Word Re-Embedding via Manifold Dimensionality Retention.
Souleiman HasanEdward CurryPublished in: EMNLP (2017)
Keyphrases
- high dimensional
- nonlinear dimensionality reduction
- dimensionality reduction
- manifold embedding
- graph embedding
- embedding space
- low dimensional
- feature space
- lower dimensional
- laplacian eigenmaps
- manifold learning
- geodesic distance
- intrinsic dimensionality
- low dimensional manifolds
- high dimension
- euclidean space
- head pose estimation
- cognitive load
- high dimensionality
- co occurrence
- long term
- locality preserving projections
- latent space
- word pairs
- high dimensional data space
- vector space
- n gram
- high dimensional data
- diffusion maps
- intrinsic dimension
- pattern recognition
- word recognition
- multidimensional scaling
- data points
- manifold structure
- locally linear embedding
- related words
- kernel pca
- high dimensional spaces
- word sense disambiguation
- similarity search
- feature selection