Toward tensor renormalization group study of lattice QCD

Atis Yosprakob

REVIEW 3 major objections 4 minor 28 references

Toward tensor renormalization group study of lattice QCD

T0 review · 3 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read This paper argues that two tensor-network constructions—multi-layer fermion placement and the armillary sphere formulation—remove the practical obstacles that block tensor renormalization group (TRG) studies of lattice QCD.

desk verdict A transparent proceedings summary of the author's own TRG program; the 3D SU(2) deconfinement benchmark is encouraging, but the path to QCD remains conditional on controlling the character expansion at N_tau>=2. read the letter →

arxiv 2501.14293 v1 pith:YKLZVHUJ submitted 2025-01-24 hep-lat

Atis Yosprakob This is my paper · ORCID

classification hep-lat MSC 81T2581T8065Z05 PACS 11.15.Ha12.38.Gc

keywords tensorrenormalizationgrouplatticeQCDnon-Abeliangaugetheoryarmillaryspherecharacterexpansionmulti-flavorfermionssignproblemdeconfinementtemperature

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper makes the case that tensor renormalization group (TRG) can become a viable route to lattice QCD by solving two practical obstacles: the explosion of tensor size when multiple fermion flavors are present, and the non-local entanglement structures that plague non-Abelian gauge theories. The multi-layer construction places each flavor in its own layer connected by a shared gauge link, shrinking the initial tensor by orders of magnitude. The armillary sphere formulation rewrites the gauge link integral via Clebsch-Gordan coefficients so that the matrix indices carrying non-local entanglement can be traced out analytically, leaving only representation indices. As evidence, the author shows that the Polyakov loop susceptibility of three-dimensional SU(2) pure gauge theory yields a deconfinement temperature consistent with Monte Carlo using only a handful of character-expansion terms and bond dimension 96. If the approach generalizes as claimed, TRG could handle full QCD at finite density without a sign problem.

What carries the argument

The armillary sphere formulation is the central mechanism: after character-expanding the plaquette action and the Polyakov loop, each link integral over the gauge group is evaluated analytically using the grand orthogonality relation and Clebsch-Gordan decomposition, yielding vertex tensors $V$ and link-conditional tensors $K$. The matrix indices (the $i,j$ indices of group elements) form closed loops around each lattice site, like the rings of an armillary sphere, and can be contracted exactly, leaving a network whose legs are only representation indices $r$. This eliminates the non-local entanglement structure that otherwise causes degeneracy in the singular value spectrum. The multi-layer construction, in parallel, splits an $N_f$-flavor action into separate layers, each with its own copy of the gauge link, connected by Kronecker deltas; this keeps the Grassmann tensor local with respect to flavor, reducing the initial tensor size from exponential to linear in $N_f$.

What would settle it

Repeat the SU(2) and SU(3) finite-temperature calculations with additional character-expansion terms ($r_{\mathrm{plaq}} = 4, 5, \dots$) and at $N_\tau = 2$ with larger $\beta$; if the Polyakov susceptibility peak shifts away from the Monte Carlo value or fails to converge as the truncation is relaxed, the central assumption of character-expansion accuracy would be refuted.

Watch

Extended reading notes

Core claim

The central claim is that the non-Abelian obstruction to tensor renormalization group—a severe degeneracy in the singular value spectrum caused by internal symmetry—can be removed by rewriting the link tensor using the Clebsch-Gordan decomposition. In the armillary sphere formulation, each link integral decomposes into vertex tensors $V$ built from Clebsch-Gordan coefficients and a link-conditional tensor $K$ enforcing representation consistency; the matrix indices assemble into a closed structure around each site whose contraction can be performed analytically. What remains is a tensor network in representation indices only, with no non-local entanglement tail. Supporting this, the author's numerical results for three-dimensional SU(2) and SU(3) pure gauge theories show a deconfinement transition, with SU(2) matching the Monte Carlo value at bond dimension 96 using character-expansion terms $r_{\mathrm{plaq}}$ up to 3 and $r_L$ up to 2. For multi-flavor systems, the multi-layer construction couples flavor layers only through a shared gauge field, so the number of tensor components grows linearly rather than exponentially in the number of flavors.

Load-bearing premise

The truncated character expansion of the gauge action remains accurate at the $\beta$ values used in the finite-temperature studies; the paper itself notes it becomes less accurate at larger $\beta$, which is why the numerical results are limited to one temporal slice.

