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

Geometric Bounds and Conditional Tensor Renormalization in Four Dimensions

Kristijan Kozic

Published Oct 20, 2026• 0 citations
Publisher

Abstract

Version 6.0 was published with the intention of making the results of my research, including its limitations, unsuccessful approaches, and open questions, publicly available to anyone who might find them useful for further scientific investigation. This work investigates mathematically controlled tensor-renormalization constructions on four-dimensional lattices with countably infinite-dimensional Hilbert-space tensor legs. The reorganized manuscript places its principal mathematical results first: an explicit four-dimensional alternating ABAB geometry with an all-scale distance-gain estimate; conditional contraction bounds for multisite interactions; anchored joint-source transport; and a weighted Hilbert–Schmidt-to-Walsh embedding. The accompanying appendices preserve the original v5.3.1 rearrangement construction, including its proposed superlinear estimate with exponent 32/31 and a technical prefactor correction. They also document lattice-field applications involving Wilson SU(2), scalar φ⁴ theory, restricted Regge geometry, numerical benchmarks, and reproducible verification procedures. The work distinguishes exact mathematical identities, conditional results, finite computational checks, and unresolved proof obligations. A general autonomous four-dimensional tensor RG and continuum field-theoretic constructions remain open. Version 6.0 — Editorial reorganization: The original mathematical constructions and results are retained and presented in a theorem-first structure. The main text emphasizes the mathematical contributions, while archival derivations, physical examples, and verification materials are organized into appendices. This revision introduces no additional mathematical claims beyond those established in the preceding version.

DOI
10.5281/zenodo.23507889

Related papers

Geometric Bounds and Conditional Tensor Renormalization in Four Dimensions

Kristijan Kozic

This work develops a rigorous, truncation-free rearrangement tensor renormalization construction in four dimensions near the high-temperature fixed-point tensor. The construction uses exact tagged cluster cancellation, dimension-independent Hilbert–Schmidt estimates, and a superlinear contraction bound with exponent 32/31. This updated version includes an application to four-dimensional Euclidean lattice scalar φ⁴ field theory. It establishes an exact infinite-dimensional tensor representation, Hilbert–Schmidt bounds for local field-insertion tensors, and a conditional RG identity for the integrated two-point susceptibility. An exponential correlation-decay estimate is also derived independently in the massive lattice regime. The manuscript is accompanied by supplementary mathematical derivations and a reproducibility package containing computational verification scripts and supporting audits. The result remains a local, radius-dependent high-temperature construction. It does not establish the continuum limit, Yang–Mills mass gap, or uniform RG control of local correlation functions across all scales.

•Sep 27, 2026•0 citations

Explicit Hilbert-Schmidt Control of the Four-Dimensional Wilson SU(2) Armillary Tensor

Kristijan Kozic

We study the exact reduced armillary tensor representation of four-dimensionalWilson SU(2) lattice gauge theory in the strong-coupling regime and derive anexplicit Hilbert-Schmidt majorant for the normalized site tensor. Two logicallyindependent sufficient-condition results are presented. First, a deliberatelyconservative quantitative estimate places the Wilson tensor inside thetruncation-free four-dimensional rearrangement-RG basin for beta < 10^-18700.Second, bypassing the rearrangement map entirely, an exact-rational directpolymer certificate gives the weighted Kotecky-Preiss condition forbeta <= 2.7 x 10^-6, leading to fixed-lattice exponential clustering and atransfer-matrix spectral-gap corollary under the standard cited structuralinputs. The two results are explicitly separated. No continuum beta -> infinity,a -> 0 result is claimed, and this work does not solve the Yang-Mills Millenniumproblem. Strong-coupling lattice mass gaps are not claimed as novel; themodel-specific contribution is the explicit four-dimensional Wilson SU(2)armillary-to-Hilbert-Schmidt control and its connection to the rigorousfour-dimensional truncation-free rearrangement basin, together with theseparate direct-polymer certificate.

