When are Overcomplete Topic Models Identifiable? Uniqueness of Tensor Tucker Decompositions with Structured Sparsity.
Anima AnandkumarDaniel J. HsuMajid JanzaminSham M. KakadePublished in: NIPS (2013)
Keyphrases
- topic models
- structured sparsity
- high order
- latent dirichlet allocation
- statistical learning
- higher order
- linear combination
- text documents
- learning problems
- sparse representation
- generative model
- latent variables
- co occurrence
- natural images
- text mining
- sparse coding
- probabilistic model
- probabilistic latent semantic analysis
- information retrieval
- dimensionality reduction
- learning algorithm