Coarse-graining support: #
Support.Book.Ch05.Theorems.Section53.JUpperBoundWeakNorms.WeightedChildren
Imported from the Apache-2.0 CoarseGraining development at commit
c7ddd76c08ade64fed1b8d2ca51be14dfee8deb4.
WeightedChildren #
Weighted child response observables and stationarity cancellations.
Private Section 5.3 stationarity cancellation for weighted child responses.
The manuscript later supplies weight R = 1 - (phi)_R; the only stochastic
input here is the Ch4 source theorem for weighted descendant response
expectations.
Manuscript cutoff child weight 1 - (phi)_R. This is deterministic;
measurability and stationarity enter only through the response observable that
it weights.
Equations
Instances For
If the cutoff has parent average one, then the finite average of the child
weights 1 - (phi)_R is zero.
Cutoff-weighted child response average from the manuscript splitting.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Childwise response integrability makes the cutoff-weighted child average integrable. The cutoff weights are deterministic scalars.
The manuscript cutoff-weighted child response has zero expectation under stationarity once the cutoff is normalized to have parent average one.
Parent centered response in the Section 5.3 notation.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The centered parent response minus the cutoff-weighted child response average. This is the stochastic side of the manuscript's centered splitting.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The expectation of the centered splitting term is just the centered parent expectation: the cutoff-weighted child contribution cancels by stationarity and normalization.
Private additivity-defect observable from the manuscript proof:
child-scale average of J minus parent-scale J.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The additivity-defect observable has expectation tau_{m,k} under
stationarity. This is the stochastic identity used in the square-root
Cauchy step of Lemma l.J.upper.bound.weak.norms.homogenization.scale.
If a scalar observable is a.e. bounded in absolute value by an integrable right-hand side, then its expectation is bounded by the expectation of that right-hand side.
Expectation assembly for the centered-minus-child split: stationarity turns the left side into the centered parent expectation, and an a.e. deterministic absolute-value bound controls it by the expectation of any integrable RHS.