Speed up of recurrent neural network language models with sentence independent subsampling stochastic gradient descent.
Yangyang ShiMei-Yuh HwangKaisheng YaoMartha A. LarsonPublished in: INTERSPEECH (2013)
Keyphrases
- language model
- recurrent neural networks
- stochastic gradient descent
- language modeling
- loss function
- n gram
- neural network
- probabilistic model
- matrix factorization
- information retrieval
- least squares
- query expansion
- retrieval model
- feed forward
- step size
- random forests
- artificial neural networks
- weight vector
- regularization parameter
- multiple kernel learning
- support vector machine
- vector space model
- relevance model
- text mining
- ensemble methods
- lower bound