Prior-aware Dual Decomposition: Document-specific Topic Inference for Spectral Topic Models.
Moontae LeeDavid BindelDavid M. MimnoPublished in: CoRR (2017)
Keyphrases
- topic models
- dual decomposition
- latent topics
- text documents
- topic modeling
- latent dirichlet allocation
- latent variables
- map inference
- generative process
- energy minimization
- max margin
- statistical topic models
- text mining
- markov logic
- probabilistic model
- graphical models
- structured prediction
- generative model
- co occurrence
- relevance model
- keywords
- markov random field
- information retrieval
- document classification
- data mining
- random variables
- prior knowledge
- efficient learning
- approximate inference
- bayesian inference
- maximum a posteriori