Fundamental Limits of Quantum Semantic Communication via Sheaf Cohomology

Christo Kurisummoottil Thomas and Mingzhe Chen

arXiv: 2601.10958v2 · PDF · 21 January 2026
Categories: cs.IT, cs.NI, math.IT
Licence: CC BY 4.0 The site shows the text.

Abstract

Semantic communication (SC) enables bandwidth-efficient coordination in multi-agent systems by transmitting meaning rather than raw bits. However, when agents employ heterogeneous sensing modalities and AI architectures, perfect bit-level transmission no longer guarantees mutual understanding. Although deep learning methods for semantic compression have advanced, the information-theoretic limits of semantic alignment under heterogeneity remain poorly understood. Notably, semantic ambiguity shares the same mathematical structure as quantum contextuality, as both arise from cohomological obstructions, motivating a quantum formulation of SC. In this paper, an information-theoretic framework for quantum semantic communication is proposed using sheaf cohomology. Multi-agent semantic networks are modeled as quantum sheaves, where agents meaning spaces are Hilbert spaces connected by quantum channels. The first sheaf cohomology group is shown to characterize irreducible semantic ambiguity, representing a fundamental obstruction to alignment that no local processing can resolve. The minimum communication rate required for semantic alignment is proven to scale with the logarithm of the dimension of the cohomological space, establishing a semantic analog of Shannon limits. For entanglement-assisted channels, the achievable capacity is shown to strictly exceed classical bounds, with each shared ebit reducing the required classical communication by one bit, providing a rigorous interpretation of shared context. Additionally, quantum contextuality is shown to reduce cohomological obstructions, and a duality between quantum discord and integrated semantic information is established, linking quantum correlations to irreducible semantic content. This framework provides rigorous foundations for quantum-enhanced semantic communication in autonomous multi-agent systems.

Items 27

Definition

Definition 1

X7F8

A quantum semantic sheaf over a communication graph G = (V,E) is a tuple (\mathcal{H}, \mathcal{F}, \rho) where:

40 \mathcal{H}_v is a finite-dimensional Hilbert space at each vertex v, representing agent v’s semantic state space.

40 \mathcal{F}_e : \mathcal{L}(\mathcal{H}_u) \to \mathcal{L}(\mathcal{H}_v) is a completely positive trace-preserving (CPTP) map for each edge e \!=\! (u,v), representing the semantic channel from agent u to v.

40 \rho_v is a density operator on \mathcal{H}_v representing agent v’s current semantic state. A density operator is a positive semi-definite matrix with \mathrm{Tr}(\rho) = 1 , generalizing probability distributions to quantum systems [Nielsen2010].

Used by Proposition 1, Proposition 3, Theorem 1, Theorem 2, Definition 5, Theorem 3

Definition

Definition 2

EZXU

A local section over a subset of agents U \subseteq V is a semantic configuration \sigma such that \mathcal{F}_e(\sigma_u) = \sigma_v, for all edges e = (u,v) with both endpoints in U. A global section is a local section over the entire vertex set V.

Used by Definition 3, Proposition 1, Proof

Definition

Definition 3

H5NW

The sheaf Laplacian is defined as

\mathcal{L} = (\delta^0)^\dagger \delta^0, \vspace{-1mm}

where \dagger denotes the adjoint operator with respect to the Hilbert–Schmidt inner product \langle A, B \rangle_{\text{HS}} = \text{Tr}(A^\dagger B). The adjoint (\delta^0)^\dagger is the unique linear operator satisfying the duality condition: \langle (\delta^0)^\dagger x, y \rangle_{\text{HS}} = \langle x, \delta^0 y \rangle_{\text{HS}}, for all semantic configurations x, y. The operator \delta^0 maps vertex assignments to edge disagreements; (\delta^0)^\dagger reverses this flow by aggregating edge errors back to vertices. The kernel of \mathcal{L} consists precisely of the global sections:

\ker(\mathcal{L}) = \{\sigma : F_e(\sigma_u) = \sigma_v \text{ for all edges } e = (u,v)\}. \vspace{-1mm}

Uses Definition 2

Proposition

Proposition 1

M783

A quantum semantic sheaf admits a global section for every locally consistent configuration if and only if H^1(G, \mathcal{S}) = 0.

Uses Definition 1, Definition 2

Proof of Proposition 1

This follows from the standard characterization of sheaf cohomology [Abramsky2012CohomologyNonlocality], which extends to our quantum setting since CPTP maps preserve the algebraic structure required for the Čech complex construction.

Definition

Definition 4

TXSZ

An entanglement-assisted quantum semantic sheaf augments Definition 1 with a bipartite entangled state \Phi_e \in \mathcal{D}(\mathcal{H}'_u \otimes \mathcal{H}'_v) for each edge e = (u, v), where \mathcal{H}'_u and \mathcal{H}'_v are local ancilla spaces held by agents u and v respectively. The state \Phi_e is maximally entangled if it achieves maximum entanglement entropy of \log_2 d ebits for d-dimensional systems (e.g., Bell states for qubits).

