PaperlayerView exact revision on arXiv

Source-linked AI summary

Universal Entanglement Distillation

Salvatore Tirone, Francesco Anna Mele, Vittorio Giovannetti, Ludovico Lami

arXiv:2609.40080v1quant-phcs.ITmath.IT

TL;DR

When Alice and Bob lack a classical description of their shared state, the usual formulation of entanglement distillation presumes one. The paper combines LOCC tomography with distillation selected from the estimate, proving asymptotically optimal-rate distillation with vanishing error. For a known family of possible states, it also characterizes the worst-case uniform rate by the minimum distillable entanglement on the family's closure.

  • Problem

    The usual formulation of entanglement distillation assumes a complete classical description of the shared state, which may be unavailable in practice.

    Source: p. 1

  • Method

    The protocol performs LOCC tomography and selects a distillation routine based on the resulting estimate.

    Source: p. 1

  • Results

    With known local dimension and identical independently prepared copies, the protocol asymptotically achieves each input state's distillable entanglement with vanishing error.

    Source: p. 1

  • Takeaways & Limitations

    Under the stated assumptions, ignorance of the state need not impose a permanent asymptotic rate penalty.

    Source: p. 1

    The existence proof is non-constructive; optimal LOCC-tomography sample complexity and extension to multipartite systems remain open.

    Source: p. 4

Abstract

from arXiv · show

We present a universal entanglement distillation protocol that asymptotically achieves the distillable entanglement of any unknown bipartite state in a fixed finite dimension. That is, its ebit yield attains every rate below the optimum that would be available if the state were known, with vanishing trace-distance error. Our approach combines tomography by local operations and classical communication with an adaptive distillation routine, using a local robustness lemma to control finite-sample estimation errors. We also characterize worst-case distillation from a family of states: the optimal uniform rate is the minimum distillable entanglement on the family's closure. These results show that ignorance of the state need not incur an asymptotic rate penalty.

I. PRELIMINARIES

The preliminaries establish the state, channel, and distance notation used in the paper, alongside background on entanglement and distillation.

I. PRELIMINARIES

  • A quantum channel is modeled as a completely positive trace-preserving map between trace-class operator spaces.The notation 𝒟(ℋ) denotes density operators on a Hilbert space ℋ.Source: §I · p. 7
  • Trace distance and fidelity are universally comparable, and fidelity equals one exactly when the two states are equal.Source: §I · pp. 7–8
  • Isotropic states include convex combinations of a maximally entangled state and the normalized projector onto its orthogonal complement.Source: §I · p. 8
  • LOCC denotes channels implementable using local operations and classical communication between bipartite systems.Source: §I · p. 8
  • The hashing bound is a proved lower bound on distillable entanglement, following its conjecture from the hashing protocol.The passage attributes the conjecture to and the proof to Devetak and Winter [10, Theorem 10].Source: §I · p. 9

II. TWO ELEMENTARY ESTIMATES

Two elementary estimates provide continuity controls used in the paper: one for tensor powers of states and one for a fixed LOCC map over finite blocks.

II. TWO ELEMENTARY ESTIMATES

  • Tensor powers of states are locally Lipschitz in trace norm.The proof invokes a telescopic identity and trace-norm multiplicativity under tensor products.Source: §II · p. 9
  • For a fixed LOCC map, the finite-block estimate uses a projector and the variational characterization of trace distance.Source: §II · p. 9
  • The finite-block continuity argument combines channel contractivity with the tensor-power Lipschitz estimate.Source: §II · p. 10

III. CLASSICAL TELEPORTATION AND TOMOGRAPHY OF BIPARTITE STATES VIA LOCCS

The section constructs LOCC tomography for arbitrary bipartite states by teleporting a noisy version of one subsystem and then applying global state tomography.

III. CLASSICAL TELEPORTATION AND TOMOGRAPHY OF BIPARTITE STATES VIA LOCCS

  • The LOCC tomography protocol first teleports a noisy version of Alice's subsystem to Bob, then applies global tomography to the resulting state.The approach is presented as applying to arbitrary mixed bipartite states, unlike prior LOCC tomography tailored to pure states.Source: §III · p. 10
  • Global tomography of an unknown D-dimensional state requires order D^2/ε^2 copies for trace-distance error ε at constant success probability.The passage attributes the optimal scaling to.Source: §III · p. 10
  • Using the separable isotropic resource W_d makes the averaged teleported output a mixture of the original state and a partially traced, maximally mixed subsystem.The resource uses the same teleportation steps as the standard protocol but replaces its maximally entangled resource with W_d.Source: §III · pp. 10–11
  • A Haar-distributed POVM on Alice's subsystem can be replaced by a finite projective two-design, allowing finite classical communication.Alice communicates the measurement outcome to Bob, who prepares the corresponding state locally.Source: §III · p. 11

A. Proof of Theorem 1 of the main text

The proof of Theorem 1 establishes an LOCC tomography guarantee by estimating a noisy transformed state and applying an inverse transformation, with correction when the raw estimate is not positive.

