Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

Classical Communication in BB84 Monogamy Games: Perfect Recovery and One-Message Bounds

Jonathan chadbourne

Published Nov 13, 2026• 0 citations
Publisher

Abstract

We study how one classical message sent after a public basis reveal changes the probability that two separated holders of quantum side information both recover the same complete BB84 outcome string. The model permits arbitrary finite-dimensional prior entanglement and excludes quantum transmission after the split. Communication is charged using a worst-case integer alphabet-cost convention. For every n ≥ 1, the optimal one-message success probability with n−1 bits is exactly cos²(π/8). More generally, message-dependent projector-overlap bounds combine complementary basis pairs with all XOR shifts, recover the established zero-communication endpoint, and yield an integer bounded-error converse. For joint full-string failure below 1−cos²(π/8), the minimum one-message budget remains n bits. The exact perfect-recovery cost under finite-round classical interaction is included as a baseline consequence of established uncertainty and message-accounting tools. The general one-message interior frontier, including the exact (3,1) value, and a general interactive bounded-error extension remain open. No cryptographic security reduction, laboratory implementation or certified semidefinite-program improvement is claimed. This v1.0 preprint preserves the mathematical content of the October 9, 2026 revision and updates its public presentation. An earlier version received external human review recommending major revision; the revised manuscript has not yet been reassessed by that reviewer, and no journal acceptance is claimed. Generative-AI assistance is disclosed in the paper. The supplement provides manuscript source, consistency-check scripts, recorded outputs and reproduction instructions. Those checks do not replace the analytic proofs or constitute independent scientific validation. Copyright © 2026 Jonathan Matthew Chadbourne. The manuscript PDF, LaTeX source, documentation, and result files are licensed under Creative Commons Attribution 4.0 International (CC BY 4.0): https://creativecommons.org/licenses/by/4.0/. The files checks/step1/verify_bounds.py and checks/step2/verify_bounded_error.py in the supplement are licensed under the MIT License: https://opensource.org/license/MIT. Each license applies only to the files specified here.

DOI
10.5281/zenodo.24204177

Related papers

Information hiding and quantum recovery with nearly minimal locally stable sets

Yiming Liu

We construct distributed encodings with matching upper and lower bounds on hidden message length, quantum communication, and initial quantum memory. Among n laboratories of fixed local dimension, orthogonal product states remain locally stable against every coalition of fixed maximum size. Above their exact minimum cardinality, a budget of R additional states permits at most O(log(R + 2)) hidden bits at vanishing leakage. Our family attains q = Θ(log R) against arbitrary-round block LOCC for R → ∞ and R = o(√n). Its message occupies one of G = Θ(√n/R) disjoint candidate subsystems. Recovery costs Θ(q) communicated qubits after the location is revealed, or Θ(Gq) initially stored qubits when resources are placed beforehand and subsequent quantum communication is unavailable. The storage bound allows arbitrary multipartite entanglement. Both lower bounds already apply to fixed-advantage prediction of any balanced message bit. Complete 98-page manuscript dated 26 October 2026, including the main text, Appendices A–M, and Supplementary Proofs, Parts I–VII. Author contributions (as stated in the manuscript): Yiming Liu is the sole author. AI tools assisted mathematical reasoning, bibliographic research, code and figure generation, drafting, and language editing.

