Metraipolis › archive › dc_faec774e60b9

Hyperbolic Graph Representation Learning: Embed in One Metric, Optimize with Another — arXiv:2610.06745

Source
arXiv
Original
https://arxiv.org/abs/2610.06745
License (as archived)
CC0-1.0
Archived
2026-10-06T05:31:47.268Z
SHA-256 of the archived text
a91cf253d05accaf4ff7a68367a60ec102dcb8e2834d3b0ed3caed2934b4cdaa
Stable URL of this page
https://metraipolis.com/documento/dc_faec774e60b9

How to cite

Hyperbolic Graph Representation Learning: Embed in One Metric, Optimize with Another — arXiv:2610.06745 — arXiv, https://arxiv.org/abs/2610.06745 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_faec774e60b9 (Metraipolis archive dc_faec774e60b9, sha256 a91cf253d05accaf4ff7a68367a60ec102dcb8e2834d3b0ed3caed2934b4cdaa)
BibTeX
@misc{dcfaec774e60b9,
  title = {Hyperbolic Graph Representation Learning: Embed in One Metric, Optimize with Another — arXiv:2610.06745},
  howpublished = {arXiv},
  url = {https://arxiv.org/abs/2610.06745},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_faec774e60b9; sha256 a91cf253d05accaf4ff7a68367a60ec102dcb8e2834d3b0ed3caed2934b4cdaa},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.06745. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Hierarchical graphs embed in hyperbolic space with lower distortion than in Euclidean space owing to its negative curvature. However, their gradient-based learning is hampered at large radii, where the Poincaré ball and the Lorentz hyperboloid models fail numerically. Polar coordinates avoid this problem, but the hyperbolic metric scales the angular step by the hyperbolic sine of the radius, freezing angular motion. We observe that this factor is a choice, silently fixed by existing implementations: the Euclidean tangent parametrization, for instance, uses the radius itself. We show that other choices are not only possible but preferable. They are endpoints of a one-parameter family of optimization preconditioners with curvatures from $-1$ to $0$, while the embedding remains at curvature $-1$. We show that since the Euclidean preconditioner rearranges a layout but refines it poorly, while an intermediate one refines far better once a layout is in place, combining them in two stages reduces the loss on real-world trees by 46-74% over the best single curvature.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG) ; Machine Learning (stat.ML).
Pagina del lavoro: https://arxiv.org/abs/2610.06745 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-06T05:27:05.910Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).