When are overcomplete topic models identifiable? uniqueness of tensor tucker decompositions with structured sparsity.
Animashree AnandkumarDaniel J. HsuMajid JanzaminSham M. KakadePublished in: J. Mach. Learn. Res. (2015)
Keyphrases
- topic models
- structured sparsity
- high order
- latent dirichlet allocation
- statistical learning
- natural images
- sparse representation
- higher order
- text mining
- linear combination
- text documents
- co occurrence
- generative model
- probabilistic model
- learning problems
- latent variables
- sparse coding
- dimensionality reduction
- singular value decomposition
- probabilistic latent semantic analysis
- image processing
- image representation
- bayesian networks
- data mining