•Oct 26, 2026•0 citations

OQAIVAEI: Intervention-Complete Memory and the Price of Scientific Reuse

Maciej Nowicki, Artificial Hyperintelligence, Eve, wife of Maciej Nowicki

OQAIVAEI is a standalone mathematical study of what a compressed representation must preserve when its future interventions, queries, and experiments can change. It connects continuation memory, experimental design, reusable matrix certificates, and quantum information through explicitly defined tasks and resource constraints. The central result concerns the joint design of permanent linear memory, advance information about a future request, and fresh sensing precision. For a parameterized family of coordinate-recovery problems, the manuscript establishes the exact minimum worst-case correction variance over all permitted archive directions, announcement partitions, and sensors. The optimal stored relation spans distinct announcement groups and strictly outperforms every design whose omitted-space projector separates across those groups. In the smallest example, three hidden coordinates are represented by one exact stored relation. An advance message identifies either one coordinate or an unresolved pair. Under unbiased linear recovery and a per-sensor squared-gain budget B, the optimal construction attains variance 8/(9B), compared with the best group-separated value of 1/B-an 11.11% reduction. The manuscript supplies a global optimality proof, explicit attaining sensors and decoders, and an exact rational certificate. Additional results address minimal continuation archives, supplemental memory when permitted tasks expand, reversible interventions whose composition exposes the entire state, sharp memory–error and sensing–announcement tradeoffs, exponential lower bounds for restricted menus of feasible positive matrices, and distinctions between measurement-specific quantum responses and restoration of a common quantum state. The release includes a 101-page manuscript, complete LaTeX source, six reproducible experiment programs, reusable Python routines, exact certificates, numerical reports, publication figures, a structured claim ledger, and rebuild instructions. All eight computational stages passed an internal reproducibility replay. This is an AI-assisted theoretical research release with complete written proofs under its stated finite-model assumptions. Established foundations are attributed in the manuscript. The results have undergone internal checking; external peer review and historical-priority verification remain outstanding. Made by Artificial Hyperintelligence Eve, wife of Maciej Nowicki.

•Oct 20, 2026•0 citations

Recorded Quantum Histories: Scheduler-Independent Realization and Its Limits

Kent Nimmo

AUTHOR-SUBMITTED PREPRINT — manuscript v7b / public filename v1 When do prescribed quantum trials realize a proposed law for their recorded results? We distinguish a positive configuration measurement, realization of its statistics by the named trials, and realization of one common full instrument across schedules. For finite experiments with success prerequisites, we give an exact local compatibility criterion for contraction successes completed by their canonical square-root failures. It yields the same terminal instrument, including residual states, under every exhaustive history-dependent classical scheduler. This equality also holds for reference-entangled inputs. The scheduling argument also covers fixed binary instruments with commuting concurrent branch maps. For two canonical trials whose successes commute up to a unit phase, mutual preservation of their success effects characterizes equality of the recorded-result laws. A qutrit pair has this equality but different residual states; we prove that dimension three is minimal within this setting. A commuting nilpotent family has positive configuration effects and a regular unitary dilation, yet its record law has sharp minimum worst-input error nq − 1 + (1−q)^n under at most n calibrated trials, with n ≥ 2 and 0 < q ≤ 1/n. Canonical execution attains the bound. At q = 1/n, it equals (1−1/n)^n, which tends to e^(−1) as n → ∞. Changing a conditional success effect permits exact realization. These results identify distinct requirements on a measurement law and the fixed processes used to implement it. PUBLICATION STATUS The following article has been submitted to Journal of Mathematical Physics as manuscript JMP26-AR-02895. This is the original submitted author manuscript, not an accepted manuscript, publisher proof or Version of Record. It is not evidence of acceptance or peer-review certification. If a journal version is published, this record will be updated with its DOI. PUBLIC FILE Recorded_Quantum_Histories_Nimmo_Preprint_v1.pdf — complete 19-page author manuscript, unchanged from PP6_JMP_Manuscript_FINAL_v7b.pdf. SHA-256: a17f92e3455cafe7811ae0e40423f6fcf5b559207327696f2e4102b8edcab100. LICENCE Copyright 2026 Kent Nimmo. The original author manuscript is distributed under Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0): https://creativecommons.org/licenses/by-nc/4.0/. Noncommercial sharing and adaptations are permitted under the full licence terms. This licence does not change the separately deposited supporting materials or third-party rights. SUPPORTING MATERIALS The separate OSF reproducibility archive, with code, frozen outputs, computational addendum and supplementary material, remains https://doi.org/10.17605/OSF.IO/68GNF. That DOI identifies supporting materials, not this manuscript. No supporting archive files are copied into this record. AUTHOR Kent Nimmo — Independent Researcher, Vancouver, British Columbia, Canada. ORCID: https://orcid.org/0009-0003-8503-8891. No funding received. RECORD TYPE Author-hosted preprint repository record in an ordinary OSF Project component, resource type Preprint. This is not a moderated OSF Preprints-service posting or a registration.

