Proof of The Law of Large Numbers for Empirical Measures is Uniform on Wasserstein-Compact Sets
corollarycor:empirical-measure-lln-uniform-wasserstein-2026aThe distance to the measure is Lipschitz in the configuration with constant , so changing the tensor-power measure costs at most one Wasserstein distance and changing the target measure costs at most another. Uniformity on a compact set follows by covering it with finitely many small balls and taking the largest of the finitely many thresholds.
Each result cited is universally quantified over the data in its own statement.
Conventions. Fix as in the statement; by Wasserstein Spaces, Random Vectors, Vector Fields and Symmetric Matrices in Every Dimension: Standing Notation, so for every . For , denotes the image of under the canonical map of The Canonical Map from the Natural Numbers to a Field, the real number written in the displayed formulas of the results cited; and by claim 3 of Properties of the Canonical Map from the Natural Numbers to an Ordered Field. For and put for . By Tensor Powers and One-Particle Marginals: Particle Laws, Product Integrals, Push-Forwards, Moments, Product Maps and Diagonal Shifts §moments, . By Law of Large Numbers for Empirical Measures in the Wasserstein Distance §finite, is integrable with respect to and is the nonnegative square root of . For , is a probability space (Probability Space, Event, and Random Variable), Borel maps (Probability Measures on Euclidean Space and Random Vectors: Standing Notation §borel-maps) are random variables on it, the expectation of a nonnegative one is its integral against (Expectation, Variance, and Moments), and square-integrability and the norm , the nonnegative square root of , are those of Square-Integrable Random Variables and the Mean-Square Inner Product; for Borel with integrable against , the norm of Wasserstein Spaces, Random Vectors, Vector Fields and Symmetric Matrices in Every Dimension: Standing Notation §fields is the same number , both being the nonnegative square root of . In particular . For nonnegative reals we use that if and only if , and if and only if (claims 2 and 3 of Monotonicity of Squaring on the Nonnegative Elements of an Ordered Field); we refer to this as (S).
Step 1 (A Lipschitz constant in the configuration). Fix and . Let be the nonnegative square root of (Existence and Uniqueness of the Nonnegative Square Root), so and . Let . By Basic Properties of Empirical Measures: Values, Integrals, Push-Forwards, Second Moment, and Lipschitz Dependence on the Configuration §distance, . By Basic Properties of Empirical Measures: Values, Integrals, Push-Forwards, Second Moment, and Lipschitz Dependence on the Configuration §lipschitz, ; multiplying by (claim 5 of Elementary Arithmetic in an Ordered Field) gives , hence by (S), both sides being nonnegative (the right one by the second axiom of Ordered Field). Therefore . Since is the Euclidean distance of Euclidean Distance on and is the metric of The Absolute Value Metric on the Real Line, is Lipschitz with constant in the sense of Lipschitz Map Between Metric Spaces.
Step 2 (Clause 1). Fix and ; write , , , and . Both lie in by the conventions.
First term. By Step 1, The Mean-Square Norm of a Lipschitz Function Depends Lipschitz-Continuously on the Measure for the Wasserstein Distance §bound applies with , , the function and the measures : is Borel, is integrable against and , and . By Tensor Powers Scale the Wasserstein Distance by the Square Root of N, and the One-Particle Marginal is Lipschitz with Constant N^{-1/2} §tensor, , so ; both and are nonnegative, so by (S). Since by the conventions,
Second term. Let . By Basic Properties of Empirical Measures: Values, Integrals, Push-Forwards, Second Moment, and Lipschitz Dependence on the Configuration §moment, , so the triangle inequality and symmetry of The Quadratic Wasserstein Distance is a Metric on the Wasserstein Space §triangle and The Quadratic Wasserstein Distance is a Metric on the Wasserstein Space §symmetry give and . Hence , i.e. by claim 6 of Properties of the Absolute Value in an Ordered Field, and by (S), since (claim 1 of Properties of the Absolute Value in an Ordered Field and claim 2 of Zero Products and Elementary Identities in a Field). On the probability space the random variables and are square-integrable (by The Mean-Square Norm of a Lipschitz Function Depends Lipschitz-Continuously on the Measure for the Wasserstein Distance §bound as just applied, and by Law of Large Numbers for Empirical Measures in the Wasserstein Distance §finite for ), so is square-integrable by Square-Integrable Random Variables and the Mean-Square Inner Product. The constant is a nonnegative simple function with integral (Simple Function and Its Integral, and the last sentence of Lebesgue Integral of a Nonnegative Measurable Function); by monotonicity of the integral (claim 1 of Linearity and Monotonicity of the Lebesgue Integral), , so by (S). By claim 2 of Cauchy-Schwarz and Triangle Inequalities for the Mean-Square Norm, applied to and to (with , as pointwise by claim 2 of Zero Products and Elementary Identities in a Field),
so by claim 6 of Properties of the Absolute Value in an Ordered Field. As by the conventions, .
Conclusion. By the triangle inequality (claim 5 of Properties of the Absolute Value in an Ordered Field) applied to and the first axiom of Ordered Field (used twice to add the two bounds),
This proves clause 1.
Step 3 (Clause 2: a finite net). Let and be as in clause 2. The objects below are chosen in this order: , then the finite set and its enumeration , then , then . Put ; by claim 8 of Elementary Order Arithmetic in an Ordered Field, applied to and to , we have , and .
By The Quadratic Wasserstein Distance is a Metric on the Wasserstein Space §metric, is a metric on ; let be the collection of its open sets, a topology by Metric Open Sets Form a Topology, and let be the subspace topology on (Subspace Topology); by Compact Topological Space and Compact Subset, is compact. For let be the open ball of Open Ball in a Metric Space; it belongs to by Open Ball in a Metric Space is Open, so . Since by The Quadratic Wasserstein Distance is a Metric on the Wasserstein Space §separation, , so the family covers . By compactness there is a finite with .
If , the union is empty, so and satisfies clause 2 vacuously. Otherwise, by Finite Set and Number of Elements of a Set there are and a bijection , . For each , Law of Large Numbers for Empirical Measures in the Wasserstein Distance §convergence, applied to , and Limit of a Sequence of Real Numbers with the positive number give such that for every with . Define and for with , the maximum for the total order of (Order on the Natural Numbers, claim 3 of Properties of the Order on the Natural Numbers), and put . By claim 1 of Elementary Properties of the Maximum of Two Elements, and , so by induction on with transitivity (claim 1 of Properties of the Order on the Natural Numbers) for all in ; in particular for every .
Step 4 (Clause 2: conclusion). Let with and let . Choose with , so , and by The Quadratic Wasserstein Distance is a Metric on the Wasserstein Space §symmetry. Since , (claim 1 of Properties of the Order on the Natural Numbers), so , and (claim 3 of Properties of the Absolute Value in an Ordered Field) gives (claim 2 of Elementary Order Arithmetic in an Ordered Field). By Step 2 with , and claim 3 of Properties of the Absolute Value in an Ordered Field,
the strict inequality by claim 3 of Elementary Order Arithmetic in an Ordered Field (from and ), combined with claim 2 there. Adding the two strict bounds (claim 3 of Elementary Order Arithmetic in an Ordered Field),
so . As and were arbitrary, and depends only on and (through and the finite set ), clause 2 holds.
Loading…
Prerequisites
d1bab53c-afc2-4d4f-b311-4807ac82e8e3