betaFeedback

WorksPhysics and Astronomy

Evaluating generative samplers for lattice field theory by scale

Abstract

Generative samplers for lattice field theory are commonly evaluated by their effective sample size (ESS). We show that a small global ESS need not imply inaccurate long-distance observables: errors spread over many short-distance modes can dominate this diagnostic. We introduce a scale-resolved diagnostic based on the overlap between coarse-grained model and target distributions. At the lattice cutoff, this overlap determines the asymptotic ESS fraction; lower cutoffs isolate long-wavelength errors. In the exactly solvable free theory, two samplers with comparable ESS yield a 32%32\% bias and no bias in the susceptibility, respectively. In two-dimensional ϕ4ϕ^4 theory near criticality, trained diffusion samplers show the same separation of scales: their path-weight ESS fraction falls rapidly over the volumes studied, while low moments of the magnetization remain close to their target values. Varying the reference parameters substantially changes infrared observables, with little effect on the log-weight variance. Given the target magnetization distribution, a marginal Metropolis test maintains an acceptance rate of order one in our tests. Combined with exact local updates, it reduces the remaining mismatch, although residual bias remains at finite relaxation length. In the free field, uniformly bounded relative errors in the proposal mode variances give a volume-independent bound on the number of updates needed for a prescribed accuracy of the chain's stationary mode variances at fixed correlation length. The target marginal must be supplied or computed separately; estimating it from the samplers' own weights can inherit the poor global overlap.

Introduction

Generative models are increasingly used to sample field configurations in lattice field theory; for reviews of machine-learning methods in lattice QCD, see Refs. [1, 2, 3]. Examples include normalizing flows [4, 5, 6, 7, 8] and diffusion models, which are closely related to stochastic quantization [9, 10]. Diffusion models have been applied to scalar field theories and sigma models [11, 12, 13, 14] and to gauge theories [15, 16, 17, 18], also with dynamical fermions [19, 20]; they have been extended to theories with a sign problem [21] and trained without reference ensembles [13, 22]. Other samplers are based on stochastic quantization formulated as optimal control (OSQ) [23], on stochastic path sampling [24] or on non-equilibrium transport with a learned drift [25]. Because the model distribution qq differs from the target distribution p=e−S/Zp=e^{-S}/Z, asymptotically exact results require a correction: reweighting with the importance weights w=p/qw=p/q [26], an independence Metropolis-Hastings (MH) test [4], or Metropolis-adjusted Langevin steps [15, 16] and MH tests [22] during generation.

The central obstacle to this program is the scaling with the number of degrees of freedom. For flow-based samplers of ϕ4\phi^4 theory, the cost of training a model to a given sampling efficiency grows extremely rapidly as the continuum limit is approached [6], and flow-based trivializing maps with a fixed architecture reduce the autocorrelation times of hybrid Monte Carlo (HMC) without changing their scaling toward the continuum [27]. This poor scaling has been attributed to the multiscale structure of the problem near critical points, which networks of moderate size do not capture efficiently [28]; in diffusion models the same structure makes the score ill-conditioned, so that the number of time steps of the generating process can grow, a difficulty mitigated by factorizing the distribution across scales [29]. The dependence on the volume is even more severe: the quality of a fixed model built from local transformations degrades exponentially under volume transfer [7], and for local theories a single global independence proposal can become impractical at realistic box sizes [30]. The scalability of flow-based sampling must therefore be assessed experimentally [7]; see also Ref. [1]. These studies suggest that, for a fixed local mismatch and finite correlation length, the ESS fraction and the acceptance rate of a global independence test can decrease as e−cVe^{-cV}, with a model-dependent constant cc. The volume dependence of the ESS of the OSQ sampler reported in Ref. [23] is consistent with this behavior qualitatively. The usual solution has been to reduce cc through more expressive architectures and improved training.

The exponential volume dependence has a simple origin. If the model deviates slightly from the target in each correlation volume, log⁡w\log w is a sum of many weakly correlated contributions, and its variance grows approximately in proportion to VV. For independent subvolumes, the population ESS fraction factorizes [7]. The useful sample size, however, depends on the observable [7], and flow-based variance reduction for specific observables has been found to be approximately independent of the volume [31]; a flow that encodes the generating functional extends it to arbitrary correlation functions [32].