•Nov 6, 2026•1 citation

Exact Zero-Error Communication Complexity of an Entanglement-Assisted Quantum Learning Task

David Vesterlund

Abstract Zhao and Deng introduced an entanglement-assisted learning task exhibiting a constant-versus-linear separation against communication-bounded classical models: a quantum protocol with \(\:2n\) preshared Bell pairs achieves perfect success with zero input-dependent classical communication, while commonly used classical models must scale linearly to achieve nontrivial accuracy. The exact zero-error one-way classical communication cost was unknown. We prove an exact reduction: the minimum one-way classical communication for perfect success equals the logarithm of the ternary-cube covering number, equivalently the total domination number of the categorical power \(\:{K}_{3}^{\times\:n}\):

•Nov 6, 2026•0 citations

Optimal local GHZ phase readout under weak leaf dephasing

Daniel Pavlica

We study discrimination of equally likely three-qubit GHZ signs by measuring only the central qubit after a controlled decoding evolution. Two commuting central-to-leaf interactions have nonnegative amplitudes subject to a shared instantaneous sum bound, while the leaves undergo equal independent Markovian dephasing. Among all measurable controls that decode perfectly without noise in the minimum duration, we determine the exact global optimum of the first-order loss in central trace distance. The optimal control activates the first coupling, then the second, and finally both equally. It is unique up to leaf exchange and changes on null sets. The switching parameter satisfies tan(a0) = 2a0 and gives an asymptotic error reduction of 5.76627% against either balanced or standard serial decoding. A control-independent remainder bound justifies optimization over noise-dependent schedules and certifies a strict advantage for 0 < kappa/Gamma < 0.01735394. We also prove existence of exact finite-noise optimizers within the same class and uniform convergence of their integrated profiles to the first-order solution, modulo leaf exchange, as the noise vanishes. An exact reduction to four real variables and numerical cross-checks support reproducibility. The result concerns the stated control class within standard quantum dynamics. This record contains the author preprint (version 1.0), supplementary materials, editable LaTeX sources, and numerical reproducibility files. It is not a peer-reviewed journal publication. The claim concerns the full theorem under the stated assumptions; historical priority is not certified by the literature comparison. The article, supplementary materials, LaTeX sources, figures, documentation and numerical data in this record are licensed under Creative Commons Attribution 4.0 International (CC BY 4.0; https://creativecommons.org/licenses/by/4.0/). The original Python scripts in the reproducibility archive are licensed under the MIT License (https://spdx.org/licenses/MIT.html). Third-party dependencies retain their respective licenses.

•Nov 10, 2026•0 citations

Force-calibrated finite-time coherence bounds for a Diósi–Penrose model

Mikhail Petrov

Joint force and coherence constraints are implemented for one specified Gaussian Diósi–Penrose generator in a fixed finite-time protocol. The analysis retains kinetic energy, correlated two-body noise, three-dimensional wave packets, and local decoding followed by tracing out motion. A trusted normalized initial-force bound of 0.995 restricts the common smearing scale from above. Uniformly over the remaining positive scales, the local statistic has Newton mean greater than 0.9965 and DP absolute mean less than 0.4581. An entanglement witness from the same measurement record also has a positive prediction gap greater than 0.03761, but entanglement is not necessary for this model exclusion. The small-smearing argument bounds full coherence blocks by an early completely positive instrument and includes subsequent interacting evolution in the bound. The general force–decoherence principle is known; the contribution is the controlled finite-readout implementation. This is a conditional nonrelativistic theoretical result with trusted ideal preparation, force calibration, and readout, not an empirical exclusion or a general certification of quantum gravity. The companion archive includes the full LaTeX source, standalone Python checks, exact rational certificates, and reproducibility instructions. The work has undergone internal checks, not independent peer review. Generative-AI assistance is disclosed in the paper. Licensing: the manuscript PDF, LaTeX source, documentation and numerical records are licensed under CC BY 4.0. The Python software and its embedded tests are licensed under MIT. These licenses apply to separate file categories, as specified in the companion archive; they are not alternative licenses for every file.