Editorial extensions

If this is right

If the armillary sphere formulation is correct, tensor renormalization group can reach non-Abelian gauge theories at small bond dimension, since the hard part of the entanglement is removed analytically.
The formulation applies to any pure gauge action that is a class function, so improved actions and theta terms can be treated by expanding each term separately in characters.
The multi-layer construction reduces the cost of flavor to roughly linear, making multi-flavor lattice QCD tensor networks feasible in principle.
Combining the multi-layer construction with the armillary sphere formulation would produce a tensor network for unquenched QCD, provided the Kronecker-delta couplings between layers can be represented efficiently.
Finite-density quantities like the Silver Blaze phenomenon, which are difficult for Monte Carlo due to the sign problem, can be computed directly with TRG, as demonstrated for $ℤ_2$ gauge theory.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A natural extension not pursued in the paper is to test the armillary sphere formulation at $N_\tau \geq 2$, where larger $\beta$ is required; the author's own footnote warns the character expansion becomes less accurate at larger $\beta$, so this is the key regime in which the method must prove itself.
Because the armillary sphere's vertex and link-conditional tensors depend only on the gauge group, the same group-theoretic data for SU(3) could be reused across many lattice actions, potentially enabling a library of precomputed tensors for future QCD calculations.
The analytic contraction of matrix indices might also open the door to TRG studies of real-time evolution or nonzero chemical potential in non-Abelian theories, regimes where Monte Carlo suffers from sign or phase problems.
The paper does not discuss volume independence, but its character-expansion route in higher dimensions could be combined with existing arguments to test volume reduction for SU(N) theories beyond two dimensions.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

This is a conference proceedings that honestly re-presents the author's prior work. The two techniques—multi-layer construction for flavors and armillary sphere for non-Abelian gauge fields—are taken from earlier papers, explicitly cited. The paper's value is not a new derivation but a compact synthesis plus one new numerical benchmark: the 3D SU(2) deconfinement transition from the armillary-sphere TRG, which matches Monte Carlo with a bond dimension of 96. That is a meaningful data point for the TRG community, and the paper deserves credit for being transparent about its limits.

The writing is clear and the logic is straightforward. The stress-test concern is valid and is actually acknowledged in footnote 1: for N_tau >= 2, larger beta is needed, but the character expansion becomes less accurate at large beta. So all finite-temperature results sit at N_tau = 1. That is not a fatal flaw, but it is load-bearing for the "toward QCD" framing. The 3D pure-gauge tests also leave the full fermion + non-Abelian + 4D case entirely prospective. The benchmark has no error bars and no code or data release; the numbers are plausible but cannot be independently checked from the paper alone. The heavy self-citation is not a problem here because this is a proceedings of one's own line of work, and the external Monte Carlo check in ref. [28] provides some independent grounding.

The character-expansion truncation is the softest spot. Figure 6 shows the susceptibility shifting as rplaq goes from {1} to {1,2} to {1,2,3}; the result stabilizes for the truncated set, but there is no extrapolation in rplaq or any estimate of the omitted-representation error. For a physical continuum result, that error must be controlled. The author knows this; it is in the footnote. Still, the armillary sphere construction itself removes a known technical obstacle (non-local entanglement / singular-value degeneracy), and that is a real conceptual step.

Who should read this: people working on tensor renormalization group for gauge theories, especially those trying to push TRG toward non-Abelian and finite-density systems. It is worth a serious referee for a proceedings, mostly to check that the benchmark claims are honestly reported and that the outlook section is not overreaching. The paper is not a full research contribution, so it would not be right to hold it to that standard. As a proceedings, it is solid work and should be accepted.

Recommendation: engage with it, but treat the N_tau = 1 limitation as the stated boundary of the current method.

Referee Report

3 major / 4 minor

Summary. This proceedings contribution summarizes two techniques developed by the author for applying tensor renormalization group methods to lattice QCD: the multi-layer construction for multi-flavor gauge theories (Sec. 3) and the armillary sphere formulation for non-Abelian gauge theories (Sec. 4). The paper presents two sets of numerical results: a finite-density study of 2D Z2 gauge theory with N_f = 1, 2, 4 Wilson fermions that reproduces Silver Blaze behavior, and a finite-temperature study of 3D SU(2) and SU(3) pure gauge theories with a Polyakov loop source, where the SU(2) deconfinement temperature is reported to be consistent with Monte Carlo data. The broader claim is that the armillary sphere formulation removes the non-local entanglement structures that cause severe singular-value degeneracy in non-Abelian tensor networks, and that the two techniques can eventually be combined for lattice QCD.

