#sequence-models
- A Hierarchical Latent Variable Encoder-Decoder Model for Generating Dialogues
- Cyclical Annealing Schedule: A Simple Approach to Mitigating KL Vanishing
- Generating Sentences from a Continuous Space
- Learning Discourse-level Diversity for Neural Dialog Models using Conditional Variational Autoencoders
- OPTIMUS: Organizing Sentences via Pre-trained Modeling of a Latent Space
- Token-Sum Versus Token-Mean NLL Aggregation
