⚛️ quantum physics

Polynomial Separation between One-Way and Two-Way LOCC in Discrimination of Maximally Entangled States

This paper establishes a polynomial separation between one-way and two-way local operations and classical communication (LOCC) by proving that a single feedback message enables the perfect discrimination of up to $\Omega(d^{1/4})$ orthogonal maximally entangled states in local dimension $d$ , whereas one-way LOCC fails for as few as four states when $d \ge 4$ .

Original authors: Youngrong Lim

Published 2026-10-06

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Youngrong Lim

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

In the strange world of quantum mechanics, particles can become linked in a way that defies our everyday experience. When two particles are "entangled," their properties are so deeply connected that measuring one instantly reveals the state of the other, no matter how far apart they are. This connection is the engine behind many proposed quantum technologies, from ultra-secure communication to powerful new computers. However, a fundamental puzzle has long challenged physicists: if two people, Alice and Bob, are far apart and share these linked particles, can they always figure out exactly which specific state they are holding just by talking to each other and measuring their own pieces?

The answer depends heavily on how they are allowed to communicate. If they can only send messages in one direction—say, Alice tells Bob what she found, and Bob makes his decision based on that—they hit a hard wall. Research has shown that for certain sets of four or more linked states, this one-way conversation is simply not enough to tell them apart with certainty. The information is hidden in the correlations between the particles, and a single message cannot unlock it all. This limitation suggests that the ability to distinguish these states is strictly bounded by the size of the system, regardless of how much they talk. But what happens if they are allowed to talk back and forth? Does a second message change everything, or is the barrier still there?

A new study by Youngrong Lim at Chungbuk National University provides a definitive answer, showing that a single round of feedback is enough to shatter that barrier. The researcher demonstrates that while one-way communication can only distinguish a fixed, small number of states no matter how large the system grows, allowing Bob to send a single message back to Alice unlocks a dramatic new capability. In this two-way scenario, the number of states they can perfectly identify grows rapidly as the system gets larger. Specifically, if the local size of the system is large enough, they can distinguish a number of states that scales with the fourth root of that size. This means that by simply adding one extra step to their conversation, they can move from being stuck with a constant limit to being able to handle a number of candidates that increases polynomially with the system's dimensions.

To understand how this works, imagine Alice and Bob sharing a set of perfectly linked particles. They know the list of possible states they might have, but they do not know which specific one was prepared. In a one-way protocol, Alice measures her part and tells Bob the result. Unfortunately, for certain groups of states, her measurement leaves Bob with a set of possibilities that are still too similar to tell apart. The new protocol changes the strategy. Alice performs a very specific type of measurement that does not try to guess the answer immediately. Instead, she rearranges the information so that Bob's remaining possibilities become much more distinct from one another, even though the overall connection between their particles remains perfectly intact. She sends the result of this rearrangement to Bob.

Bob then uses this information to narrow the field down significantly. He can now reduce the list of possible states to either a single option or a pair of options. If it is a single option, he knows the answer and tells Alice. If it is a pair, he performs a further local measurement that distinguishes between the two, and then sends his finding back to Alice. Because the initial steps preserved the exact mathematical relationship between the states, Alice can now complete the job with certainty. The key insight is that Alice's first move preserves the "orthogonality"—the perfect distinctness—of the states while making the overlap between Bob's options very small. This small overlap is what allows Bob to make a decisive choice with just one more message.

The study proves that this method works for any set of states, provided the local dimension is large enough relative to the number of states. The researchers derived a precise mathematical condition showing that as long as the system size is big enough, this two-way conversation will always succeed. This result is significant because it establishes a clear, polynomial separation between what is possible with one-way communication and what is possible with two-way. It shows that the ability to distinguish these quantum states is not a fixed limit but a resource that grows with the size of the system when feedback is allowed.

Furthermore, the paper extends this finding to a practical scenario involving multiple copies of the same state. If Alice and Bob have several identical sets of these linked particles, they can process them together to improve their chances. The study shows that for sufficiently large systems, a universal number of copies—specifically nine—is enough to perfectly distinguish any set of orthogonal states, regardless of how large the system is or how many states are in the set. This is a powerful result because previous methods often required a number of copies that grew with the complexity of the problem. Here, the number of copies needed stays constant, independent of the system size.

The researchers also explored whether this protocol works for states that are not perfectly "maximally" entangled, meaning the link between the particles is slightly uneven. They found that the same two-way strategy works even in these imperfect cases, as long as the imbalance in the entanglement is not too extreme. This suggests the method is robust and applicable to a wider range of physical situations than just the ideal theoretical models. The work does not claim to solve every possible case in small systems, nor does it provide a fast, easy algorithm for building these measurements in a lab today. Instead, it provides a rigorous proof that such measurements exist and that the gap between one-way and two-way communication is vast and fundamental.

In the broader context of quantum information, this finding clarifies the power of classical communication in quantum tasks. It shows that the direction of communication is not just a minor detail but a critical factor that determines the limits of what can be known. By proving that a single feedback message can raise the distinguishable limit from a constant to a growing quantity, the study settles a long-standing question about the hierarchy of quantum operations. It confirms that the restriction imposed by one-way communication is not a fundamental law of nature but a limitation of the protocol, one that can be overcome with a simple, additional step. The result stands as a clear demonstration that in the quantum realm, the ability to listen as well as speak can fundamentally expand the horizon of what is knowable.