Significance. The armillary sphere construction is a notable structural advance: by analytically contracting the matrix indices of the link integrals via Clebsch-Gordan coefficients, it converts non-local index loops into local representation indices. If the numerical benchmarks are reliable, this would allow TRG treatment of non-Abelian gauge theories with modest bond dimensions, and the SU(2) deconfinement-temperature match at D_cut = 96 is a genuinely encouraging falsifiable check. The Silver Blaze benchmark in the 2D Z2 model demonstrates the sign-problem-free capability and provides a quantitative comparison against an analytic benchmark. The paper is appropriately cautious in some places (especially footnote 1), but the overall 'toward QCD' claim rests on evidence from 2D Abelian and 3D pure-gauge examples; the combination of non-Abelian gauge symmetry, dynamical fermions, and 4D spacetime remains unvalidated.

major comments (3)

[Sec. 4, footnote 1; Eqs. (13)-(14)] The finite-temperature benchmarks are all performed at N_tau = 1 because, as stated in footnote 1, N_tau >= 2 requires larger beta values at which the truncated character expansion becomes less accurate. Since the character expansion truncation (rplaq up to 3, and r_L up to 2 for SU(2)) is central to the armillary sphere construction, this limitation directly affects the claim that the method can be extended to lattice QCD, where N_tau >= 2 and larger beta are unavoidable. The paper should either quantify the truncation error at the beta values used (for example, by reporting the magnitudes of f_r for the omitted representations), present a test at N_tau = 2 in a simpler non-Abelian theory, or explicitly revise the claim to state that the current formulation is not expected to reach the continuum regime.
[Sec. 4, Fig. 6] The key numerical evidence is the match of the SU(2) deconfinement temperature to the Monte Carlo result [28], but the figure has no error bars and no convergence study in D_cut (only D_cut = 96 is reported). The rplaq truncation is varied as {1}, {1,2}, {1,2,3}, but the r_L truncation is fixed and no quantitative measure of the spread among these curves is given. In addition, the SU(3) result is not benchmarked against an independent Monte Carlo result, so the claim that the deconfinement transition is correctly captured is not supported to the same standard for SU(3).
[Sec. 5, combining the techniques] The paper asserts that with fermions 'the armillary tensor can still be constructed' and that the multi-layer construction can be incorporated 'as long as an efficient representation of the Kronecker delta is provided,' but no numerical demonstration of the combined formulation is given. Since the title promises a step toward lattice QCD, which requires non-Abelian gauge symmetry, dynamical fermions, and four spacetime dimensions, the gap between the separate demonstrations and the combined claim should be explicitly acknowledged as prospective. A proof-of-principle in a partially simplified setting, such as 2+1D SU(2) with one Wilson flavor or a 4D Abelian model with the armillary sphere, would substantially strengthen the central claim.

minor comments (4)

[Sec. 4, Fig. 6 right panel] The legend 'rplaq, r_L ∈ {1,3,3̄}' is ambiguous, and the caption states that a dashed line indicates the transition temperature from Monte Carlo without specifying whether this applies to SU(3) or only to SU(2); please clarify both the legend and the caption.
[Sec. 2, Eq. (15)] The notation '{r1,...,s_D ∈ {trv,fund}}' is nonstandard; please state explicitly that r_mu and s_mu are independent representation labels for each direction mu and specify the ranges of the matrix indices i_mu, j_mu, k_mu, l_mu.
[Sec. 3] The multi-layer construction is demonstrated only for the 2D Z2 gauge theory (not a general Z_K theory) with N_f = 1, 2, 4; the text should state this explicitly so that the numerical benchmark is not overgeneralized.
[Sec. 4, Eq. (23)] The symbol Λ_2 is used in the Polyakov loop source term without a definition; please define it as the spatial lattice (the xy-plane) before first use.

