ott.geometry.costs.SqPNorm.transport_map#
- SqPNorm.transport_map(g)#
Get an optimal transport map for a concave function \(g\).
Uses Proposition 1 from [Klein et al., 2024] to define an OT map \(x - (\nabla h^*) \circ \nabla \bar g^h(x)\), where \(h^*\) is the Legendre transform of \(h\) and \(\bar g^h\) is the
h_transform()
of a concave function \(g\).