Estimating class separability of text embeddings with persistent homology.
Kostis GourgouliasNajah GhalyanMaxime LabonneYash SatsangiSean MoranJoseph SabeljaPublished in: Trans. Mach. Learn. Res. (2024)
Keyphrases
- class separability
- dimensionality reduction
- persistent homology
- discriminant analysis
- feature space
- feature selection
- linear discriminant analysis
- topological features
- manifold learning
- vector space
- low dimensional
- computational geometry
- pattern recognition
- data sets
- principal component analysis
- data analysis
- multiscale
- machine learning