TheoremBase

Plan Jets of a Series of Squared Wasserstein Distances at Optimal Couplings of Bounded Noncommutative Laws

For a summable series of squared Wasserstein distances to bounded laws, realised along optimal couplings in one tracial W*-probability space, the law of the position with the summed displacement momentum is a plan superjet, its negative momentum gives a plan subjet of the negative, and the plan pairing splits into the coupling pairings.

Statement

In the setting of Plan Jets and Hamiltonians on Square-Integrable Noncommutative Laws: Standing Notation, bounded plans and the momentum norm ∣⋅∣mom|\cdot|_{\mathrm{mom}} are those of Marginal Isometries, Bounded Plans and Displacement Pairings for Noncommutative Laws §plans, the plan pairing J\mathcal{J} that of Marginal Isometries, Bounded Plans and Displacement Pairings for Noncommutative Laws §plan-pairing and the coupling pairings Jγ1\mathcal{J}^{1}_{\gamma} those of Marginal Isometries, Bounded Plans and Displacement Pairings for Noncommutative Laws §coupling-pairing; J+J^{+} and J−J^{-} are the plan superjet and the plan subjet with slack 00. Self-adjoint dd-tuples ss in MM, their vacuum tuples sΩs\Omega and their laws λs\lambda_{s} are those of Self-Adjoint Tuples in a Tracial W*-Probability Space and Their Laws, and (s,t)(s,t) is the self-adjoint 2d2d-tuple of Laws of Self-Adjoint Tuples in a Tracial W*-Probability Space: Moments, Affine Images, Couplings, Embeddings and L^2 Approximation §coupling; vacuum tuples of self-adjoint dd-tuples are L2L^{2} dd-tuples by Calculus of Laws of Square-Integrable Tuples: Bounded Tuples, the Lipschitz Bound, Moments, Affine Push-Forwards, Couplings and Embeddings §bounded, and sums, differences, real multiples and the L2L^{2} norm ∥⋅∥2\lVert\cdot\rVert_{2} of L2L^{2} dd-tuples are those of Plan Jets and Hamiltonians on Square-Integrable Noncommutative Laws: Standing Notation §pairing.

Let R>0R>0 be real and μ∈Σd,R\mu\in\Sigma_{d,R}; for every k∈Nk\in\mathbb{N} let νk∈Σd,R\nu_{k}\in\Sigma_{d,R} and let γk∈Π(μ,νk)\gamma_{k}\in\Pi(\mu,\nu_{k}) be an optimal coupling; and let (ck)k∈N(c_{k})_{k\in\mathbb{N}} be a sequence of nonnegative reals whose series converges. Let (H,M,Ω)(H,M,\Omega) be a tracial W*-probability space and let ss and tkt^{k} (k∈Nk\in\mathbb{N}) be self-adjoint dd-tuples in MM with ∥sj∥op≤R\lVert s_{j}\rVert_{\mathrm{op}}\le R and ∥tjk∥op≤R\lVert t^{k}_{j}\rVert_{\mathrm{op}}\le R for all j∈[d]j\in[d] and k∈Nk\in\mathbb{N}, λs=μ\lambda_{s}=\mu, and λ(s,tk)=γk\lambda_{(s,t^{k})}=\gamma_{k} for every k∈Nk\in\mathbb{N}; such data exist by Gluing Countably Many Noncommutative Couplings with a Common First Marginal in One Tracial W*-Probability Space §glue.

1. (Momentum) The series ∑k=1∞ckW2(μ,νk)\sum_{k=1}^{\infty}c_{k}W_{2}(\mu,\nu_{k}) converges. For every j∈[d]j\in[d] the partial sums ∑k=1n2ck(sj−tjk)\sum_{k=1}^{n}2c_{k}(s_{j}-t^{k}_{j}) converge in operator norm, as n→∞n\to\infty, to a self-adjoint element PjP_{j} of MM; thus P=(P1,…,Pd)P=(P_{1},\dots,P_{d}) and −P=(−P1,…,−Pd)-P=(-P_{1},\dots,-P_{d}) are self-adjoint dd-tuples in MM. Moreover, for every n∈Nn\in\mathbb{N},

∥PΩ−∑k=1n2ck(sΩ−tkΩ)∥2≤2(∑k=1∞ckW2(μ,νk)−∑k=1nckW2(μ,νk)),∥PΩ∥2≤2∑k=1∞ckW2(μ,νk).\Bigl\lVert P\Omega-\sum_{k=1}^{n}2c_{k}\bigl(s\Omega-t^{k}\Omega\bigr)\Bigr\rVert_{2}\le2\Bigl(\sum_{k=1}^{\infty}c_{k}W_{2}(\mu,\nu_{k})-\sum_{k=1}^{n}c_{k}W_{2}(\mu,\nu_{k})\Bigr),\qquad\lVert P\Omega\rVert_{2}\le2\sum_{k=1}^{\infty}c_{k}W_{2}(\mu,\nu_{k}).

2. (Plans) π=λ(s,P)\pi=\lambda_{(s,P)} and π−=λ(s,−P)\pi^{-}=\lambda_{(s,-P)} are bounded plans at μ\mu, κ2d(π)\kappa_{2d}(\pi) and κ2d(π−)\kappa_{2d}(\pi^{-}) are plans at κd(μ)\kappa_{d}(\mu), and ∣π∣mom=∣π−∣mom=∥PΩ∥2|\pi|_{\mathrm{mom}}=|\pi^{-}|_{\mathrm{mom}}=\lVert P\Omega\rVert_{2}.

3. (Superjet) For every λ∈Σd2\lambda\in\Sigma^{2}_{d} the series ∑k=1∞ckW^2(λ,κd(νk))2\sum_{k=1}^{\infty}c_{k}\widehat{W}_{2}(\lambda,\kappa_{d}(\nu_{k}))^{2} converges. With φ:Σd2→R\varphi:\Sigma^{2}_{d}\to\mathbb{R} its sum, φ(κd(μ))=∑k=1∞ckW2(μ,νk)2\varphi(\kappa_{d}(\mu))=\sum_{k=1}^{\infty}c_{k}W_{2}(\mu,\nu_{k})^{2} and κ2d(π)∈J+φ(κd(μ))\kappa_{2d}(\pi)\in J^{+}\varphi\bigl(\kappa_{d}(\mu)\bigr).

4. (Subjet) With φ\varphi as in claim 3, κ2d(π−)∈J−(−φ)(κd(μ))\kappa_{2d}(\pi^{-})\in J^{-}(-\varphi)\bigl(\kappa_{d}(\mu)\bigr).

5. (Pairing) For every ζ∈Hμd\zeta\in\mathcal{H}_{\mu}^{d} the series ∑k=1∞2ck Jγk1(ζ)\sum_{k=1}^{\infty}2c_{k}\,\mathcal{J}^{1}_{\gamma_{k}}(\zeta) converges, and

J(ζ,π)=∑k=1∞2ck Jγk1(ζ),J(ζ,π−)=−J(ζ,π).\mathcal{J}(\zeta,\pi)=\sum_{k=1}^{\infty}2c_{k}\,\mathcal{J}^{1}_{\gamma_{k}}(\zeta),\qquad\mathcal{J}(\zeta,\pi^{-})=-\mathcal{J}(\zeta,\pi).

Proofs

Log in to submit a proof.

Loading...

Citations

Loading…

Dependencies

Loading…

Related

0 relations

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.

No relations recorded yet.

Comments

Log in to comment.

Loading…