Used by Proposition 2, Proof, Theorem 2

Proposition

Proposition 2

XJ9I

For an entanglement-assisted semantic sheaf with maximally entangled states on each edge:

\dim H^{1}_{\mathrm{EA}}(G,S) \;\le\; \dim H^{1}(G,S), \vspace{-1mm}

with strict inequality for generic sheaves over graphs with cycles.

Uses Definition 4

Proof

RWQS

Maximally entangled states enable superdense coding, effectively doubling the classical capacity of each edge. This additional capacity can eliminate cocycles in H^1 that arise from holonomy obstructions around loops.

Uses Definition 4

Proposition

Proposition 3

Q78Y

Let \mathcal{M} be a quantum measurement scenario with sheaf \mathcal{S}_\mathcal{M} whose vertices correspond to measurement contexts and whose stalks assign outcome probabilities. There exists a quantum semantic sheaf \mathcal{S} such that H^1(\mathcal{S}) \cong H^1(\mathcal{S}_\mathcal{M}). Contextual measurement scenarios correspond precisely to semantic sheaves with irreducible ambiguity.

Uses Definition 1

Theorem

Theorem 1

8TOI

For a quantum semantic sheaf \mathcal{S} over graph G, the minimum communication rate required for perfect semantic alignment is:

R^*_{\mathrm{sem}} = \log_2 \dim H^1(G, \mathcal{S}) \vspace{-1mm}

measured in qubits (or bits for classical sheaves).

Uses Definition 1 · Used by Proof, Proof

Proof of Theorem 1

See Appendix A of extended version in arxiv [ThomasISIT2026].

Corollary

Corollary 1

O8Y2

The semantic capacity of a classical channel with capacity C is:

C_{\mathrm{sem}} = \frac{C}{\log_2\dim H^1} \vspace{-1mm}

semantic messages per channel use (when H^1 \neq 0).

Proof

YU3X

Each semantic alignment requires R^{\ast} = \log_2 \,\text{dim}\, H^1 bits (Theorem 1). A channel with capacity C transmits C bits per use. Therefore, C_{\mathrm{sem}} =\frac{C}{R^{\ast}} = \frac{C}{\log_2\dim H^1}.

Uses Theorem 1

Theorem

Theorem 2

ZETV

For a quantum semantic sheaf \mathcal{S} over graph G = (V, E) with entanglement-assisted sheaf \mathcal{S}_{EA}, where edge e shares entangled state \Phi_e with Schmidt rank r_e. The Schmidt rank is the number of terms in the Schmidt decomposition of a bipartite state, quantifying its entanglement. A product state has r=1; a maximally entangled d-dimensional state has r=d:

\dim H^1_{EA}(G, \mathcal{S}) = \max\left(0, \dim H^1(G, \mathcal{S}) - \sum_{e \in E} \log_2 r_e\right). \vspace{-1mm}

For separable (classical) correlations where r_e = 1 for all e, no reduction occurs: \dim H^1_{EA} = \dim H^1.

Uses Definition 1, Definition 4

Proof

NVLT

See Appendix B of extended version in arxiv [ThomasISIT2026].

Definition

Definition 5

EQPP

A quantum semantic sheaf \mathcal{S} exhibits contextual correlations if there exists no classical, i.e., local hidden variable model that reproduces the joint distributions over all semantic measurements. Formally, \mathcal{S} is contextual if H^1(G, \mathcal{S}) \neq 0 when restricted to any classical sub-sheaf.

Uses Definition 1 · Used by Theorem 3, Proof

Theorem

Theorem 3

7G8C

Let \mathcal{S}_C be a classical semantic sheaf and \mathcal{S}_Q be a quantum semantic sheaf over the same graph G, where \mathcal{S}_Q admits contextual correlations. Then:

\dim H^1(G, \mathcal{S}_Q) < \dim H^1(G, \mathcal{S}_C) \vspace{-2mm}

whenever the contextual fraction is nonzero.

Uses Definition 1, Definition 5 · Used by Proof

Proof

0KGR

For classical sheaves, the gluing axiom implies \ker(\delta^1) = \mathrm{im}(\delta^0), hence H^1 = 0 when correlations are non-contextual. For quantum sheaves with contextual correlations, let \{c_i\} be independent contextual witnesses. Each c_i eliminates one generator of H^1: \dim H^1(S_Q) = \dim H^1(S_C) - |\{c_i\}|. The contextual fraction \mathrm{CF}(e)\! >\! 0 guarantees |\{c_i\}| \geq 1.

Uses Definition 5

Proposition

Proposition 4

7YK8

The semantic rate reduction due to contextuality satisfies: R_{\text{classical}} - R_{\text{quantum}} \geq CF(e) · \log_2\, \text{dim}\, H^1_{\text{classical}}.

Proof

LFZG

From Theorem 3, contextuality reduces \dim H^1. The contextual fraction \mathrm{CF}(e) quantifies what fraction of classical correlations are replaced by contextual ones. Since each generator of H^1 eliminated saves \log_2 \dim H^1_{\mathrm{classical}} bits of communication (by Theorem 1), the bound follows.