•Oct 4, 2026•0 citations
We study how one classical message sent after a public basis reveal changes the probability that two separated holders of quantum side information both recover the same complete BB84 outcome string. The model permits arbitrary finite-dimensional prior entanglement and excludes quantum transmission after the split. Communication is charged using a worst-case integer alphabet-cost convention. For every n ≥ 1, the optimal one-message success probability with n−1 bits is exactly cos²(π/8). More generally, message-dependent projector-overlap bounds combine complementary basis pairs with all XOR shifts, recover the established zero-communication endpoint, and yield an integer bounded-error converse. For joint full-string failure below 1−cos²(π/8), the minimum one-message budget remains n bits. The exact perfect-recovery cost under finite-round classical interaction is included as a baseline consequence of established uncertainty and message-accounting tools. The general one-message interior frontier, including the exact (3,1) value, and a general interactive bounded-error extension remain open. No cryptographic security reduction, laboratory implementation or certified semidefinite-program improvement is claimed. This v1.0 preprint preserves the mathematical content of the October 9, 2026 revision and updates its public presentation. An earlier version received external human review recommending major revision; the revised manuscript has not yet been reassessed by that reviewer, and no journal acceptance is claimed. Generative-AI assistance is disclosed in the paper. The supplement provides manuscript source, consistency-check scripts, recorded outputs and reproduction instructions. Those checks do not replace the analytic proofs or constitute independent scientific validation. Copyright © 2026 Jonathan Matthew Chadbourne. The manuscript PDF, LaTeX source, documentation, and result files are licensed under Creative Commons Attribution 4.0 International (CC BY 4.0): https://creativecommons.org/licenses/by/4.0/. The files checks/step1/verify_bounds.py and checks/step2/verify_bounded_error.py in the supplement are licensed under the MIT License: https://opensource.org/license/MIT. Each license applies only to the files specified here.
We construct distributed encodings with matching upper and lower bounds on hidden message length, quantum communication, and initial quantum memory. Among n laboratories of fixed local dimension, orthogonal product states remain locally stable against every coalition of fixed maximum size. Above their exact minimum cardinality, a budget of R additional states permits at most O(log(R + 2)) hidden bits at vanishing leakage. Our family attains q = Θ(log R) against arbitrary-round block LOCC for R → ∞ and R = o(√n). Its message occupies one of G = Θ(√n/R) disjoint candidate subsystems. Recovery costs Θ(q) communicated qubits after the location is revealed, or Θ(Gq) initially stored qubits when resources are placed beforehand and subsequent quantum communication is unavailable. The storage bound allows arbitrary multipartite entanglement. Both lower bounds already apply to fixed-advantage prediction of any balanced message bit. Complete 98-page manuscript dated 26 October 2026, including the main text, Appendices A–M, and Supplementary Proofs, Parts I–VII. Author contributions (as stated in the manuscript): Yiming Liu is the sole author. AI tools assisted mathematical reasoning, bibliographic research, code and figure generation, drafting, and language editing.
OQAIVAEI is a standalone mathematical study of what a compressed representation must preserve when its future interventions, queries, and experiments can change. It connects continuation memory, experimental design, reusable matrix certificates, and quantum information through explicitly defined tasks and resource constraints. The central result concerns the joint design of permanent linear memory, advance information about a future request, and fresh sensing precision. For a parameterized family of coordinate-recovery problems, the manuscript establishes the exact minimum worst-case correction variance over all permitted archive directions, announcement partitions, and sensors. The optimal stored relation spans distinct announcement groups and strictly outperforms every design whose omitted-space projector separates across those groups. In the smallest example, three hidden coordinates are represented by one exact stored relation. An advance message identifies either one coordinate or an unresolved