Circularity Check

0 steps flagged · score 2.0 of 10

No load-bearing circularity: the central benchmark is an independent Monte Carlo comparison, and the self-cited armillary-sphere construction is supported by in-paper numerical evidence.

full rationale

The paper is a proceedings summary built on the author's prior work (refs [17]-[19]), so self-citation is frequent, but the central claims do not reduce to those citations. The key numerical claim, the SU(2) deconfinement temperature in Section 4, is obtained by evaluating the partition function with the character-expanded action (Eqs. (13)-(14)) and comparing the resulting susceptibility peak with an independent Monte Carlo result [28]. The coefficients f_r(beta/N) are defined by the exact Haar integral in Eq. (14), not fitted to the benchmark; the representation truncation r_plaq in {1,2,3} is an approximation whose error is assessed by comparing different truncation orders in Fig. 6, not by construction. The armillary-sphere decomposition (Eq. (22)) is cited from [18,19], but the paper also presents direct evidence -- the singular value spectra in Fig. 5 from HOSVD of the initial tensor -- that the expected degeneracy is absent, so the claim is not carried by the citation alone. No uniqueness theorem is imported from the authors' prior work to forbid alternatives. The acknowledged scope limitation in footnote 1, that for N_tau >= 2 'the character expansion becomes less accurate at large beta', is an honest statement about the truncation's validity, not a circular step; it limits the physical regime but does not make the derivation equivalent to its inputs. I therefore find no circular reduction; the mild self-citation density is contextual and not load-bearing.

Assumptions & free parameters 7 free parameters · 5 assumptions · 0 invented entities

The constructive steps rely on standard group theory and Grassmann integration, plus algorithm-specific truncation. The key unproven-to-paper element is the convergence of the truncated character expansion at the beta values needed for QCD, which the author explicitly flags as a limitation.

free parameters (7)

HOTRG bond dimension (flavor direction) = 64 (Nf=1,2); 32 (Nf=4)
Truncation rank in higher-order TRG along the flavor direction; chosen by hand to balance cost and accuracy, not fitted to data.
TRG bond dimension (space-time) = 64
Levin-Nave TRG bond dimension for the 2D space-time plane; chosen by hand.
Levin-Nave bond dimension D_cut (3D) = 96
Bond dimension for the 3D pure gauge TRG; chosen by hand.
Character expansion truncation (plaquette) = rplaq in {1}, {1,2}, {1,2,3}
Number of representation terms retained in the character expansion of the plaquette action; convergence checked by adding terms, chosen by hand.
Character expansion truncation (Polyakov source) = rL in {1,2} for SU(2); rL in {1,3,3bar} for SU(3)
Representation terms for the Polyakov loop source; chosen by hand.
Polyakov source coupling kappa and finite-difference step Delta_kappa = kappa=0.01, Delta_kappa=0.01
Source strength and numerical derivative step for susceptibility; chosen by hand.
Number of temporal slices N_tau = 1
Set to 1 to access high temperature at small beta; a limitation since larger N_tau needs larger beta where character expansion is less accurate.

assumptions (5)

domain assumption The site-to-link fermion transformation and Grassmann integration of F_x are valid, following Ref [20].
The construction relies on the exactness of the transformation of site fermions to link fermions and the analytic evaluation of the Grassmann integral; not proven in this paper.
ad hoc to paper The character expansion of the gauge action, truncated to a finite set of representations, converges sufficiently for the beta values used.
The paper explicitly notes this becomes less accurate at large beta (footnote 1), limiting N_tau >= 2.
standard math The Clebsch-Gordan decomposition and orthogonality relation for the link integral in Eq. (22) are valid and the resulting V and K tensors can be contracted exactly.
Group theory identities; assumed from Ref [18].
domain assumption The tensor network truncations (HOSVD, HOTRG, TRG) at specified bond dimensions give converged results for the observables shown.
No systematic convergence study is reported; only variation of character expansion terms is shown.
domain assumption The multi-layer construction with auxiliary gauge copies and delta functions reproduces the original multi-flavor theory exactly.
Equation (20) defines the equivalence via delta(U^(alpha)-U); this is straightforward but not explicitly proven in this proceedings.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Toward tensor renormalization group study of lattice QCD." pith.science (2026). https://pith.science/paper/YKLZVHUJ

@misc{pith2026250114293,
  author       = {Pith},
  title        = {Pith review of: Toward tensor renormalization group study of lattice QCD},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YKLZVHUJ}},
  note         = {Machine review of arXiv:2501.14293}
}

read the original abstract

The tensor renormalization group is a promising complementary approach to traditional Monte Carlo methods for lattice systems, as it is inherently free from the sign problem. We discuss recent developments crucial for its application to lattice QCD: the multi-layer construction for multi-flavor gauge theory and the armillary sphere formulation for non-Abelian gauge theory. These techniques are important for reducing the size of the initial tensor and for eliminating non-local entanglement structures within the tensor network. We present selected numerical results and discuss potential generalizations to lattice QCD.

