks. Using the smae regression classes as MLLR, it is shown that the proposed approach reduces the risk fo overfitting and exploit the adaptation data much more efficiently then MLLR, leading to a significant reduction of the word error rate with as little as one adaptioatn utterance.