paper-with-me

GEOMANCER

Geometric Manifold Component Estimator

2000년 도입 · 논문 1편에서 사용

Geomancer is a nonparametric algorithm for symmetry-based disentangling of data manifolds. It learns a set of subspaces to assign to each point in the dataset, where each subspace is the tangent space of one disentangled submanifold. This means that geomancer can be used to disentangle manifolds for which there may not be a global axis-aligned coordinate system.

출처: Disentangling by Subspace Diffusion

소개 논문: Disentangling by Subspace Diffusion

Manifold Disentangling · General