[Adapt] [Seminar]Convoluational sequence to sequence learning

[Yizhu Liu]

Hi Adapters, 

In the weekly seminar tomorrow, I'll give a talk on Convolutional sequence to sequence learning. The prevalent approach to sequence to sequence learning maps an input sequence to a variable length output sequence via recurrent neural networks. I will introduce an architecture based entirely on convolutional neural networks. Compared to recurrent models, computations over all elements can be fully parallelized during training and optimization is easier since the number of non-linearities is fixed and independent of the input length. CNN seq2seq outperforms the accuracy of the deep LSTM setup of Wu et al. (2016) on both WMT'14 English-German and WMT'14 English-French translation at an order of magnitude faster speed, both on GPU and CPU.


There are several sources that you can refer to:

1. Convolutional Sequence to Sequence Learning https://arxiv.org/abs/1705.03122

2. Language modeling with gated linear units https://arxiv.org/abs/1612.08083

3. Deep Residual Learning for Image Recognition https://arxiv.org/abs/1512.03385



Time: Wed 5:00 pm
Venue: Room 3-517

See you then!

Best,
Yizhu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://cs.sjtu.edu.cn/pipermail/adapt/attachments/20180612/9d1cf319/attachment.html>