Other proposals modify the structure of the sampler. Stochastic normalizing flows interleave flow layers with out-of-equilibrium Monte Carlo updates and inherit the favorable scaling of non-equilibrium Monte Carlo with the number of degrees of freedom [33, 34, 35], also when the intermediate distributions are those of a trained energy-based diffusion model [36], closely related to sequential Monte Carlo along the generating process [37]. Continuous normalizing flows improve the sampling efficiency, in particular on larger lattices [38], and learned gradient flows reduce the number of trainable parameters by orders of magnitude [39]. Localized updates restrict the generative proposal to subdomains [30, 40], and renormalization-group-inspired constructions treat different length scales separately [41, 42, 43, 44, 45, 46], also in diffusion and flow-matching models, which generate the field scale by scale [14, 29] or along a renormalization-group flow [47, 48, 49]. Physics-informed kernels (PIKs) build the generative map layer by layer from renormalization-group flows, so that learning reduces to solving a set of independent linear differential equations for the kernels of the transformation [50]; the construction has been extended to complex distributions with a sign problem [51]. Inverse blocking avoids the global test altogether: a coarse ensemble matched by a renormalization-group blocking supplies the long-distance modes, a conditional flow reconstructs the short-distance degrees of freedom, and rethermalization with the fine action reduces the residual mismatch [52, 53]. Short-distance mismatches then relax rapidly, whereas a mismatch in the relevant thermal direction relaxes much more slowly [52], as expected for an error that moves the ensemble away from the critical surface. Similarly, the residual bias of a diffusion-model sampler for ϕ4\phi^4 theory has been found to be concentrated in the zero mode [12].

These observations motivate two quantitative questions:

  1. Which part of the mismatch determines the ESS, and which part affects physical observables?

  1. Given the target distribution of selected infrared variables, how can it be used to reduce errors in full-configuration proposals?

We develop a framework for evaluating sampler errors by scale, the main subject of this paper, and examine both questions in the free field. The asymptotic ESS fraction is e−D2(p∥q)e^{-D_2(p\Vert q)}, where D2D_2 is the Rényi divergence of order two. Removing short-distance modes gives a monotone overlap flow D2(Λ)D_2(\Lambda): its value at the lattice cutoff determines the global ESS, while smaller cutoffs isolate errors at longer wavelengths. We construct two samplers with similar global ESS: a mass error biases the susceptibility by 32%, whereas a short-distance lattice artifact leaves it unbiased. The distinction also explains why weighting only the modes on which an observable depends avoids noise from other modes, in the spirit of the local factorization of Ref. [54].

For the second question, we derive the mean acceptance rate of an independence test and show that a test on a few collective variables (low-dimensional functions of the configuration such as the magnetization), which we call the marginal test, accepts at least as often as the full test at their respective stationary distributions. Its acceptance stays bounded away from zero if the overlap of the chosen marginal distributions does so as the volume grows. We then use exact local updates to relax the errors in the remaining modes. In the free field, the resulting composite chain reaches a given accuracy of the mode variances with a number of updates independent of volume at fixed correlation length and uniformly bounded relative errors in the proposal mode variances. Thus, once the infrared input is supplied, poor global overlap need not imply exponentially costly relaxation to a prescribed accuracy.

We test this picture with trained OSQ samplers for two-dimensional ϕ4\phi^4 theory near criticality. With reference parameters fixed from the benchmark, they reproduce ⟨∣M∣⟩\langle|M| \rangle and ⟨M2⟩\langle M^2 \rangle to a few percent while their path-weight ESS fractions fall below 10−310^{-3} at L=64L=64; the test on MM still accepts about 0.8 of the proposals. Changing the reference parameters substantially changes the infrared observables while barely changing the log-weight variance. In the composite chain, about ten HMC trajectories remove most of the ultraviolet mismatch at every volume studied. The order of the two corrections matters when the proposal’s ultraviolet is far from the target: its relaxation shifts the tested variables, so the ultraviolet should be relaxed before the infrared is corrected, or with the tested variables held fixed. The remaining limitations concern the untested low modes, their coupling to the tested variables, and errors in the estimated marginal densities. We also estimate the target distribution of MM from the action by integrating the mean force at fixed MM at L=32L=32 and 64, whereas the samplers’ own weights give it reliably only at L=16L=16 (Appendix C).

