Feature space maximum a posteriori linear regression for adaptation of deep neural networks

21 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Overcoming Catastrophic Forgetting by Incremental Moment Matching (2017, methodology): “yes was used to learn topic models from stream databy Broderick et al. [6].Huang et al. applied sequential Bayesian to adapt a deep neural network to the speciﬁc user in the speech recognition domain [15, 16]. They assigned the layer for the user adaptation and applied MAP estimation to this single ”
Maximum a posteriori adaptation of network parameters in deep models (2015, methodology): “We observed that the histograms for weights of the adapted LHN over speakers are quite like Gaussian, so we assume that the distribution of the weights in Wlhn to be joint Gaussian [26].”
Speaker Adaptive Training of Deep Neural Network Acoustic Models Using I-Vectors (2015, methodology): “[44] estimate the adaptation parameters via maximum a posteriori (MAP) linear regression, which naturally incorporates prior knowledge into the adaptation process.”