Morphological Priors for Probabilistic Neural Word Embeddings.
Parminder BhatiaRobert GuthrieJacob EisensteinPublished in: EMNLP (2016)
Keyphrases
- learned from training data
- co occurrence
- generalized em algorithm
- prior probabilities
- probabilistic model
- network architecture
- bayesian networks
- multiscale
- keywords
- neural network
- uncertain data
- word sense disambiguation
- mathematical morphology
- sparck jones
- generative model
- structuring elements
- high dimensional
- prior knowledge
- dimensionality reduction
- distance measure
- vector space
- bayesian framework
- high dimensional data
- belief networks
- n gram
- prior information
- morphological operators
- learning rules
- semi supervised
- probability distribution
- metropolis hastings algorithm
- information theoretic