Superadditivity of Zero-Error Capacity in Noisy Classical and Perfect Quantum Channel Pairs

Ambuj, Anushko Chattopadhyay, Kunika Agarwal, Rakesh Das, Amit Mukherjee · 2026

Read Superadditivity of Zero-Error Capacity in Noisy Classical and Perfect Quantum Channel Pairs by Ambuj, Anushko Chattopadhyay, Kunika Agarwal, Rakesh Das, Amit Mukherjee (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-08913.txt.

Introduction

Ambuj and four co-authors show a gain when a noisy plain link and a pure quantum link are used as a pair. Each link on its own can send a fixed count of notes with no flaw at all. One would guess the pair can send at most the product of those counts. The team shows that guess is wrong, in a good way.

Used as a pair, the links can send more flaw free notes than the product allows. The gain is called superadditivity. It means joint use beats the sum of parts in a strict sense. A reader in comms work gets a clean proof of that gain. It hints that mixed plain and quantum links may carry more than we thought.

Catalog record

CREATORAmbuj, Anushko Chattopadhyay, Kunika Agarwal, Rakesh Das, Amit Mukherjee
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE34 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-08913

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Superadditivity of Zero-Error Capacity in Noisy Classical and Perfect Quantum Channel Pairs Ambuj, Anushko Chattopadhyay, Kunika Agarwal, Rakesh Das, Amit Mukherjee Abstract We demonstrate superadditivity of one-shot zero-error classical capacity in an asymmetric communication setting where a noisy classical channel is used in parallel with a perfect quantum channel. Each channel individually supports only a fixed number of perfectly distinguishable messages. Their joint use enables transmission of strictly more messages than permitted by the product of the individual capacities. We present explicit constructions achieving this enhancement and establish that replacing the perfect quantum channel with a perfect classical channel eliminates the effect. Finally, we identify a structural criterion on the noisy channel governing this effect and show that the quantum advantage is rooted in Kochen–Specker contextuality. Introduction.- Can the combined performance of two imperfect resources exceed the sum of what each can achieve individually. The origin of this philosophy traces back to Aristotle Aristotle 1924 . In information theory, this question takes a more concrete form in the study of communication channels, where performance is quantified by channel capacity Shannon 1948 . When two noisy channels are used together, their joint capacity is generally expected to equal the sum of their individual capacities. Remarkably, quantum information theory violates this intuition. Certain pairs of noisy quantum channels exhibit superadditivity, whereby their combined capacity exceeds the sum of the capacities achievable in isolation. A quantum channel supports the transmission of quantum information, private information, or classical information, leading to distinct notions of capacity Schumacher and Westmoreland 1997 ; Holevo 1998 ; Lloyd 1997 ; Devetak 2005 ; Shor 2002 . Superadditivity has been rigorously established for quantum and private capacities Li et al. 2009 ; Smith and Smolin 2009 ; Smith and Yard 2008 ; Smith et al. 2011 , and indications of a similar phenomenon for classical capacity have also been reported Hastings 2009 . Zero-error information theory, introduced by Shannon Shannon 1956 , addresses communication under the stringent requirement of absolute certainty, allowing no probability of error. Within this setup, interestingly even noisy classical channels can display superadditive behavior Shannon 1956 ; Lovasz 1979 ; Alon 1998 ; Korner and Orlitsky 1998 . More recently, quantum extension of zero-error information theory has emerged as an active area of research Cubitt et al. 2010 ; Duan 2009 ; Cubitt et al. 2011a ; Chen et al. 2010 ; Beigi 2010 ; Cubitt et al. 2011b ; Chen et al. 2011 ; Duan et al. 2013 ; Duan et al. 2016 ; Wang and Duan 2018 ; Prevedel et al. 2011 ; Yadavalli and Kunjwal 2022 ; Alimuddin et al. 2023 ; Agarwal et al. 2024 . In this work, we present a previously unexplored instance of superadditivity phenomenon having a quantum origin. Specifically, we show that a noisy classical channel, when paired in parallel with a perfect quantum channel, exhibits superadditive capacity. Namely, the capacity of the joint noisy classical and perfect quantum channels supersedes sum of their individual capacities. We also establish that this enhancement is fundamentally quantum, i.e. , replacing the perfect quantum channel with a perfect classical channel eliminates the superadditive advantage. Along the way, we derive general criteria characterizing when such superadditive behavior can occur, thereby expanding the conceptual landscape of zero-error communication theory and connecting it to the foundational notion of Kochen Specker contextuality Kochen and Specker 1975 ; Budroni et al. 2022 . Framework.