⚛️ quantum physics

Exact certification of a positive-order Rényi additivity violation for an explicit channel pair

This paper provides the first rigorous, computer-verifiable certification of a strict positive-order Rényi additivity violation for the explicit quantum channel pair originally proposed by Cubitt et al., establishing that the violation holds for all orders $0 < p \le 1/22$ through a complete proof based on small rational witness matrices and elementary interval arguments.

Original authors: Artus Krohn-Grimberghe

Published 2026-08-19

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Artus Krohn-Grimberghe

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 and counterintuitive world of quantum information, scientists study how much information can be packed into the smallest possible containers. A central question for decades has been whether combining two separate quantum systems allows them to hold more information together than the sum of what they could hold individually. For a long time, researchers believed that information capacity was strictly additive: two systems together would simply equal the sum of their parts. However, this belief was shattered when it was discovered that for certain types of quantum channels, or pathways for information, the combined system can actually be more efficient than the simple sum suggests. This phenomenon, known as a violation of additivity, implies that entangled inputs—where the two systems are linked in a way that has no classical equivalent—can unlock hidden storage potential. While this violation was proven to exist at a specific mathematical limit, a lingering uncertainty remained for the range of values just above that limit, where numerical simulations hinted at a violation but lacked a rigorous, unshakeable proof.

A new study by Artus Krohn-Grimberghe of Percivio Ltd. finally closes this gap for a specific, well-known pair of quantum channels. The researcher provides a completely verified, computer-checkable proof that these two channels violate the additivity rule for a continuous range of values, extending from zero up to a specific point. Unlike previous work that relied on numerical approximations or suggested violations without a hard mathematical boundary, this paper delivers a certificate of truth. It uses a set of small, rational numbers and matrices that anyone with a basic computer program can run to verify the result instantly. The proof confirms that for every real number order up to one twenty-second, the combined system of these two specific channels produces less entropy—a measure of disorder or uncertainty—than the sum of their individual minimums. This means the channels are indeed more efficient together than apart, and this fact is now established with absolute mathematical certainty rather than just strong numerical evidence.

The work focuses on a pair of channels originally identified by a researcher known as CHLMW. These channels were constructed from specific geometric subspaces, and while they were known to break the additivity rule at a single point, their behavior just above that point was a mystery. The new paper takes the exact same pair of channels and applies a rigorous method to prove they continue to break the rule for a small but significant interval. The proof relies on three concrete facts. First, the researcher established strict upper and lower limits for the energy levels, or eigenvalues, of the output from each channel. Second, they identified a single, specific entangled input that produces a joint output with a precise, rational spectrum of eight distinct values. Third, they used two independent mathematical arguments to show that the combined output of this specific input is strictly more ordered than the best possible separate outputs of the individual channels.

The verification process is designed to be transparent and foolproof. Every step of the calculation reduces to comparing whole numbers, eliminating any need for floating-point approximations that could introduce error. The researcher used an AI-assisted search to find the specific matrices needed to set the bounds, but the proof itself does not depend on the search. Instead, it depends entirely on the printed certificate: a few small matrices and a list of rational numbers. Two separate computer programs, written independently, checked every claim. One program worked with exact mathematical fields, while the other reconstructed the data from the paper's printed text and enclosed every value in safe intervals. Both programs confirmed that the inequalities hold true and that the specific input produces the predicted result. This dual verification ensures that the conclusion is not an artifact of a single software bug or a numerical glitch.

The study is careful to define its own limits. The researcher does not claim to have found the absolute maximum point where the violation stops, nor do they claim to have found the perfect input that minimizes entropy for all cases. The proof establishes a guaranteed interval where the violation is certain, but it acknowledges that the violation likely persists further, perhaps up to a value of roughly 0.11 as suggested by earlier numerical work. The current method hits a wall at one twenty-second because the bounds used in the proof are not sharp enough to go further without new mathematical tools. Nevertheless, the achievement is significant because it provides the first printed, independently verifiable endpoint for this famous example. It transforms a result that was once a "numerical suggestion" into a "rigorous fact," giving the scientific community a solid foundation to build upon.

This work stands as a testament to the power of exact verification in an era where complex calculations often rely on black-box simulations. By reducing the problem to a series of integer comparisons and providing the raw data for anyone to check, the paper removes doubt from a long-standing question in quantum information theory. It confirms that for this explicit pair of channels, the quantum world continues to defy classical intuition, offering a combined efficiency that is strictly greater than the sum of its parts, and it does so with a level of certainty that can be checked by hand or by a simple script. The result is a clear, unambiguous confirmation that the additivity violation is not just a numerical curiosity, but a robust mathematical reality within a defined range.

