Sequence Tutor: Conservative Fine-Tuning of Sequence Generation Models with KL-control
ICML, pp. 1645-1654, 2017.
This paper proposes a general method for improving the structure and quality of sequences generated by a recurrent neural network (RNN), while maintaining information originally learned from data, as well as sample diversity. An RNN is first pre-trained on data using maximum likelihood estimation (MLE), and the probability distribution ov...More
PPT (Upload PPT)