Differences

This shows you the differences between two versions of the page.

Link to this comparison view

convolution_recurrent_network [2016/11/27 18:46] (current)
Line 1: Line 1:
 +====== Convolution LSTM Network ======
  
 +
 +https://​openreview.net/​pdf?​id=H1zJ-v5xl ​ QUASI-RECURRENT NEURAL NETWORKS
 +
 + We
 +introduce quasi-recurrent neural networks (QRNNs), an approach to neural sequence
 +modeling that alternates convolutional layers, which apply in parallel
 +across timesteps, and a minimalist recurrent pooling function that applies in parallel
 +across channels. Despite lacking trainable recurrent layers, stacked QRNNs
 +have better predictive accuracy than stacked LSTMs of the same hidden size. Due
 +to their increased parallelism,​ they are up to 16 times faster at train and test time.
 +
 +{{::​quasi.png|}}
 +
 +{{:​wiki:​quasi2.png}}
 +
 +http://​arxiv.org/​pdf/​1606.05262v1.pdf ​ Convolutional Residual Memory Networks
 +
 +{{https://​ai2-s2-public.s3.amazonaws.com/​figures/​2016-11-01/​003dba0ca2abf582ecf42a0ee88d6b831f1b24aa/​3-Figure2-1.png?​600x600}}