The marginal test alone has the correct stationary distribution for the tested variables when their density ratio is known; at fixed values of these variables it retains the model’s conditional distribution. The composite chain is generally biased at finite relaxation length. We therefore distinguish the cost set by full microscopic overlap from that set by a few infrared variables and the local relaxation time. The practical question is whether the relevant infrared errors can be identified and corrected using a number of variables that does not grow with volume, and whether their target distribution can be obtained at controlled cost as the volume grows. This approach is an alternative to inverse blocking [52, 53]: instead of a matched coarse ensemble, which supplies the joint distribution of the long-distance modes implicitly, the marginal test targets the joint distribution of a few chosen variables, preferably those that local relaxation repairs most slowly, and needs it as input. The two could be combined, but such a hybrid has not been tested (Sec. VI).

The paper is organized as follows. Sections II and III give the general framework; Sec. II relates the ESS to the Rényi divergence and introduces the overlap flow, and Sec. III constructs the marginal test and the composite chain. Sections IV and V test both parts in parallel, exactly in the free field and numerically in two-dimensional ϕ4\phi^4 theory near its critical point. Section VI discusses what evaluating samplers by scale implies for their assessment and design, and Sec. VII concludes. Derivations are collected in Appendix A, Appendix B studies the composite chain at fixed correlation length, and Appendix C examines how to estimate and use the target marginal.

Effective sample size and coarse graining

We first relate the effective sample size to the Rényi divergence and define a coarse-grained version of it. The overlap identities apply to normalized nonnegative target and model distributions; the scaling statements require the assumptions specified below.

Rényi divergence

Let w=p/qw=p/q, with q>0q>0 wherever p>0p>0, so Eq[w]=1\mathbb{E}_q[w]=1. For independent samples ϕi∼q\phi_i\sim q and finite Eq[w2]\mathbb{E}_q[w^2], the usual estimate ESS^=(∑iwi)2/∑iwi2\widehat{\mathrm{ESS}}=(\sum_i w_i)^2/\sum_i w_i^2 satisfies

ESS^N→N→∞1Eq[w2]=e−D2(p∥q),(1)\frac{\widehat{\mathrm{ESS}}}{N}\xrightarrow{N\to\infty}\frac{1}{\mathbb{E}_q[w^2]}=e^{-D_2(p\Vert q)}, \tag*{(1)}
D2(p∥q)=log⁡∫dϕ p2q.D_2(p\Vert q)=\log\int d\phi\,\frac{p^2}{q}.

The ratio follows from Eq[w]=1\mathbb{E}_q[w]=1 and Eq[w2]=∫p2/q\mathbb{E}_q[w^2]=\int p^2/q. The normalization of pp or qq need not be known: with ℓ=log⁡w\ell=\log w given up to an additive constant,

D2=log⁡Eq[e2ℓ]−2log⁡Eq[eℓ],(2)D_2=\log\mathbb{E}_q[e^{2\ell}]-2\log\mathbb{E}_q[e^\ell], \tag*{(2)}

which does not change when a constant is added to ℓ\ell. If ℓ\ell is Gaussian with variance σ2\sigma^2, Eq. (2) gives D2=σ2D_2=\sigma^2. Thus ESS/N≃e−Var⁡(log⁡w)\mathrm{ESS}/N\simeq e^{-\operatorname{Var}(\log w)} is exact for Gaussian log weights; otherwise the exponential moments in Eq. (2), especially their sensitivity to the upper tail, must be considered.

