Matrix representations, linear transformations, and kernels for disambiguation in natural language.
Tapio PahikkalaSampo PyysaloJorma BobergJouni JärvinenTapio SalakoskiPublished in: Mach. Learn. (2009)
Keyphrases
- linear transformations
- natural language
- linear transformation
- positive definite
- natural language processing
- positive semidefinite
- linear model
- parameter space
- highly ambiguous
- matrix valued
- locally optimal
- wordnet
- machine learning
- information extraction
- kernel function
- support vector
- machine translation
- co occurrence
- word sense disambiguation
- learning tasks
- higher order
- object recognition