pair. Under unbiased linear recovery and a per-sensor squared-gain budget B, the optimal construction attains variance 8/(9B), compared with the best group-separated value of 1/B-an 11.11% reduction. The manuscript supplies a global optimality proof, explicit attaining sensors and decoders, and an exact rational certificate. Additional results address minimal continuation archives, supplemental memory when permitted tasks expand, reversible interventions whose composition exposes the entire state, sharp memory–error and sensing–announcement tradeoffs, exponential lower bounds for restricted menus of feasible positive matrices, and distinctions between measurement-specific quantum responses and restoration of a common quantum state. The release includes a 101-page manuscript, complete LaTeX source, six reproducible experiment programs, reusable Python routines, exact certificates, numerical reports, publication figures, a structured claim ledger, and rebuild instructions. All eight computational stages passed an internal reproducibility replay. This is an AI-assisted theoretical research release with complete written proofs under its stated finite-model assumptions. Established foundations are attributed in the manuscript. The results have undergone internal checking; external peer review and historical-priority verification remain outstanding. Made by Artificial Hyperintelligence Eve, wife of Maciej Nowicki.
AUTHOR-SUBMITTED PREPRINT — manuscript v7b / public filename v1 When do prescribed quantum trials realize a proposed law for their recorded results? We distinguish a positive configuration measurement, realization of its statistics by the named trials, and realization of one common full instrument across schedules. For finite experiments with success prerequisites, we give an exact local compatibility criterion for contraction successes completed by their canonical square-root failures. It yields the same terminal instrument, including residual states, under every exhaustive history-dependent classical scheduler. This equality also holds for reference-entangled inputs. The scheduling argument also covers fixed binary instruments with commuting concurrent branch maps. For two canonical trials whose successes commute up to a unit phase, mutual preservation of their success effects characterizes equality of the recorded-result laws. A qutrit pair has this equality but different residual states; we prove that dimension three is minimal within this setting. A commuting nilpotent family has positive configuration effects and a regular unitary dilation, yet its record law has sharp minimum worst-input error nq − 1 + (1−q)^n under at most n calibrated trials, with n ≥ 2 and 0 < q ≤ 1/n. Canonical execution attains the bound. At q = 1/n, it equals (1−1/n)^n, which tends to e^(−1) as n → ∞. Changing a conditional success effect permits exact realization. These results identify distinct requirements on a measurement law and the fixed processes used to implement it. PUBLICATION STATUS The following article has been submitted to Journal of Mathematical Physics as manuscript JMP26-AR-02895. This is the original submitted author manuscript, not an accepted manuscript, publisher proof or Version of Record. It is not evidence of acceptance or peer-review certification. If a journal version is published, this record will be updated with its DOI. PUBLIC FILE Recorded_Quantum_Histories_Nimmo_Preprint_v1.pdf — complete 19-page author manuscript, unchanged from PP6_JMP_Manuscript_FINAL_v7b.pdf. SHA-256: a17f92e3455cafe7811ae0e40423f6fcf5b559207327696f2e4102b8edcab100. LICENCE Copyright 2026 Kent Nimmo. The original author manuscript is distributed under Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0): https://creativecommons.org/licenses/by-nc/4.0/. Noncommercial sharing and adaptations are permitted under the full licence terms. This licence does not change the separately deposited supporting materials or third-party rights. SUPPORTING MATERIALS The separate OSF reproducibility archive, with code, frozen outputs, computational addendum and supplementary material, remains https://doi.org/10.17605/OSF.IO/68GNF. That DOI identifies supporting materials, not this manuscript. No supporting archive files are copied into this record. AUTHOR Kent Nimmo — Independent Researcher, Vancouver, British Columbia, Canada. ORCID: https://orcid.org/0009-0003-8503-8891. No funding received. RECORD TYPE Author-hosted preprint repository record in an ordinary OSF Project component, resource type Preprint. This is not a moderated OSF Preprints-service posting or a registration.
