Alleviating "Posterior Collapse" in Deep Topic Models via Policy Gradient.
Yewen LiChaojie WangZhibin DuanDongsheng WangBo ChenBo AnMingyuan ZhouPublished in: NeurIPS (2022)
Keyphrases
- topic models
- policy gradient
- probabilistic model
- topic modeling
- latent variables
- latent dirichlet allocation
- posterior distribution
- reinforcement learning
- optimal control
- function approximation
- generative model
- gradient method
- co occurrence
- text documents
- text mining
- posterior probability
- probability distribution
- reinforcement learning algorithms
- bayesian framework
- gaussian process
- graphical models
- information retrieval
- approximate inference
- least squares
- state action
- image segmentation