•Sep 27, 2026•0 citations

Recent progress in the tensor renormalization group

Daisuke Kadoh

The tensor renormalization group is a very attractive approach for numerically calculating lattice field theory. It does not suffer from the sign problem, and it is a method that enables calculation with an extremely large volume lattices and direct calculation of free energy. I report on recent developments in this field and clarify the problems to be solved in the future.

International Symposium on Lattice Field Theory (LATTICE)•May 25, 2022•6 citations

Renormalization, Entanglement and Continuous Tensor Networks

Adrián Franco-Rubio

The study of the ground states of local Hamiltonians in quantum many-body and quantum field theoretic systems is a source of many research problems of great complexity. Almost in its entirety, this thesis deals with constructions related to the field of tensor networks, which are efficient parametrizations of mathematical objects from quantum many-body theory (such as quantum states and operators) that exploit our knowledge of their entanglement structures in order to obtain improvements in our computational abilities. In the past decade, the new research program of continuous tensor networks has arisen with the goal of reproducing, in the setting of quantum field theory, the success that tensor network techniques have enjoyed on the lattice. First, we present a number of results from research performed on non-interacting examples of the continuous Multiscale Entanglement Renormalization Ansatz, or cMERA, a continuous tensor network that provides a variational wavefunctional for a quantum field theoretic ground state. Initially, we study the correlations and entanglement structure of cMERA states, providing evidence that such states can be interpreted as UV regularized versions of the physical states they approximate. After that, we study what modifications of the formalism are necessary in order to consider systems with gauge symmetry, or in the presence of boundaries and defects. We obtain prescriptions for the treatment of these cases which, we expect, may also hold in the interacting setting. We then abandon briefly the area of continuous tensor networks to present a piece of research within a different though related research program: that of the study of manifestations of universal emergent behaviour in critical lattice systems. In particular we bring attention to the existence of an approximate representation of the Virasoro algebra supported on the eigenvectors of the reduced density matrices of critical lattice systems. We call this the entanglement algebra, and study its accuracy in the particular case of the Ising model. Lastly, we present a proposal for a continuous version of the Tensor Network Renormalization (TNR) algorithm, which we dub continuous TNR, or cTNR. cTNR operates on UV regularized classical partition functions or quantum Euclidean path integrals, and generates a real-space renormalization group flow via continuous coarse-graining and rescaling operations. We show that a UV regularized version of the free boson path integral can be made a fixed point of such a renormalization flow, in a way that allows for recovery of the conformal data associated to the free boson conformal field theory.

•Jul 7, 2020•0 citations

Formal Pressure Testing a Relation-First Ontology: Representability, Non-Canonicity, and Structural Limits - The Universal Recursive Dynamics Case

D. Berkla

This release presents an Open Specialist Assessment Version of a formal pressure test concerning selected commitments of Universal Recursive Dynamics (URD) and categorical, dependent, history-sensitive, and process-theoretic resources used to represent them without collapsing structural conditioning into temporal succession, constitution into restriction, or participant continuation into current-state identity. The investigation originated when the author recognized a possible structural correspondence between unresolved URD questions and split-idempotent/Karoubi resources encountered in Hefford and Wilson's work on hyper-decoherence from a causally indefinite process theory to quantum theory. The author is not a specialist in the mathematical fields employed and could not independently determine whether that apparent correspondence was mathematically substantive. Sustained human–AI inquiry was therefore used to translate the initial question into explicit definitions, candidate constructions, propositions, proofs, counterexamples, negative results, and unresolved formal obligations. The manuscript reports both positive and negative results. Split-idempotent/Karoubi machinery is retained as a resource for restriction and invariance but does not supply relation-priority or constitution; compatible-idempotent order does not by itself encode Paper-A structural conditioning. Dependent/displayed structures provide a bounded representation of specifiability dependence and display-relative polarity; an inductive-inductive witness supplies a limited joint-generation capability and a paper-local specification non-separability result. History-sensitive individuation, register, and continuation resist reduction to endpoint or current state, while the proposed Part III interface remains representational rather than derivational and retains explicit open obligations. The surviving formal material is released for independent specialist scrutiny. AI-assisted convergence is not treated as a substitute for disciplinary verification, and the author does not claim specialist mathematical competence sufficient to certify the formal results independently. In this assessment version, PROVED means that a result is claimed to follow from the stated assumptions by the proof or argument supplied, or is a directly used standard result; it does not imply prior independent specialist verification. Corrections, counterexamples, typing objections, provenance challenges, scope objections, and identification of prior mathematical formulations are specifically invited. A separate Technical Assessment Guide accompanies the manuscript so that readers may evaluate only the portions within their expertise.

