Diagonal Quadratic Profiles on the Noise Wasserstein Space: Moment Bounds, a Lipschitz Estimate, the Gradient Field, and Noise Intrinsic Test Functions
Integrals of diagonal quadratics with suitably summable coefficients are finite on all measures noise-connected to a diagonal Gaussian, depend Lipschitz-continuously on the measure, and are noise intrinsic test functions whose gradient is the corresponding linear field.
1. (Moments) Let t=(tk)k∈N be a sequence of nonnegative real numbers such that ∑k=1∞tkck converges, and let M∈R satisfy 0≤M and tkak≤M for every k∈N. For every μ∈Pρa the series St(μ)=∑k=1∞tk∫Xxk2μ(dx) converges, and
St(μ)≤2k=1∑∞tkck+2MWa(μ,ρ)2.
2. (A Lipschitz estimate) Let t and M be as in claim 1, and let μ,ν∈Pρa and π∈Πa(μ,ν). Then the series ∑k=1∞tk∫X×X(yk−xk)2π(d(x,y)) converges with sum at most MIa(π), and
St(μ)−St(ν)≤MIa(π).
3. (The gradient field) Let b be admissible and μ∈Pρa. There is exactly one Vb(μ)∈L2(μ;Xa) whose coordinate along fk is, for every k∈N, the class of the function x↦ak1/2bkxk, and
∥Vb(μ)∥μ2=k=1∑∞akbk2∫Xxk2μ(dx).
4. (The profile) Let b be admissible. For every μ∈Pρa the series ∑k=1∞bk∫Xxk2μ(dx)converges absolutely, and the diagonal quadratic profile with coefficients b is the function
Φb:Pρa→R,Φb(μ)=21k=1∑∞bk∫Xxk2μ(dx).
5. (First-order expansion) Let b be admissible with bound B. For μ,ν∈Pρa and π∈Πa(μ,ν),
Φb(ν)−Φb(μ)−Ja(Vb(μ),π)≤2BIa(π).
6. (Test function) Let b be admissible. For every subset Q⊆Pρa, Φb is a noise intrinsic test function on Q, and ∇Φb(μ)=Vb(μ) for every μ∈Q.
7. (Scaling) Let b be admissible and s∈R. Then the sequence sb=(sbk)k∈N is admissible, and Φsb(μ)=sΦb(μ) and Vsb(μ)=sVb(μ) for every μ∈Pρa.
Curated associations between results. These are editable and subjective — they do not replace the dependency graph, which is derived from the references in the text.