Figures

Figures reproduced from arXiv: 2501.14293 by the authors.

**Figure 1.** The transformation of site fermions 𝜓 into link fermions 𝜂. The gauge link variable 𝑈 is shown for reference. taken to be the imaginary time direction, and 𝛾𝜇 denotes the Gamma matrices. It is convenient to separate the fermion bilinears into on-site terms and hopping terms. 𝜓¯ (𝛼) 𝑥 𝐷/ (𝛼)𝜓 (𝛼) 𝑥 = 𝜓¯ (𝛼) 𝑥 𝑊 (𝛼) 𝑥 𝜓 (𝛼) 𝑥 + ∑︁ 𝜈 n 𝜓¯ (𝛼) 𝑥 𝐻 (𝛼) 𝑥,𝜈 𝜓 (𝛼) 𝑥+𝜈ˆ + 𝜓¯ (𝛼) 𝑥 𝐻 (𝛼) 𝑥,−𝜈𝜓 (𝛼) 𝑥−𝜈ˆ o ; (4) 𝑊 (𝛼) 𝑥 := 𝑚˜ … view at source ↗

**Figure 2.** The tensor connections in a unit cell on site 𝑥; a) with the direct group integral, b) with the character expansion, and c) the multi-layer construction of a). Single lines without arrows correspond to bosonic link 𝑈, single lines with arrows correspond to fermionic links 𝜂, and double lines correspond to the representation links (𝑟, 𝑖, 𝑗). The black nodes are delta functions ensuring that all incoming links take th… view at source ↗

**Figure 3.** The number density 𝜌 = 1 𝑉 𝑑 𝑑𝜇˜ log 𝑍 per flavor as a function of 𝜇˜, with 𝑁𝑓 = 1, 2, and 4 for (left) free electron gas and (right) Z2 gauge theory. 4. The armillary sphere formulation for non-Abelian gauge theories It was shown in Ref. [6] that pure Yang-Mills theory suffers from a severe degeneracy of singular values due to non-local entanglement structures in the tensor network. In two dimensions, using charact… view at source ↗

Figures from the paper (3 more)

**Figure 4.** Figure 4: a) The link tensor 𝐿𝑥,𝜇 expressed in terms of vertex tensors 𝑉 and a link-conditioned tensor 𝐾. b) A three-dimensional example of a tensor network for a pure gauge theory, represented in terms of 𝑉 and 𝐾. An armillary sphere is shown around the lattice site, with 𝑉 as …

**Figure 5.** Figure 5: The singular value spectra of SU(2) and SU(3) gauge theories obtained from the HOSVD of the initial tensor without truncation. No severe degeneracy of singular values associated with the internal symmetry is observed here. 𝑓𝑟 and the diagonal Kronecker deltas (black no…

**Figure 6.** Figure 6: Polyakov susceptibility as a function of temperature 𝑇/𝑔 2 = 𝛽/2𝑁𝑁𝜏 for SU(2) (left) and SU(3) (right) gauge theories. Different symbols represent the data with different numbers of terms in the character expansions: 𝑟plaq corresponds to the representations from the pl…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

28 extracted references · 8 canonical work pages

[28]

M.Teper, TheFinitetemperaturephasetransitionofSU(2)gaugefieldsin(2+1)-dimensions , Phys. Lett. B313 (1993) 417. 10

work page 1993
[1]

Levin and C.P

M. Levin and C.P. Nave,Tensor renormalization group approach to 2D classical lattice models,Phys. Rev. Lett.99 (2007) 120601 [cond-mat/0611687]

arXiv 2007
[2]

Kadoh, Y

D. Kadoh, Y. Kuramashi, Y. Nakamura, R. Sakai, S. Takeda and Y. Yoshimura,Investigation of complex𝜙4 theory at finite density in two dimensions using TRG,JHEP 02(2020) 161 [1912.13092]

arXiv 2020
[3]