•Aug 17, 2026•0 citations

Tensor renormalization group and the solution of classical lattice models

Michael Levin, Cody P. Nave

•Nov 27, 2006•0 citations
Version 6.0 was published with the intention of making the results of my research, including its limitations, unsuccessful approaches, and open questions, publicly available to anyone who might find them useful for further scientific investigation. This work investigates mathematically controlled tensor-renormalization constructions on four-dimensional lattices with countably infinite-dimensional Hilbert-space tensor legs. The reorganized manuscript places its principal mathematical results first: an explicit four-dimensional alternating ABAB geometry with an all-scale distance-gain estimate; conditional contraction bounds for multisite interactions; anchored joint-source transport; and a weighted Hilbert–Schmidt-to-Walsh embedding. The accompanying appendices preserve the original v5.3.1 rearrangement construction, including its proposed superlinear estimate with exponent 32/31 and a technical prefactor correction. They also document lattice-field applications involving Wilson SU(2), scalar φ⁴ theory, restricted Regge geometry, numerical benchmarks, and reproducible verification procedures. The work distinguishes exact mathematical identities, conditional results, finite computational checks, and unresolved proof obligations. A general autonomous four-dimensional tensor RG and continuum field-theoretic constructions remain open. Version 6.0 — Editorial reorganization: The original mathematical constructions and results are retained and presented in a theorem-first structure. The main text emphasizes the mathematical contributions, while archival derivations, physical examples, and verification materials are organized into appendices. This revision introduces no additional mathematical claims beyond those established in the preceding version.
This work develops a rigorous, truncation-free rearrangement tensor renormalization construction in four dimensions near the high-temperature fixed-point tensor. The construction uses exact tagged cluster cancellation, dimension-independent Hilbert–Schmidt estimates, and a superlinear contraction bound with exponent 32/31. This updated version includes an application to four-dimensional Euclidean lattice scalar φ⁴ field theory. It establishes an exact infinite-dimensional tensor representation, Hilbert–Schmidt bounds for local field-insertion tensors, and a conditional RG identity for the integrated two-point susceptibility. An exponential correlation-decay estimate is also derived independently in the massive lattice regime. The manuscript is accompanied by supplementary mathematical derivations and a reproducibility package containing computational verification scripts and supporting audits. The result remains a local, radius-dependent high-temperature construction. It does not establish the continuum limit, Yang–Mills mass gap, or uniform RG control of local correlation functions across all scales.
We study the exact reduced armillary tensor representation of four-dimensionalWilson SU(2) lattice gauge theory in the strong-coupling regime and derive anexplicit Hilbert-Schmidt majorant for the normalized site tensor. Two logicallyindependent sufficient-condition results are presented. First, a deliberatelyconservative quantitative estimate places the Wilson tensor inside thetruncation-free four-dimensional rearrangement-RG basin for beta < 10^-18700.Second, bypassing the rearrangement map entirely, an exact-rational directpolymer certificate gives the weighted Kotecky-Preiss condition forbeta <= 2.7 x 10^-6, leading to fixed-lattice exponential clustering and atransfer-matrix spectral-gap corollary under the standard cited structuralinputs. The two results are explicitly separated. No continuum beta -> infinity,a -> 0 result is claimed, and this work does not solve the Yang-Mills Millenniumproblem. Strong-coupling lattice mass gaps are not claimed as novel; themodel-specific contribution is the explicit four-dimensional Wilson SU(2)armillary-to-Hilbert-Schmidt control and its connection to the rigorousfour-dimensional truncation-free rearrangement basin, together with theseparate direct-polymer certificate.
