Neural Network Models

Catherine discusses the evolution from gaussian mixture models to neural networks in acoustic modeling, leading to significant accuracy improvements in speech recognition systems. The acoustic model and language model are statistical, machine-learned models trained on audio and text data, respectively.