Can Quantum Entanglement Increase the Capacity of a Classical Relay Channel?

AI summary

Overview

Research area: Quantum information theory and network information theory, specifically entanglement-assisted classical communication over relay channels. Technical level: Advanced. Scope: The paper determines whether pre-shared quantum entanglement between a sender and a relay can increase the capacity of a classical relay channel, contrasting strictly causal unit-delay relaying with causal zero-delay relaying.

What This Paper Is About

A relay helps a sender reach a distant destination, and in a classical relay channel the relay’s transmission may be based either only on past observations (unit-delay, strictly causal encoding) or also on the current observation (zero-delay, causal encoding). The paper asks whether giving the sender and relay pre-shared quantum entanglement can increase the classical message capacity in each of these two models. The answer turns out to depend sharply on which relay encoding model is used: entanglement gives no gain in the strictly causal model, but can give a significant gain in the zero-delay model.

Key Contributions

1. Cutset bound with entanglement for unit-delay relays. The authors prove that the standard cutset upper bound, (I_{\text{cut}}^{(1)} = \max_{p_{XX_r}} \min{I(XX_r;Y),, I(X;YY_r|X_r)}), remains valid for unit-delay relay channels even when the sender and relay share arbitrary entanglement. Hence (C_{\text{ET}}^{(1)} \le I_{\text{cut}}^{(1)}).

2. No capacity gain for several relay classes. Since the cutset bound is known to be tight for degraded, reversely degraded, semi-deterministic, and orthogonal-sender-component relay channels, the result implies entanglement does not improve capacity for any of these classes.

3. A zero-delay relay channel with an entanglement advantage. The authors construct a classical zero-delay relay channel from the magic-square nonlocal game and show that entanglement between sender and relay enables rates exceeding the cutset upper bound by at least 12.5%.

4. A converse technique for the unassisted zero-delay case. To establish the separation, the authors bound rates achievable without entanglement using an approach inspired by Seshadri et al., translating locality into optimization constraints and exploiting the parity structure of classical magic-square responses.

Main Findings

• Strictly causal relaying: no entanglement gain. For van der Meulen’s unit-delay relay model, the cutset upper bound continues to hold with entanglement assistance. Entanglement does not increase capacity for the relay-channel classes where the cutset bound is tight.

• Causal relaying: entanglement can help substantially. For a zero-delay relay, where the relay may use the current observation (Y_{r,i}) as well as past observations, the authors exhibit a channel where entanglement enables rates at least 12.5% above the cutset upper bound.

• The dividing line is encoding delay. The sharp contrast is between strictly causal encoding, where the relay sees only (Y_r^{i-1}), and causal encoding, where the relay sees (Y_r^{i}). The same classical channel model with different synchronization produces qualitatively different entanglement-assisted behavior.

• Game-based channel construction. The zero-delay example uses a relay channel built from the magic-square game: the sender’s input includes a question (q_0) and answer (a_0), the relay observes a question (q_1), the relay transmits an answer (a_1), and the destination output equals (q_0) when the answers satisfy the game’s winning condition and is otherwise randomized uniformly over (\mathcal{Q}_0).

• Quantum strategy wins the magic-square game with certainty. The players share two maximally entangled qubit pairs, (|\Phi^+\rangle_{A'G'} \otimes |\Phi^+\rangle_{A''G''}), and measure commuting observables from the rows and columns of the magic-square table. Classical deterministic strategies cannot win for every question pair, which underlies the rate separation.

• Connection to prior entanglement advantages. Earlier work showed entanglement can increase the sum-rate of a two-user classical multiple-access channel, with a later improvement to at least 4.7%. This paper shows a larger 12.5% separation in a single-user relay setting, but only for zero-delay relaying.

Methodology in Plain English

The authors model the relay channel as a discrete memoryless channel with finite alphabets and a conditional distribution (P_{YY_r|XX_r}) for the unit-delay case, or (P_{Y_r|X}P_{Y|XX_rY_r}) for the zero-delay case. Entanglement assistance is described by a bipartite state shared between the sender and relay, with the sender’s encoding and the relay’s encoding implemented as quantum measurements on their respective shares.

For the negative result, the proof tracks the joint quantum-classical state through the protocol using completely positive trace-preserving maps and quantum instruments. Fano’s inequality bounds the message rate by (I(M;Y^n)), and the authors derive two bounds: a multiple-access cut (I(X_iX_{r,i};Y_i)) and a broadcast cut (I(X_i;Y_iY_{r,i}|X_{r,i})). The broadcast cut is proved with a telescoping argument over auxiliary systems that include the relay’s residual quantum memory, using the data-processing inequality and strong subadditivity. Summing over channel uses and applying a time-sharing argument yields the cutset bound.

For the positive result, the authors build a classical relay channel whose noiseless behavior is controlled by whether the sender’s and relay’s symbols form a winning play of the magic-square game. Entanglement lets the sender and relay coordinate their answers to satisfy the winning condition, while without entanglement the locality constraints of the game limit achievable rates. The converse for the unassisted case exploits the specific parity relations among classical magic-square responses. The provided paper text is truncated partway through the definition of the zero-delay channel, so the complete rate expressions and full converse details are not present in the excerpt.

Why This Matters

• Sharpens the theory of entanglement in networks. The result shows that entanglement assistance is not uniformly beneficial or uniformly useless in classical networks; the outcome depends on causal structure and encoding delay.

• Complements multiple-access and broadcast results. Prior work found an entanglement gain for a classical multiple-access channel and no gain for a broadcast channel with entangled receivers. This paper adds a relay-channel setting where both positive and negative results coexist depending on the relay model.

• Identifies when entanglement can be safely ignored. For unit-delay relays in the degraded, reversely degraded, semi-deterministic, and orthogonal-sender-component classes, the classical cutset capacity remains the correct benchmark despite available entanglement.

• Potential applications in relay-assisted communication. Relay links appear in wireless mesh networks, satellite and deep-space links, cellular cooperative communication, and multi-hop infrastructure. The results indicate that entanglement-assisted coordination could matter in low-delay relay architectures but not in unit-delay architectures.

• Relevance to quantum-enabled network design. The 12.5% separation gives a concrete target for architectures that combine classical relay transmission with shared quantum resources, and it suggests that synchronization and delay assumptions must be part of any performance analysis.

Future Directions

• Determine the entanglement-assisted capacity of general zero-delay relay channels, beyond the specific magic-square construction.

• Characterize broader classes of relay channels, or general causal network topologies, where entanglement provides a strict capacity advantage.

• Extend the analysis to multi-relay, multi-user, and interference settings, where the interaction between causal encoding and entanglement is not yet understood.

• Investigate larger separations using other nonlocal games, and study the effect of imperfect entanglement, finite blocklengths, and practical measurement constraints.

Target Audience

Quantum information theorists, network information theory researchers, and graduate students working on entanglement-assisted classical communication. The paper is also relevant to communication engineers and architects interested in the fundamental limits of relay-assisted networks with quantum resources, particularly those studying causal versus strictly causal relay operation.

Authors’ abstract

We show positive and negative results on classical relay channels with entanglement assistance, revealing a sharp contrast between strictly causal and causal relay encoding. First, we consider van der Meulen's setting of a unit-delay relay, with strictly causal encoding. We show that the standard cutset upper bound continues to hold with entanglement assistance, and deduce that entanglement does not increase the capacity for several classes of relay channels. Second, we consider a zero-delay relay, where the encoding is causal. We present an example in which entanglement enables rates that exceed the cutset upper bound by at least 12.5%. Namely, entanglement shared between the source and relay significantly increases the capacity of a classical zero-delay relay channel.