For translation-invariant targets at fixed finite correlation length, a weak local mismatch gives an extensive D2D_2. A sampler defines an effective action −log⁡q-\log q; if it differs from SS by a local term, q=e−S−cA/Z(c)q=e^{-S-cA}/Z(c) with A=∑xO(x)A=\sum_x O(x), then, provided Z(c)Z(c) is finite in a neighborhood of c=0c=0 containing ±c\pm c (Appendix A § 3.0.1),

D2=log⁡Z(c)+log⁡Z(−c)−2log⁡Z(0)≃c2VχO,(3)D_2=\log Z(c)+\log Z(-c)-2\log Z(0)\simeq c^2V\chi_O, \tag*{(3)}
χO=∑x⟨O(x)O(0)⟩conn.\chi_O=\sum_x\langle O(x)O(0)\rangle_{\mathrm{conn}}.

The next term is c4κ4(A)/12c^4\kappa_4(A)/12, with κ4\kappa_4 the fourth cumulant. For several perturbations ∑iciAi\sum_i c_iA_i the leading term is ∑ijcicj⟨AiAj⟩conn\sum_{ij}c_ic_j\langle A_iA_j\rangle_{\mathrm{conn}}, which is the Fisher metric of the Gibbs family. The negative logarithm of the population ESS fraction is thus a second difference of the free energy and grows like VV when χO\chi_O and the higher cumulant densities remain finite. This need not imply inaccurate long-distance observables: D2D_2 simply counts how many regions carry the same small error.

Overlap flow

Let RΛR_\Lambda keep the Fourier modes with lattice momentum k^≤Λ\hat{k}\le\Lambda, where k^2=∑μ4sin⁡2(kμ/2)\hat{k}^2=\sum_\mu4\sin^2(k_\mu/2), and define D2(Λ)=D2(RΛp∥RΛq)D_2(\Lambda)=D_2(R_\Lambda p\Vert R_\Lambda q). This compares the Wilsonian effective actions after removing modes above Λ\Lambda. Coarse graining averages over the discarded fluctuations: RP/RQ=Eq[w∣R(U)]R_P/R_Q=\mathbb{E}_q[w\mid R(U)]. Jensen’s inequality then gives ∫Rq (RP/RQ)2≤Eq[w2]\int Rq\,(R_P/R_Q)^2\le\mathbb{E}_q[w^2] (Appendix A § 3.0.1; general Markov kernels, including other block-spin maps, are covered in Ref. [55]). Lowering Λ\Lambda therefore decreases D2(Λ)D_2(\Lambda) monotonically, from the global value at the lattice cutoff to the zero-mode value at Λ=0\Lambda=0. The proof needs neither Gaussianity nor translation invariance and holds with interactions.

Relevant operators and sectors

For a weak mismatch c∑xO(x)c\sum_xO(x) with scaling dimension Δ\Delta in the regime a≪ξ≪La\ll\xi\ll L, Eq. (3) gives D2≃c2VχOD_2\simeq c^2V\chi_O, and power counting gives χO\chi_O scaling as ξd−2Δ\xi^{d-2\Delta} (IR dominated) for Δ<d/2\Delta<d/2, as log⁡(ξ/a)\log(\xi/a) for Δ=d/2\Delta=d/2, and as ad−2Δa^{d-2\Delta} (UV dominated) for Δ>d/2\Delta>d/2. This boundary differs from the RG boundary y=d−Δ=0y=d-\Delta=0: long-distance scaling is especially sensitive to relevant couplings (y>0y>0), marginal directions, the zero mode and sector weights, while irrelevant couplings generally give suppressed corrections. Operators with Δ>d/2\Delta>d/2 contribute to D2D_2 mainly at the cutoff scale in this approximation. The mass mismatch (Δ=0\Delta=0) and the lattice artifact (Δ=4\Delta=4) of the two-dimensional free field (Sec. IV) are examples of the two extremes. If the configuration space splits into sectors aa (vacua, topological charge),

eD2(p∥q)=∑apa2qaeD2(p(⋅∣a)∥q(⋅∣a)),(4)e^{D_2(p\Vert q)}=\sum_a\frac{p_a^2}{q_a}e^{D_2(p(\cdot\mid a)\Vert q(\cdot\mid a))}, \tag*{(4)}

so a missing sector (qa=0q_a=0, pa>0p_a>0), as in a mode-collapsed sampler [56], makes D2D_2 infinite. Reweighting can correct wrong sector probabilities only if the proposal reaches every sector populated by the target.

