Distributed Instrument Simulation with Quantum Side Information in the One-Shot Regime

Technical Summary: Distributed Instrument Simulation with Quantum Side Information in the One-Shot Regime

Problem Statement
This paper addresses the problem of distributed quantum instrument simulation in a network setting involving three parties: two transmitters (Alice$_1$ and Alice$_2$) and a receiver (Charlie). The parties share a tripartite quantum state $\rho_{A_1 A_2 C}$, where Alice$_i$ holds system $A_i$ and Charlie holds the quantum side information (QSI) system $C$. The objective is to simulate the action of a separable instrument $\mathcal{J}$ acting on the joint system $A_1 A_2$. Unlike a standard Positive Operator Valued Measure (POVM), which only yields a classical outcome, an instrument produces both a classical outcome and a post-measurement quantum state.

The simulation protocol requires:

1. Local Encoding: Each Alice$_i$ applies a local encoding instrument to her system $A_i$, utilizing shared randomness at rate $C_i$.
2. Communication: Each Alice$_i$ sends a classical message to Charlie over a noiseless bit pipe at rate $R_i$.
3. Decoding: Charlie, utilizing his QSI $C$, the received messages, and the shared randomness, reconstructs the classical outcome $Y$ and ensures the post-instrument quantum state (held by the Alices and the reference system) is indistinguishable from the target state.

The authors investigate this problem in two regimes: the one-shot regime (single use of the channel/state) and the asymptotic regime (independent and identically distributed (IID) uses).

Methodology
The paper employs Shannon-theoretic techniques adapted for quantum information, specifically leveraging Sen's smooth multiparty covering and simultaneous decoding.

1. One-Shot Regime (IID Codes):

   • Protocol Construction: The authors propose a protocol based on unstructured IID codes and likelihood POVMs. The encoding instruments are constructed using a multi-codebook structure involving common randomness indices ($K_i$), message indices ($M_i$), and bin indices ($B_i$).
   • Decoding: Charlie employs a hypothesis-testing-based decoder (similar to Classical-Quantum Multiple Access Channel decoding) to recover bin indices, followed by stochastic post-processing to generate the final outcome.
   • Analytical Tools: The proof utilizes a "proxy state" strategy to decompose the simulation error into four components: distributed quantum classical (QC) covering, Classical-Quantum Multiple Access Channel (CQMAC) packing, and two quantum-only covering terms.
   • Key Technical Innovation: To handle the distributed nature and the lack of time-sharing in the one-shot regime, the authors introduce a "compatible operator sliding trick." This technique allows for the transformation of references and the removal of random inverse operators (arising from likelihood POVMs) while preserving the trace distance bounds. This overcomes limitations found in prior work (e.g., Atif et al.) where one-shot analysis was restricted or sub-optimal.

2. Asymptotic Regime (Coset Codes):

   • Protocol Construction: The authors utilize coset codes endowed with joint algebraic properties over a finite field. This approach allows the receiver to recover the sum of hidden indices (algebraic structure) rather than individual indices, potentially yielding larger inner bounds.
   • Decoding: The receiver uses the QSI $C$ and a hypothesis-testing decoder to recover the algebraic component (the sum of the bin indices).
   • Analysis: The proof combines a point-to-point coset-structured likelihood POVM block with the distributed binning mechanism. It isolates the error caused by hidden bin indices and utilizes the finite-field structure to achieve packing gains.

Key Contributions and Results

1. New Inner Bounds:

   • Theorem 1 (One-Shot): Characterizes a new set of inner bounds for the rate quadruples $(R_1, R_2, C_1, C_2)$ required for $\eta$-simulation. The bounds are expressed in terms of smooth entropic quantities (e.g., $I^\epsilon_2$, $D^\epsilon_2$) involving the QSI and the post-measurement states.
   • Theorem 2 (Asymptotic): Provides an asymptotic inner bound using coset codes. This result demonstrates that utilizing the algebraic structure of coset codes can improve rates, particularly when the target instrument admits an efficient decomposition via post-processing of a bivariate function.

2. Generalization of Instrument Simulation:

   • The paper extends the measurement compression problem (MCP) to separable instruments. Unlike prior works that focused on POVMs (which discard the post-measurement phase), this formulation preserves the accessible component's phase. This is crucial for tasks requiring phase preservation, such as purity and entanglement distillation.
   • The formulation explicitly incorporates Quantum Side Information (QSI) at the receiver, requiring simultaneous decoding of bin indices, a feature not fully addressed in previous one-shot studies.

3. Recovery of Known Results:

   • The derived one-shot bounds recover all previously known inner bounds for instrument and measurement simulation in scenarios where those results apply (e.g., single transmitter, no QSI, or asymptotic limits).

Significance and Claims
The authors claim that their work addresses a central problem in quantum information theory by relaxing common restrictions found in prior literature. Specifically:

• Beyond POVMs: By simulating general instruments rather than just POVMs, the results are applicable to a broader class of quantum tasks that require the preservation of the post-measurement quantum state's phase.
• One-Shot Generality: The paper provides a rigorous one-shot analysis for distributed scenarios, overcoming the difficulty of lacking simplifying techniques like time-sharing. The "compatible operator sliding trick" is presented as a novel tool to handle the distributed-component scenario.
• Efficiency with QSI: The work demonstrates how to efficiently utilize QSI to compress communication rates through simultaneous decoding, a necessary step for network scenarios involving multiple transmitters.
• Algebraic Gains: The asymptotic result via coset codes highlights that structured codes can yield higher rates than unstructured IID codes in distributed settings, analogous to findings in classical distributed source coding.

The paper positions itself as a foundational step in understanding the information content of distributed quantum measurements with side information, providing the necessary mathematical tools (smooth covering, simultaneous decoding, operator sliding) to tackle these complex network scenarios.