Technical Summary: Exact Certification of a Positive-Order Rényi Additivity Violation

Problem Statement
The paper addresses the additivity of the minimum output Rényi entropy for quantum channels, defined as $S_{\min}^p(\mathcal{N}) = \min_\rho S_p(\mathcal{N}(\rho))$ . While the additivity conjecture $S_{\min}^p(\mathcal{N}_1 \otimes \mathcal{N}_2) = S_{\min}^p(\mathcal{N}_1) + S_{\min}^p(\mathcal{N}_2)$ was known to fail for many orders, a specific gap remained for the explicit channel pair constructed by Cubitt, Harrow, Leung, Montanaro, and Winter (CHLMW). CHLMW demonstrated a violation at order $p=0$ and reported numerical evidence of violations for positive orders up to approximately $p \approx 0.11$ . However, as noted by Leung, Lovitz, and Wu, no rigorous, verifiable endpoint for this positive-order violation had been established for this explicit pair; previous claims relied on semidefinite programming arguments without printed endpoints or certificates.

Methodology
The author provides a rigorous, computer-verifiable proof that the CHLMW channel pair violates additivity for all real orders $0 < p \leq 1/22$ . The methodology relies on elementary linear algebra and interval arithmetic, avoiding complex optimization solvers in the final verification step. The proof structure consists of three main components:

Eigenvalue Bounds (Floor and Cap):
The author establishes strict rational bounds on the eigenvalues of the output states for the individual channels. Using "witness matrices" (small rational Hermitian matrices), they prove that for any input state $\rho$ , every eigenvalue $\lambda$ of the output $\mathcal{N}_T(\rho)$ (where $T \in \{R, \bar{S}\}$ ) lies within the interval $[\ell, u]$ , where:
- $\ell = 301/100000$
- $u = 2/3$
  These bounds are derived via a decomposition lemma involving partial transposes, ensuring that the output spectrum is uniformly bounded away from zero and one.
Single-Channel Lower Bound:
Using the established eigenvalue bounds and the concavity of the function $f(t) = t^p$ for $0 < p < 1$ , the author derives a lower bound for the trace of the output power, $\text{Tr}(\mathcal{N}_T(\rho)^p)$ . They define an "extreme admissible spectrum" $e = (u, 1-u-\ell, \ell)$ and prove that for any valid output spectrum $\lambda$ , $\sum \lambda_i^p \geq \sum e_i^p$ . This yields a lower bound $A(p)$ for the single-channel entropy.
Joint Input Upper Bound:
The author constructs a single explicit entangled input state $\Psi$ (derived from the inverse-input-marginal normalization of the CHLMW subspaces). They compute the exact rational spectrum of the joint output $(\mathcal{N}_R \otimes \mathcal{N}_{\bar{S}})(\Psi\Psi^*)$ , which has rank eight. This provides an upper bound $B(p)$ for the joint minimum output entropy.

Key Results
The core result is the proof of the strict inequality:
$S_{\min}^p(\mathcal{N}_R \otimes \mathcal{N}_{\bar{S}}) < S_{\min}^p(\mathcal{N}_R) + S_{\min}^p(\mathcal{N}_{\bar{S}})$
for all $0 < p \leq 1/22$ .

This is achieved by demonstrating that $A(p)^2 > B(p)$ over the interval. The verification of this transcendental inequality is reduced to finite comparisons of integers. The author employs two independent arguments to certify this interval:

The Staircase Argument: Partitioning the interval $[0, 1/22]$ into rational sub-intervals and using the monotonicity of $x^p$ to bound the functions $A(p)$ and $B(p)$ at the endpoints.
The Derivative Argument: Enclosing the function $D(p) = A(p)^2 - B(p)$ and its derivative in rational intervals to prove $D(p)$ is positive and decreasing throughout the interval.

Contributions and Claims
The paper explicitly limits its novelty to the certification and explicit constants rather than the underlying phenomenon or the eigenvalue-floor mechanism, both of which are attributed to CHLMW. The specific contributions are:

First Rigorous Endpoint: Providing the first printed, computer-verifiable, and rigorously certified positive-order endpoint ( $p = 1/22$ ) for the CHLMW explicit pair.
Explicit Certificates: Supplying small rational witness matrices and a rational spectrum that allow any reader to verify the proof using a short program or, for a single order, by hand.
Independence: The proof relies on exact arithmetic (integer comparisons) rather than numerical approximations. The witness matrices were found via AI-assisted search, but the search plays no role in the proof itself; only the resulting certificates are used.

Significance
The paper closes a specific gap in the literature regarding the CHLMW example, moving from "numerical evidence" and "unprinted semidefinite arguments" to a fully explicit, checkable proof. The author notes that the endpoint $1/22$ is not claimed to be optimal; the method is limited by the current certified floor and cap values, and extending the interval toward the numerically observed $0.11$ would require sharper bounds rather than finer interval partitioning. The work serves as a template for rigorous certification of quantum information theoretic inequalities using elementary, verifiable certificates.

Technical Summary: Exact Certification of a Positive-Order Rényi Additivity Violation

More like this