0.25
0.5
0.75
1.25
1.5
1.75
2
CRNN's
Published on Jul 27, 20173540 Views
Related categories
Chapter list
SeaRNN: training RNNs with global-local losses00:00
RNNs: models for sequential data00:20
Encoder-decoder architecture00:38
Standard training01:37
Structured prediction03:55
Learning to Search, a close relative? 04:53
L2S, roll-in/roll-out06:00
Links to RNNs06:46
Our approach: SeaRNN07:37
Roll-outs in RNNs08:19
The devil in the details09:19
Expected benefits11:38
Experimental results12:48
Experimental takeaways13:34
Future work14:54
Questions15:55