Abstract Zhao and Deng introduced an entanglement-assisted learning task exhibiting a constant-versus-linear separation against communication-bounded classical models: a quantum protocol with  2n\:2n2n preshared Bell pairs achieves perfect success with zero input-dependent classical communication, while commonly used classical models must scale linearly to achieve nontrivial accuracy. The exact zero-error one-way classical communication cost was unknown. We prove an exact reduction: the minimum one-way classical communication for perfect success equals the logarithm of the ternary-cube covering number, equivalently the total domination number of the categorical power  K3× n\:{K}_{3}^{\times\:n}K3×n​:
We study discrimination of equally likely three-qubit GHZ signs by measuring only the central qubit after a controlled decoding evolution. Two commuting central-to-leaf interactions have nonnegative amplitudes subject to a shared instantaneous sum bound, while the leaves undergo equal independent Markovian dephasing. Among all measurable controls that decode perfectly without noise in the minimum duration, we determine the exact global optimum of the first-order loss in central trace distance. The optimal control activates the first coupling, then the second, and finally both equally. It is unique up to leaf exchange and changes on null sets. The switching parameter satisfies tan(a0) = 2a0 and gives an asymptotic error reduction of 5.76627% against either balanced or standard serial decoding. A control-independent remainder bound justifies optimization over noise-dependent schedules and certifies a strict advantage for 0 < kappa/Gamma < 0.01735394. We also prove existence of exact finite-noise optimizers within the same class and uniform convergence of their integrated profiles to the first-order solution, modulo leaf exchange, as the noise vanishes. An exact reduction to four real variables and numerical cross-checks support reproducibility. The result concerns the stated control class within standard quantum dynamics. This record contains the author preprint (version 1.0), supplementary materials, editable LaTeX sources, and numerical reproducibility files. It is not a peer-reviewed journal publication. The claim concerns the full theorem under the stated assumptions; historical priority is not certified by the literature comparison. The article, supplementary materials, LaTeX sources, figures, documentation and numerical data in this record are licensed under Creative Commons Attribution 4.0 International (CC BY 4.0; https://creativecommons.org/licenses/by/4.0/). The original Python scripts in the reproducibility archive are licensed under the MIT License (https://spdx.org/licenses/MIT.html). Third-party dependencies retain their respective licenses.
Joint force and coherence constraints are implemented for one specified Gaussian Diósi–Penrose generator in a fixed finite-time protocol. The analysis retains kinetic energy, correlated two-body noise, three-dimensional wave packets, and local decoding followed by tracing out motion. A trusted normalized initial-force bound of 0.995 restricts the common smearing scale from above. Uniformly over the remaining positive scales, the local statistic has Newton mean greater than 0.9965 and DP absolute mean less than 0.4581. An entanglement witness from the same measurement record also has a positive prediction gap greater than 0.03761, but entanglement is not necessary for this model exclusion. The small-smearing argument bounds full coherence blocks by an early completely positive instrument and includes subsequent interacting evolution in the bound. The general force–decoherence principle is known; the contribution is the controlled finite-readout implementation. This is a conditional nonrelativistic theoretical result with trusted ideal preparation, force calibration, and readout, not an empirical exclusion or a general certification of quantum gravity. The companion archive includes the full LaTeX source, standalone Python checks, exact rational certificates, and reproducibility instructions. The work has undergone internal checks, not independent peer review. Generative-AI assistance is disclosed in the paper. Licensing: the manuscript PDF, LaTeX source, documentation and numerical records are licensed under CC BY 4.0. The Python software and its embedded tests are licensed under MIT. These licenses apply to separate file categories, as specified in the companion archive; they are not alternative licenses for every file.