FDP-LDA: Inherent Privacy Amplification of Collapsed Gibbs Sampling via Group Subsampling.
Tao HuangHong ChenSuyun ZhaoPublished in: APWeb/WAIM (3) (2022)
Keyphrases
- collapsed gibbs sampling
- latent dirichlet allocation
- topic models
- variational bayesian inference
- gibbs sampling
- variational inference
- topic modeling
- iterative deepening
- generative model
- approximate inference
- probabilistic model
- text mining
- linear discriminant analysis
- belief propagation
- bayesian inference
- parameter estimation
- markov chain
- dimensionality reduction
- latent variables
- evaluation function
- feature selection
- gaussian process
- graphical models
- co occurrence
- probability distribution
- data analysis
- face recognition