A. Proof of Theorem 1 of the main text

  • O(d^6/ε^2) copies suffice for constant-failure-probability LOCC tomography of an unknown d-dimensional-per-party bipartite state.The theorem states that the protocol outputs a density-operator estimator with trace-distance error controlled by ε.Source: A. Proof of Theorem 1 of the main text · pp. 12, 14
  • The proof estimates the noisy transformed state, applies the inverse channel, then discards the raw estimate's negative part and renormalizes.The raw inverse estimate need not be positive semidefinite because the inverse is not a quantum channel.Source: A. Proof of Theorem 1 of the main text · p. 13
  • For unequal local dimensions, the analysis also applies by teleporting the smaller subsystem.Source: A. Proof of Theorem 1 of the main text · p. 14

IV. UNIFORM HASHING FOR AN ISOTROPIC SHELL

This section constructs a hashing protocol for isotropic states with performance uniform over a fidelity interval. It uses sequential parity extraction to decode Bell-error strings and convert successful decoding into ebits.

IV. UNIFORM HASHING FOR AN ISOTROPIC SHELL

  • The hashing map is independent of the isotropic state's fidelity and achieves a rate across the specified shell.The finite-block estimate supplies the same LOCC map throughout the interval.Source: §IV · p. 15
  • The rate check relates the retained Bell pairs to the shell's smallest coherent information, then discards surplus pairs to meet the target rate.The coherent information is identified with ℓ − H(p_q) for the isotropic states considered.Source: §IV · pp. 16, 20
  • Each round consumes one Bell pair while local measurements reveal a chosen parity of the Bell-error label.The bilateral Clifford operation makes the measured parity a known linear function of the label; the residual label is updated for subsequent rounds.Source: §IV · pp. 16–17
  • After the rounds, candidates matching the observed transcript are retained, and a correction is applied when they share one residual Bell label.Decoding fails only if a distinct candidate has the same transcript but a different final residual label.Source: §IV · p. 19
  • The decoding-error bound tends to zero uniformly over the fidelity interval for fixed positive slack.The bound is stated to be independent of the fidelity parameter.Source: §IV · pp. 15, 19

V. LOCAL ROBUSTNESS AND LOWER SEMICONTINUITY OF THE DISTILLABLE ENTANGLEMENT

Local robustness extends any rate below a state's distillable entanglement to a neighborhood using one fixed sequence of LOCC maps. This yields lower semicontinuity of distillable entanglement in trace distance.

V. LOCAL ROBUSTNESS AND LOWER SEMICONTINUITY OF THE DISTILLABLE ENTANGLEMENT

  • For any distillable state and any rate below its distillable entanglement, one fixed LOCC sequence works uniformly on a sufficiently small trace-distance ball.The radius can depend on both the state and the chosen rate.Source: §V · p. 20
  • The robustness construction fixes the distillation map using the center state and rate, then uses twirling and hashing to handle nearby inputs.The fixed map does not depend on the unknown nearby state.Source: §V · pp. 20–21
  • Distillable entanglement is lower semicontinuous: states sufficiently close to a state ρ retain every rate below Ed(ρ).The result also gives the corresponding asymptotic lower bound for any sequence converging to ρ.Source: §V · p. 22

VI. PROOF OF THEOREM 3 OF THE MAIN TEXT

The proof builds a single state-independent LOCC protocol that estimates the input locally and selects among robust distillation routines. For every state, it achieves each rate below that state's distillable entanglement with vanishing error.

VI. PROOF OF THEOREM 3 OF THE MAIN TEXT

  • The active menu grows slowly while tomography becomes more accurate, so tomography costs o(n) copies and the worst active-protocol error vanishes.Accuracy is chosen to remain within each active entry's robustness margin.Source: §VI · pp. 23, 25, 27
  • For every state ρ and rate R < Ed(ρ), one state-independent LOCC sequence supports a readout yielding ⌈Rn⌉ pairs with vanishing error and shortfall probability.The universal output includes a declared length; the rate-specific readout keeps the requested number when available.Source: §VI · pp. 23, 29
  • The protocol uses a countable menu of local robust routines, each guaranteeing a rate on a ball around its center.For a fixed rate label, the balls cover the corresponding open set of states with greater distillable entanglement.Source: §VI · pp. 23–24
  • The active routines' worst-case error and the tomography failure probability both vanish, while the probability of declaring fewer than ⌈Rn⌉ pairs tends to zero.The stated tomography failure probability is at most n^-2.Source: §VI · pp. 27, 29
  • After LOCC tomography, the protocol selects the highest-rate admissible active entry; its safety margin ensures that the selected routine also applies to the true state.The estimate-to-center distance and tomography error together place the true state inside that routine's robustness ball.Source: §VI · pp. 27–29

VII. PROOF OF THEOREM 4 OF THE MAIN TEXT

For a family of states, the proof combines local robust protocols with tomography to establish a uniform achievable rate. It also uses trace-norm continuity of the family-based figure of merit to pass from closed families to their closures.

VII. PROOF OF THEOREM 4 OF THE MAIN TEXT

  • The family-based figure of merit is unchanged when X is replaced by its closure, allowing the closed-family result to extend to general families.The passage states continuity in trace norm and equality of the corresponding suprema over X and its closure.Source: §VII · pp. 30, 32
  • Any universally achievable rate on a family X is no greater than Ed(ρ) for every ρ in X.Thus it is bounded above by the infimum of distillable entanglement over the family.Source: §VII · p. 30
  • For a closed family, finitely many local robustness balls cover X, and tomography selects a corresponding LOCC routine for the remaining copies.The finite subcover follows from compactness; the selected routine works at the fixed rate on its ball.Source: §VII · pp. 31–32

Related papers

Loading 2609.40080v1…