Uses Theorem 3, Theorem 1

Definition

Definition 6

PTNL

For a semantic state \rho_{AB} over a bipartite semantic sheaf for any two nodes in G, the integrated semantic information is:

\Phi_{\mathrm{sem}}(A:B) = I(A:B) - \max_{\mathrm{partitions}} \sum_i I(A_i:B_i) \vspace{-2mm}

where the maximization is over all partitions of A and B into independent subsystems.

Theorem

Theorem 4

7GE2

For bipartite quantum semantic states: \delta(A:B) = \phi_{\text{sem}}(A:B).

Proof

187D

We show the equality by demonstrating that both quantities measure the same irreducible correlations from dual perspectives. The quantum mutual information decomposes as:

I(A:B) = J(A:B) + \delta(A:B),

where J(A:B) = \max_{M_B}[S(\rho_A) - S(\rho_A|M_B)] is the classical correlation extractable by measuring subsystem B, and \delta(A :B) is the quantum discord. The classical correlation J(A:B) represents the maximum information about A obtainable through local measurements on B. This corresponds precisely to \max_{\mathrm{partitions}} \sum_i I(A_i :B_i) in the definition of \Phi_{\mathrm{sem}}, since the optimal measurement on B induces the partition that maximizes extractable local information. Substituting into the definition of \Phi_{\mathrm{sem}}:

\begin{aligned} \Phi_{\mathrm{sem}}(A :B) &= I(A:B) - \max_{\mathrm{partitions}} \sum_i I(A_i:B_i) \\ &= I(A:B) - J(A:B) = \delta(A:B). \end{aligned} \vspace{-1mm}

Thus both quantities answer the same question from dual perspectives: discord asks “what correlations survive any local measurement?" while integrated information asks “what information cannot be decomposed into parts?" The equality holds because the optimal measurement basis for extracting classical correlations coincides with the minimum information partition.

Lemma

Lemma 1

I3TF

There exists a semantic alignment protocol using exactly k = \dim H^1(G, \mathcal{S}) transmitted symbols.

Proof of Lemma 1

Let k = \dim H^1(G, \mathcal{S}) and choose a basis \{[\omega_1], \ldots, [\omega_k]\} for H^1, where each \omega_i \in \ker(\delta^1) is a representative cocycle. Given any locally consistent configuration \omega \in \ker(\delta^1), its cohomology class can be uniquely expressed as:

[\omega] = \sum_{i=1}^{k} c_i [\omega_i], \quad c_i \in \mathbb{C}

The encoder (e.g., a designated coordinator) transmits the coefficient vector \mathbf{c} = (c_1, \ldots, c_k) \in \mathbb{C}^k to all agents that require alignment. The receiver reconstructs \tilde{\omega} = \sum_{i=1}^{k} c_i \omega_i. Since [\omega] = [\tilde{\omega}] under noiseless transmission, we have \omega - \tilde{\omega} \in \mathrm{im}(\delta^0). Thus there exists \sigma \in C^0 with \delta^0(\sigma) = \omega - \tilde{\omega}. The receiver solves this linear system (guaranteed to have a solution by construction) to obtain the global section \sigma achieving semantic alignment. Here, the protocol transmits exactly k = \dim H^1 symbols.

Uses Definition 2

Lemma

Lemma 2

HF9M

Any semantic alignment protocol requires at least k = \dim H^1(G, \mathcal{S}) transmitted symbols.

Proof of Lemma 2

Suppose, for contradiction, that a protocol achieves perfect semantic alignment using r < k transmitted symbols. The encoder is then a function \mathrm{Enc}: \ker(\delta^1) \to \mathbb{C}^r mapping locally consistent configurations to r-symbol codewords. For perfect alignment, configurations with different cohomology classes must map to different codewords. If [\omega] \neq [\omega'] in H^1, then \mathrm{Enc}(\omega) \neq \mathrm{Enc}(\omega'); otherwise the decoder cannot distinguish them. Consider the induced map:

\overline{\mathrm{Enc}}: H^1 \to \mathbb{C}^r, \quad [\omega] \mapsto \mathrm{Enc}(\omega)

This is well-defined: if [\omega] = [\omega'], then \omega - \omega' = \delta^0(\sigma) for some \sigma, and both configurations lead to the same aligned state without communication. For perfect alignment, \overline{\mathrm{Enc}} must be injective. But an injective linear map from a k-dimensional space to an r-dimensional space requires r \geq k. Since r < k by assumption, \overline{\mathrm{Enc}} cannot be injective—there exist [\omega] \neq [\omega'] with identical codewords, making alignment impossible.

Source and licence

The text on this page is from arXiv:2601.10958v2 by Christo Kurisummoottil Thomas and Mingzhe Chen, under the licence CC BY 4.0. The corpus made these changes: it read the items out of the LaTeX source, it expanded the macros of the paper, it put each proof after its statement, and it converted the text to HTML. The formulas are the TeX of the paper, and KaTeX renders them. A formula or a reference can look different from the PDF. The page is published under CC BY 4.0.

The record of each item is in analysis/objects/2601/2601.10958.jsonl.