Metropolis test on collective variables

Exact results for all observables require that the short-distance errors be removed as well. The inverse population ESS fraction of global reweighting, N/ESS≃eD2N/\mathrm{ESS}\simeq e^{D_2}, grows exponentially with VV when D2D_2 is extensive (Sec. II). For an observable depending only on a few variables, reweighting by their marginal density ratio instead incurs the ESS loss of that marginal; Sec. IV C demonstrates this in the free field. We use the same kind of low-dimensional information in a Metropolis test on a few collective variables and let exact local updates relax the remaining errors (Sec. III C).

Acceptance rate

In an independence MH chain with proposal y∼qy \sim q and current state x∼px \sim p, the mean acceptance rate at stationarity is exactly (Appendix A § 3.0.2)

References

  1. [1]K. Cranmer, G. Kanwar, S. Racanière, D. J. Rezende, and P. E. Shanahan, Advances in machine-learning-based sampling motivated by lattice quantum chromodynamics, Nat. Rev. Phys. 5, 526 (2023).
  2. [2]G. Aarts, K. Fukushima, T. Hatsuda, A. Ipp, S. Shi, L. Wang, and K. Zhou, Physics-driven learning for inverse problems in quantum chromodynamics, Nat. Rev. Phys. 7, 154 (2025).DOI
  3. [3]A. Tomiya, Machine learning for lattice QCD, J. Phys. Soc. Jpn. 94, 031006 (2025).DOI
  4. [4]M. S. Albergo, G. Kanwar, and P. E. Shanahan, Flow-based generative models for Markov chain Monte Carlo in lattice field theory, Phys. Rev. D 100, 034515 (2019).DOI
  5. [5]G. Kanwar, M. S. Albergo, D. Boyda, K. Cranmer, D. C. Hackett, S. Racanière, D. J. Rezende, and P. E. Shanahan, Equivariant flow-based sampling for lattice gauge theory, Phys. Rev. Lett. 125, 121601 (2020).DOI
  6. [6]L. Del Debbio, J. Marsh Rossney, and M. Wilson, Efficient modeling of trivializing maps for lattice ϕ4 theory using normalizing flows: A first look at scalability, Phys. Rev. D 104, 094507 (2021).DOI
  7. [7]R. Abbott, M. S. Albergo, A. Botev, D. Boyda, K. Cranmer, D. C. Hackett, A. G. D. G. Matthews, S. Racanière, A. Razavi, D. J. Rezende, F. Romero-López, P. E. Shanahan, and J. M. Urban, Aspects of scaling and scalability for flow-based sampling of lattice QCD, Eur. Phys. J. A 59, 257 (2023).arxiv.org/abs/2211.07541
  8. [8]M. C. N. Cheng and N. Stratikopoulou, Lecture notes on normalizing flows for lattice quantum field theories, SciPost Phys. Lect. Notes 110, 10.21468/SciPostPhysLectNotes.110 (2026).DOI
  9. [9]L. Wang, G. Aarts, and K. Zhou, Diffusion models as stochastic quantization in lattice field theory, J. High Energy Phys. 2024 (05), 060.DOI
  10. [10]K. Fukushima and S. Kamata, Stochastic quantization and diffusion models, J. Phys. Soc. Jpn. 94, 031010 (2025).DOI
  11. [11]O. Vega, J. Komijani, A. El-Khadra, and M. Marinkovic, Group-equivariant diffusion models for lattice field theory, J. High Energy Phys. 2026 (07), 213.DOI
  12. [12]Y.-y. Tan, G. Aarts, D. E. Habibi, B. Lucini, and L. Wang, Diffusion models for sampling near criticality in lattice field theories (2026), arXiv:2607.08505.arxiv.org/abs/2607.08505
  13. [13]D. Grinwald, A. Singha, K. Kahouli, K. Jansen, and S. Nakajima, Data free training of diffusion models for lattice field theories (2026), arXiv:2610.18863.arxiv.org/abs/2610.18863
