RegD embeds hierarchical data in Euclidean space using arbitrary regions and a depth-based dissimilarity that emulates hyperbolic expressiveness including exponential growth.
Sincereg ′ 2 ⊆reg 2, of course we have dbd(reg1,reg ′ 2)≤d bd(reg1,reg 2)
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
RegD: Hierarchical Embeddings via Dissimilarity between Arbitrary Euclidean Regions
RegD embeds hierarchical data in Euclidean space using arbitrary regions and a depth-based dissimilarity that emulates hyperbolic expressiveness including exponential growth.