Technical Summary: Polynomial Separation between One-Way and Two-Way LOCC in Discrimination of Maximally Entangled States

Problem Statement
The paper addresses the local distinguishability of orthogonal maximally entangled states (MESs) in a bipartite system $\mathbb{C}^d \otimes \mathbb{C}^d$ . While global measurements can perfectly distinguish any set of orthogonal states, local operations and classical communication (LOCC) face fundamental restrictions. It is established that one-way LOCC (communication in a single direction, e.g., Alice to Bob) fails to distinguish certain sets of four orthogonal MESs for all local dimensions $d \ge 4$ . Consequently, the number of candidates distinguishable by one-way LOCC remains bounded by a constant (specifically, at most 3) regardless of how large the dimension $d$ grows.

The central question is whether allowing feedback (two-way LOCC) allows the number of distinguishable candidates to scale with the local dimension. Previous results showed that two-way LOCC could distinguish specific ensembles or relied on structural properties of generalized Bell sets, but no general scaling law existed for arbitrary sets of MESs.

Methodology
The author proposes a universal two-way LOCC protocol with a specific communication pattern: $A \to B \to A$ (Alice measures, sends a message to Bob; Bob measures and sends a feedback message; Alice completes the discrimination).

Alice's Initial Measurement (Sector Decomposition):
Alice performs a projective measurement $\{P_a^T\}$ on her subsystem. The goal is not to extract the label immediately (as the local reduced states are identical and maximally mixed) but to transform the problem. The measurement must satisfy two conditions for every pair of candidate unitaries $U_i, U_j$ (where $|\Psi_i\rangle = (I \otimes U_i)|\Phi_d\rangle$ ):
- Exact Orthogonality Preservation: The trace of the compressed relative unitary must vanish: $\text{Tr}(P_a A_{ij}) = 0$ , where $A_{ij} = U_i^\dagger U_j$ . This ensures the conditional states remain perfectly orthogonal.
- Weak Residual Coupling: The operator norm of the compressed relative unitary must be small: $\|P_a A_{ij} P_a\| \le \eta$ , where $\eta = 1/[4(k-1)]$ . This ensures that the conditional subspaces on Bob's side have small overlaps.
The existence of such a measurement is proven constructively using a recursive refinement argument. The author starts with a random sector decomposition (Haar random unitaries) which provides small operator compressions but only approximate trace cancellation. They then apply small collective rotations to the projectors to cancel the block traces exactly while preserving the compression bounds. This process is repeated recursively to achieve the required suppression of overlaps without incurring the full dimension cost in a single step.
Bob's Reduction (Singleton or Pair):
Upon receiving Alice's outcome $a$ , Bob's conditional states are supported on subspaces with small overlaps. Bob performs a Positive Operator-Valued Measure (POVM) designed to decompose the Gram matrix of the conditional states into positive terms supported on single labels or pairs of labels.
- The Gram matrix $G$ is shown to be strictly positive definite ( $G \succeq \frac{3}{4}I$ ) due to the small overlaps.
- $G$ is decomposed into singleton terms $X_i$ and pair terms $X_{ij}$ .
- Bob implements a measurement that yields either a "singleton" outcome (identifying the state uniquely) or a "pair" outcome (identifying the state is one of two specific candidates).
- Crucially, for pair outcomes, Bob performs the first step of a Walgate protocol (measuring first) to ensure the remaining conditional states for Alice are orthogonal.
Alice's Final Discrimination:
Bob sends the outcome (either the label $i$ or the pair $\{i, j\}$ plus his measurement result) to Alice. If a pair remains, Alice's conditional states are orthogonal, allowing her to distinguish them perfectly with a local measurement.

Key Contributions and Results

Theorem 1 (Universal Two-Way LOCC Discrimination): The paper proves that there exist universal constants $C, \nu > 0$ such that any set of $k$ pairwise orthogonal MESs in $\mathbb{C}^d \otimes \mathbb{C}^d$ is perfectly distinguishable by two-way LOCC (specifically the $A \to B \to A$ protocol) whenever:
$d \ge C k^4 [\log(2k)]^\nu$
This establishes a polynomial separation between one-way and two-way LOCC. While one-way LOCC is limited to a constant number of candidates ( $m_\to(d) \le 4$ ), two-way LOCC allows the number of distinguishable candidates to scale as $m_\leftrightarrow(d) = \Omega(d^{1/4} / (\log d)^{\nu/4})$ .
Corollary 1 (Universal Copy Bound): By applying the dimension bound to multiple copies (where $N$ copies effectively create a system of dimension $d^N$ ), the author shows that a universal number of copies suffices for perfect discrimination of any orthogonal MES set, independent of the set size $k$ and dimension $d$ . Specifically, for sufficiently large $d$ , nine copies are sufficient. This removes the basis restrictions found in previous results (e.g., for canonical Bell bases).
Extensions:
- The protocol extends to non-maximally entangled states with full Schmidt rank $d$ and bounded Schmidt coefficients (ratio $\kappa$ ), requiring $d \ge C \kappa^2 k^4 [\log(2\kappa k)]^\nu$ .
- The results provide lower bounds for Separable (SEP) and Positive Partial Transpose (PPT) measurements, though PPT bounds are superseded by existing stronger results.

Significance and Claims
The paper demonstrates that a single feedback message is sufficient to break the constant barrier of one-way LOCC, raising the capacity of local discrimination from a constant to a quantity scaling polynomially with the local dimension. The author emphasizes that this is a result regarding perfect discrimination of exact pure states. They explicitly state that:

The result does not imply efficient synthesis of the required measurements (the construction is an existence proof via convergent corrections).
The universal constants $C$ and $\nu$ have not been optimized or numerically evaluated.
The result does not address average error rates or stability under experimental noise.
The optimal growth rate of $m_\leftrightarrow(d)$ (e.g., whether it is linear in $d$ ) remains an open question.
While Theorem 1 settles the triple problem (distinguishability of three states) in all sufficiently large dimensions, it does not determine the remaining cases in smaller dimensions.

More like this