Akiyama, D

S. Akiyama, D. Kadoh, Y. Kuramashi, T. Yamashita and Y. Yoshimura,Tensor renormalization group approach to four-dimensional complex𝜙4 theory at finite density, JHEP09 (2020) 177 [2005.04645]

arXiv 2020
[4]

Akiyama, Y

S. Akiyama, Y. Kuramashi and Y. Yoshimura,Phase transition of four-dimensional lattice𝜙4 theory with tensor renormalization group,Phys. Rev. D104 (2021) 034507 [2101.06953]

arXiv 2021
[5]

Kuramashi and Y

Y. Kuramashi and Y. Yoshimura,Tensor renormalization group study of two-dimensional U(1) lattice gauge theory with a𝜃 term, JHEP04(2020) 089 [1911.06480]

arXiv 2020
[6]

Fukuma, D

M. Fukuma, D. Kadoh and N. Matsumoto,Tensor network approach to two-dimensional Yang–Mills theories, PTEP2021(2021) 123B03 [2107.14149]

arXiv 2021
[7]

Hirasawa, A

M. Hirasawa, A. Matsumoto, J. Nishimura and A. Yosprakob,Tensor renormalization group and the volume independence in 2D U(N) and SU(N) gauge theories, JHEP12(2021) 011 [2110.05800]

arXiv 2021

Show all 28 references

[8]

Kuwahara and A

T. Kuwahara and A. Tsuchiya,Toward tensor renormalization group study of three-dimensional non-Abelian gauge theory,PTEP2022 (2022) 093B02 [2205.08883]

2022 arXiv
[9]

Gu,Efficient simulation of Grassmann tensor product states, Phys

Z.-C. Gu,Efficient simulation of Grassmann tensor product states, Phys. Rev. B88(2013) 115139 [1109.4470]

2013 arXiv
[10]

Akiyama, Y

S. Akiyama, Y. Meurice and R. Sakai,Tensor renormalization group for fermions,J. Phys. Condens. Matter36(2024) 343002 [2401.08542]

2024 arXiv
[11]

Shimizu and Y

Y. Shimizu and Y. Kuramashi,Grassmann tensor renormalization group approach to one-flavor lattice Schwinger model,Phys. Rev. D90(2014) 014508 [1403.0642]

2014 arXiv
[12]

Shimizu and Y

Y. Shimizu and Y. Kuramashi,Critical behavior of the lattice Schwinger model with a topological term at𝜃=𝜋 using the Grassmann tensor renormalization group, Phys. Rev. D 90 (2014) 074503 [1408.0897]

2014 arXiv
[13]

Shimizu and Y

Y. Shimizu and Y. Kuramashi,Berezinskii-Kosterlitz-Thouless transition in lattice Schwinger model with one flavor of Wilson fermion,Phys. Rev. D97 (2018) 034502 [1712.07808]

2018 arXiv
[14]

Z.Y. Xie, J. Chen, M.P. Qin, J.W. Zhu, L.P. Yang and T. Xiang,Coarse-graining renormalization by higher-order singular value decomposition, Phys. Rev. B86 (2012) 045139 [1201.1144]. 9 Toward tensor renormalization group study of lattice QCD Atis Yosprakob

2012 arXiv
[15]

Adachi, T

D. Adachi, T. Okubo and S. Todo,Anisotropic Tensor Renormalization Group,Phys. Rev. B 102 (2020) 054432 [1906.02007]

2020 arXiv
[16]

Kadoh and K

D. Kadoh and K. Nakayama,Renormalization group on a triad network, 1912.02414

1912 arXiv
[17]

Yosprakob, J

A. Yosprakob, J. Nishimura and K. Okunishi,A new technique to incorporate multiple fermion flavors in tensor renormalization group method for lattice gauge theories, JHEP11 (2023) 187 [2309.01422]

2023 arXiv
[18]

Yosprakob,Reduced Tensor Network Formulation for Non-Abelian Gauge Theories in Arbitrary Dimensions, PTEP2024(2024) 073B05 [2311.02541]

A. Yosprakob,Reduced Tensor Network Formulation for Non-Abelian Gauge Theories in Arbitrary Dimensions, PTEP2024(2024) 073B05 [2311.02541]

2024 arXiv
[19]

