In this paper, we propose a novel acoustic model training method which is suitable for speaker adaptation in speech recognition. Our method is based on feature generation from a small amount of speakers' data. For decades, speaker adaptation methods have been widely used. Such adaptation methods need some amount of adaptation data and if the data is not sufficient, speech recognition performance degrade significantly. If the seed models to be adapted to a specific speaker can widely cover more speakers, speaker adaptation can perform robustly. To make such robust seed models, we adopt inverse maximum likelihood linear regression (MLLR) transformation-based feature generation, and then train our seed models using these features. First we obtain MLLR transformation matrices from a limited number of existing speakers. Then we extract the bases of the MLLR transformation matrices using PCA. The distribution of the weight parameters to express the MLLR transformation matrices for the existing speakers is estimated. Next we generate pseudo-speaker MLLR transformations by sampling the weight parameters from the distribution, and apply the inverse of the transformation to the normalized existing speaker features to generate the pseudo-speakers' features. Finally, using these features, we train the acoustic seed models. Using this seed models, we obtained better speaker adaptation results than using simply environmentally adapted models.

The author proposes a self-routing fault-tolerant switching network for asynchronous transfer mode (ATM) switching systems. The network has many subswitches to enhance the fault tolerance of the conventional multistage interconnection network which only has a unique path. The subswitches provide large numbers of alternative paths between switching stages and allow the network to tolerate multiple paths. The routing algorithm is quite simple. The paths can also be used to route cells under the condition that internal cell contentions occur in switching elements. A reliability analysis shows a quantitative measurement of the improvement in fault tolerance as compared with previously presented fault-tolerant networks. A performance analysis and simulation results show that the proposed network has a high level of maximum throughput. In addition, that level of throughput is maintained with reasonable cell delay even though the number of faulty components increases in the network.>