Tuning of Acoustic Modeling and Adaptation Technique for a Real Speech Recognition Task

At the beginning, we had started to develop a Czech telephone acoustic model by evaluating various Kaldi recipes. We had a 500-h Czech telephone Switchboard-like corpus. We had selected the Time-Delay Neural Network (TDNN) model variant “d” with the i-vector adaptation as the best performing model on the held-out set from the corpus. The TDNN architecture with an asymmetric time-delay window also fulfilled our real-time application constrain. However, we were wondering why the model totally failed on a real call center task. The main problem was in the i-vector estimation procedure. The training data are split into short utterances. In the recipe, 2-utterance pseudospeakers are made and i-vectors are evaluated for them. However, the real call center utterances are much longer, in order of several minutes or even more. The TDNN model was trained from i-vectors that did not match the test ones. We propose two ways how to normalize statistics used for the i-vector estimation. The test data i-vectors with the normalization are better compatible with the training data i-vectors. In the paper, we also discuss various additional ways of improving the model accuracy on the out-of-domain real task including using LSTM based models.