- A discrete memoryless classical channel N is specified by a finite input alphabet \mathcal{X} , a finite output alphabet \mathcal{Y} , and a collection of conditional probabilities P(y|x) , where \sum_{y\in\mathcal{Y}}P(y|x)=1 for all x\in\mathcal{X} , and y\in\mathcal{Y} . For a fixed input symbol x\in\mathcal{X} , only a subset of output symbols can occur with nonzero probability Shannon 1948 . We denote this support of input by \Gamma(x):=\{y\in\mathcal{Y}\mid P(y|x)>0\} , Operationally, \Gamma(x) represents all outputs that the receiver may observe when x is transmitted in a single use of the channel. Two distinct input symbols x,x^{\prime}\in\mathcal{X} are said to be confusable in single channel use if their supports in \mathcal{Y} overlap, that is, if \Gamma(x)\cap\Gamma(x^{\prime})\neq\varnothing Shannon 1956 . In this case, there exists at least one output that could have arisen from either input, and therefore the receiver cannot infer with certainty which input was sent. Conversely, two input symbols are perfectly distinguishable in a single channel use if their supports are disjoint. A one-shot zero-error code for the channel is a subset \mathcal{C}\subseteq\mathcal{X} such that every pair of distinct input symbols in \mathcal{C} is perfectly distinguishable by the receiver. Equivalently, for all x\neq x^{\prime} in \mathcal{C} , we require \Gamma(x)\cap\Gamma(x^{\prime})=\varnothing . The size of such a code, |\mathcal{C}| represents the number of classical messages that can be transmitted with absolute certainty using a single invocation of the channel. The confusability relations among input symbols admit a natural graph-theoretic representation. The confusability graph G(N) of the channel has vertex set V(G) which coincides with input set \mathcal{X} . Two distinct vertices are connected by an edge if and only if the corresponding input symbols are confusable and the set of all edges is denoted by E(G) . By construction, a one-shot zero-error code \mathcal{C} corresponds precisely to a set of vertices with no edges between them. In graph-theoretic terms, a code \mathcal{C} is an independent set of the graph G(N) . The one-shot zero-error classical capacity \mathfrak{C}_{0}(N) of the channel is therefore given by the independence number \alpha(G(N)) , defined as the maximum cardinality of an independent set of G(N) . Although capacity is often defined as the logarithm of this quantity, throughout this work we identify the one-shot zero-error capacity with \alpha(G(N)) itself, corresponding directly to the maximum number of messages that can be transmitted with certainty in a single channel use. As a simple example, a perfect classical channel with d input symbols has an edgeless confusability graph ( \overline{K}_{d} ) and hence \mathfrak{C}_{0}(\overline{K}_{d})=\alpha(\overline{K}_{d})=d . When two classical channels N and N^{\prime} are used in parallel, denoted as N\otimes N^{\prime} , the receiver receives both outputs simultaneously. Two input pairs are confusable if the receiver cannot distinguish at least one component of the pair in the corresponding channel perfectly. This joint confusability graph corresponds precisely to the strong product of the individual confusability graphs, denoted G(N)\boxtimes G(N^{\prime}) . Two distinct vertices (u,i) and (v,j) in G(N)\boxtimes G(N^{\prime}) are adjacent if and only if at least one of the following holds: (i) u=v and i is adjacent to j in G(N^{\prime}) ; (ii) u is adjacent to v in G(N) and i=j ; (iii) u is adjacent to v in G(N) and i is adjacent to j in G(N^{\prime}) . In this framework, a one-shot zero-error code for the parallel channel corresponds to an independent set in the strong product graph G(N)\boxtimes G(N^{\prime}) . Consequently, the joint one-shot zero-error capacity is given by \mathfrak{C}_{0}(N\otimes N^{\prime})=\alpha\bigl(G(N)\boxtimes G(N^{\prime})\bigr) . A noisy classical channel N can equivalently be represented by its channel hypergraph, denoted as H(N) . A vertex set of H(N) is the input alphabet \mathcal{X} of the channel. The output symbol y\in\mathcal{Y} , is a hyperedge h_{y} which consists of vertex subset \{x\in\mathcal{X}\mid P(y|x)>0\}\subseteq\mathcal{X} . Two inputs are confusable if and only

Read the complete text