ott.problems.quadratic.quadratic_costs.GWLoss

class ott.problems.quadratic.quadratic_costs.GWLoss(f1, f2, h1, h2)

Efficient decomposition of the Gromov-Wasserstein loss function.

The loss function \(L\) is assumed to match the form given in eq. 5. of [Peyré et al., 2016]:

\[L(x, y) = f_1(x) + f_2(y) - h_1(x) h_2(y)\]

Parameters:

• f1 (Loss) – First linear term.

• f2 (Loss) – Second linear term.

• h1 (Loss) – First quadratic term.

• h2 (Loss) – Second quadratic term.

Methods

count(value, /)	Return number of occurrences of value.
index(value[, start, stop])	Return first index of value.

Attributes

f1	Alias for field number 0
f2	Alias for field number 1
h1	Alias for field number 2
h2	Alias for field number 3