Multiplicative separability of the positive linear functional contrast

Let AA be a real-valued positive definite operator representing an attribute, let MM represent the action of a mechanism, and let UU be a unitary operator diagonalizing AA. Let G\mathcal{G} be the generic subgroup of operators, and define the positive linear functional contrast by

C(A)=A,e,C(A)=\langle A,e\rangle,

where ee is the identity operator. The action is mA=MAMmA=MAM^*, where * denotes complex conjugation.

Multiplicative separability conjecture. For all positive definite attributes AA, whenever UU belongs to G\mathcal{G}, the contrast is multiplicatively separable and

Cm,x=C(MM)C(A).\langle C\rangle_{m,x}=C(MM^*)C(A).

Such a multiplicative form for the contrast would support the derivation of a forward-backward inequality under one additional assumption. The source does not provide evidence that the conjecture has been resolved.

Sources & referencesView supporting material

Primary source

Michel Besserve, Naji Shajarisales, Bernhard Schölkopf and Dominik Janzing, “Group invariance principles for causal generative models”, arXiv:1705.02212 (2017).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.