Noisy objective functions based on the f-divergence

Markus Nussbaum-Thom; Ralf Schlüter; Vaibhava Goel; Hermann Ney

doi:10.1109/ICASSP.2017.7952572

ICASSP 2017

Conference paper

16 Jun 2017

Noisy objective functions based on the f-divergence

View publication

Abstract

Dropout, the random dropping out of activations according to a specified rate, is a very simple but effective method to avoid over-fitting of deep neural networks to the training data.

Conference paper

Co-occurring directions sketching for approximate matrix multiply

Youssef Mroueh, Etienne Marcheret, et al.

AISTATS 2017

Conference paper

Data augmentation for deep convolutional neural network acoustic modeling

Xiaodong Cui, Vaibhava Goel, et al.

ICASSP 2015

Conference paper

State of the art discriminative training of subspace constrained Gaussian mixture models in big training corpora

Jing Huang, Peder A. Olsen, et al.

ICASSP 2013

Conference paper

Detecting audio-visual synchrony using deep neural networks

Etienne Marcheret, Gerasimos Potamianos, et al.

INTERSPEECH 2015

View all publications

Abstract

Related

Co-occurring directions sketching for approximate matrix multiply

Data augmentation for deep convolutional neural network acoustic modeling

State of the art discriminative training of subspace constrained Gaussian mixture models in big training corpora

Detecting audio-visual synchrony using deep neural networks