End-to-End latent-variable task-oriented dialogue system with exact log-likelihood optimization.
Haotian XuHaiyun PengHaoran XieErik CambriaLiuyang ZhouWeiguo ZhengPublished in: World Wide Web (2020)
Keyphrases
- end to end
- dialogue system
- log likelihood
- latent variables
- natural language
- maximum likelihood
- information theoretic
- scoring function
- tutorial dialogue
- density estimation
- latent variable models
- closed form
- em algorithm
- random variables
- probabilistic model
- congestion control
- topic models
- user model
- human users
- information retrieval
- machine learning
- hidden variables
- parameter estimation
- association rules
- end to end distortion