When Cohesion Lies in the Embedding Space: Embedding-Based Reference-Free Metrics for Topic Segmentation.
Iacopo GhinassiLin WangChris NewellMatthew PurverPublished in: LREC/COLING (2024)
Keyphrases
- embedding space
- topic segmentation
- euclidean space
- low dimensional
- graph embedding
- manifold learning
- dimensionality reduction
- high dimensional
- input space
- geometric structure
- data points
- nonlinear dimensionality reduction
- geodesic distance
- high dimensional data
- shape analysis
- machine learning
- euclidean distance
- lexical chains
- semi supervised
- topic detection
- keywords
- co occurrence
- feature space
- feature vectors
- feature extraction
- locally linear embedding
- training set
- query processing
- nearest neighbor
- discriminant analysis
- data sets