The study of the ground states of local Hamiltonians in quantum many-body and quantum field theoretic systems is a source of many research problems of great complexity. Almost in its entirety, this thesis deals with constructions related to the field of tensor networks, which are efficient parametrizations of mathematical objects from quantum many-body theory (such as quantum states and operators) that exploit our knowledge of their entanglement structures in order to obtain improvements in our computational abilities. In the past decade, the new research program of continuous tensor networks has arisen with the goal of reproducing, in the setting of quantum field theory, the success that tensor network techniques have enjoyed on the lattice. First, we present a number of results from research performed on non-interacting examples of the continuous Multiscale Entanglement Renormalization Ansatz, or cMERA, a continuous tensor network that provides a variational wavefunctional for a quantum field theoretic ground state. Initially, we study the correlations and entanglement structure of cMERA states, providing evidence that such states can be interpreted as UV regularized versions of the physical states they approximate. After that, we study what modifications of the formalism are necessary in order to consider systems with gauge symmetry, or in the presence of boundaries and defects. We obtain prescriptions for the treatment of these cases which, we expect, may also hold in the interacting setting. We then abandon briefly the area of continuous tensor networks to present a piece of research within a different though related research program: that of the study of manifestations of universal emergent behaviour in critical lattice systems. In particular we bring attention to the existence of an approximate representation of the Virasoro algebra supported on the eigenvectors of the reduced density matrices of critical lattice systems. We call this the entanglement algebra, and study its accuracy in the particular case of the Ising model. Lastly, we present a proposal for a continuous version of the Tensor Network Renormalization (TNR) algorithm, which we dub continuous TNR, or cTNR. cTNR operates on UV regularized classical partition functions or quantum Euclidean path integrals, and generates a real-space renormalization group flow via continuous coarse-graining and rescaling operations. We show that a UV regularized version of the free boson path integral can be made a fixed point of such a renormalization flow, in a way that allows for recovery of the conformal data associated to the free boson conformal field theory.
This release presents an Open Specialist Assessment Version of a formal pressure test concerning selected commitments of Universal Recursive Dynamics (URD) and categorical, dependent, history-sensitive, and process-theoretic resources used to represent them without collapsing structural conditioning into temporal succession, constitution into restriction, or participant continuation into current-state identity. The investigation originated when the author recognized a possible structural correspondence between unresolved URD questions and split-idempotent/Karoubi resources encountered in Hefford and Wilson's work on hyper-decoherence from a causally indefinite process theory to quantum theory. The author is not a specialist in the mathematical fields employed and could not independently determine whether that apparent correspondence was mathematically substantive. Sustained human–AI inquiry was therefore used to translate the initial question into explicit definitions, candidate constructions, propositions, proofs, counterexamples, negative results, and unresolved formal obligations. The manuscript reports both positive and negative results. Split-idempotent/Karoubi machinery is retained as a resource for restriction and invariance but does not supply relation-priority or constitution; compatible-idempotent order does not by itself encode Paper-A structural conditioning. Dependent/displayed structures provide a bounded representation of specifiability dependence and display-relative polarity; an inductive-inductive witness supplies a limited joint-generation capability and a paper-local specification non-separability result. History-sensitive individuation, register, and continuation resist reduction to endpoint or current state, while the proposed Part III interface remains representational rather than derivational and retains explicit open obligations. The surviving formal material is released for independent specialist scrutiny. AI-assisted convergence is not treated as a substitute for disciplinary verification, and the author does not claim specialist mathematical competence sufficient to certify the formal results independently. In this assessment version, PROVED means that a result is claimed to follow from the stated assumptions by the proof or argument supplied, or is a directly used standard result; it does not imply prior independent specialist verification. Corrections, counterexamples, typing objections, provenance challenges, scope objections, and identification of prior mathematical formulations are specifically invited. A separate Technical Assessment Guide accompanies the manuscript so that readers may evaluate only the portions within their expertise.