Sequence to Sequence
Sequence to sequence (seq2seq) model was first developed by Ilya Sutskever at Google1. It can learn the translation of one input sequence to the output sequence.
Problem of RNN
The RNN can easily map sequences to sequences whenever the alignment between the inputs and the outputs is known ahead of time. However, it is not clear how to apply an RNN to problems whose input and the output sequences have different lengths with complicated and nonmonotonic relationships.
Sequence to sequence
- Convolutional Sequence to Sequence Learning (ICML 2017)
- Sequence Tutor: Conservative Fine-Tuning of Sequence Generation Models with KL-control (ICML 2017)
- Sequence to Better Sequence: Continuous Revision of Combinatorial Structures (ICML 2017)