Yosprakob and K

A. Yosprakob and K. Okunishi,Tensor renormalization group study of the three-dimensional SU(2) and SU(3) gauge theories with the reduced tensor network formulation, 2406.16763

arXiv
[20]

Akiyama and D

S. Akiyama and D. Kadoh,More about the Grassmann tensor renormalization group,JHEP 10 (2021) 188 [2005.07570]

2021 arXiv
[21]

Y. Liu, Y. Meurice, M.P. Qin, J. Unmuth-Yockey, T. Xiang, Z.Y. Xie et al.,Exact Blocking Formulas for Spin and Gauge Models, Phys. Rev. D88 (2013) 056005 [1307.6543]

2013 arXiv
[22]

Asaduzzaman, S

M. Asaduzzaman, S. Catterall, Y. Meurice, R. Sakai and G.C. Toga,Tensor network representation of non-abelian gauge theory coupled to reduced staggered fermions,JHEP05 (2024) 195 [2312.16167]

2024 arXiv
[23]

K.H. Pai, S. Akiyama and S. Todo,Grassmann tensor renormalization group approach to (1+ 1)-dimensional two-color lattice QCD at finite density, 2410.09485

arXiv
[24]

Kanno, S

H. Kanno, S. Akiyama, K. Murakami and S. Takeda,Grassmann Tensor Renormalization Group for𝑁𝑓 = 2 massive Schwinger model with a𝜃 term, 2412.08959
[25]

S. Akiyama,Matrix product decomposition for two- and three-flavor Wilson fermions: BenchmarkresultsinthelatticeGross-Neveumodelatfinitedensity ,Phys.Rev.D 108(2023) 034514 [2304.01473]

2023 arXiv
[26]

De Lathauwer, B

L. De Lathauwer, B. De Moor and J. Vandewalle,A multilinear singular value decomposition,SIAM Journal on Matrix Analysis and Applications21(2000) 1253

2000
[27]

Yosprakob,GrassmannTN: a Python package for Grassmann tensor network computations, SciPost Phys

A. Yosprakob,GrassmannTN: a Python package for Grassmann tensor network computations, SciPost Phys. Codebases (2023) 20(2023) [2309.07557]

2023 arXiv

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.

[1] [28]

M.Teper, TheFinitetemperaturephasetransitionofSU(2)gaugefieldsin(2+1)-dimensions , Phys. Lett. B313 (1993) 417. 10

work page 1993

[2] [1]

Levin and C.P

M. Levin and C.P. Nave,Tensor renormalization group approach to 2D classical lattice models,Phys. Rev. Lett.99 (2007) 120601 [cond-mat/0611687]

arXiv 2007

[3] [2]

Kadoh, Y

D. Kadoh, Y. Kuramashi, Y. Nakamura, R. Sakai, S. Takeda and Y. Yoshimura,Investigation of complex𝜙4 theory at finite density in two dimensions using TRG,JHEP 02(2020) 161 [1912.13092]

arXiv 2020

[4] [3]

Akiyama, D

S. Akiyama, D. Kadoh, Y. Kuramashi, T. Yamashita and Y. Yoshimura,Tensor renormalization group approach to four-dimensional complex𝜙4 theory at finite density, JHEP09 (2020) 177 [2005.04645]

arXiv 2020

[5] [4]

Akiyama, Y

S. Akiyama, Y. Kuramashi and Y. Yoshimura,Phase transition of four-dimensional lattice𝜙4 theory with tensor renormalization group,Phys. Rev. D104 (2021) 034507 [2101.06953]

arXiv 2021

[6] [5]

Kuramashi and Y

Y. Kuramashi and Y. Yoshimura,Tensor renormalization group study of two-dimensional U(1) lattice gauge theory with a𝜃 term, JHEP04(2020) 089 [1911.06480]

arXiv 2020

[7] [6]

Fukuma, D

M. Fukuma, D. Kadoh and N. Matsumoto,Tensor network approach to two-dimensional Yang–Mills theories, PTEP2021(2021) 123B03 [2107.14149]

arXiv 2021

[8] [7]

Hirasawa, A

M. Hirasawa, A. Matsumoto, J. Nishimura and A. Yosprakob,Tensor renormalization group and the volume independence in 2D U(N) and SU(N) gauge theories, JHEP12(2021) 011 [2110.05800]

arXiv 2021