  14. [14]D. Schubring, Diffusion models for the O(3) nonlinear sigma model (2026), arXiv:2610.39568.arxiv.org/abs/2610.39568
  15. [15]Q. Zhu, G. Aarts, W. Wang, K. Zhou, and L. Wang, Physics-conditioned diffusion models for lattice gauge theory, J. High Energy Phys. 2026 (03), 111.arxiv.org/abs/2502.05504
  16. [16]G. Aarts, D. E. Habibi, A. Ipp, D. I. Müller, T. R. Ranner, L. Wang, W. Wang, and Q. Zhu, Generalizable equivariant diffusion models for non-Abelian lattice gauge theory (2026), arXiv:2601.19552.arxiv.org/abs/2601.19552
  17. [17]H. Alharazin, Y. Yu. Panteleeva, and B.-D. Sun, Diffusion models for SU(2) lattice gauge theory in two dimensions, Phys. Rev. D 114, 014522 (2026).DOI
  18. [18]J. Komijani, M. K. Marinkovic, and L. Turgut, Diffusion model for SU(N) gauge theories (2026), arXiv:2605.06134.arxiv.org/abs/2605.06134
  19. [19]O. Vega and A. X. El-Khadra, Sampling the Schwinger model with gauge-equivariant diffusion, in Conference on Physics and AI (PAI 2026), Stanford University (2026) arXiv:2606.27481.DOI
  20. [20]A. Tomiya, Continuous Hasenbusch transport towards gauge diffusion with fermions (2026), arXiv:2609.15959.arxiv.org/abs/2609.15959
  21. [21]G. Aarts, D. E. Habibi, L. Wang, and K. Zhou, Combining complex Langevin dynamics with score-based and energy-based diffusion models, J. High Energy Phys. 2025 (12), 160.DOI
  22. [22]A. Tomiya, Lattice configuration generation with a self-learning diffusion model (2026), arXiv:2607.12587.arxiv.org/abs/2607.12587
  23. [23]L. Wang, Stochastic quantization as optimal control (2026), arXiv:2607.21436.arxiv.org/abs/2607.21436
  24. [24]S. Chen, M. Qian, G. Aarts, B. Lucini, and K. Zhou, Stochastic path sampler for lattice field theory (2026), arXiv:2606.13790.arxiv.org/abs/2606.13790
  25. [25]M. S. Albergo and E. Vanden-Eijnden, NETs: A non-equilibrium transport sampler, in Proceedings of the 42nd International Conference on Machine Learning, Proc. Mach. Learn. Res., Vol. 267 (2025) pp. 1026–1055, arXiv:2410.02711.arxiv.org/abs/2410.02711
  26. [26]K. A. Nicoli, S. Nakajima, N. Strodt hoff, W. Samek, K.-R. Müller, and P. Kessel, Asymptotically unbiased estimation of physical observables with neural samplers, Phys. Rev. E 101, 023304 (2020).
  27. [27]D. Albandea, L. Del Debbio, P. Hernández, R. Kenway, J. Marsh Rossney, and A. Ramos, Learning trivializing flows, Eur. Phys. J. C 83, 676 (2023).
  28. [28]J. Komijani and M. K. Marinkovic, Generative models for scalar field theories: how to deal with poor scaling?, Proc. Sci. LATTICE2022, 019 (2023).
  29. [29]F. Guth, S. Coste, V. De Bortoli, and S. Mallat, Wavelet score-based generative modeling, in Advances in Neural Information Processing Systems, Vol. 35 (2022) pp. 478–491.
  30. [30]J. Finkenrath, Fine grinding localized updates via gauge equivariant flows in the 2D Schwinger model, Proc. Sci. LATTICE2023, 022 (2024).
  31. [31]R. Abbott, D. Boyda, Y. Fu, D. C. Hackett, G. Kanwar, F. Romero-López, P. E. Shanahan, and J. M. Urban, Variance reduction in lattice QCD observables via normalizing flows, Phys. Rev. D 114, 014513 (2026).
  32. [32]R. Abbott, Y. Fu, D. C. Hackett, G. Kanwar, F. Romero-López, and P. E. Shanahan, Learning the generating functional for variance reduction in lattice QCD (2026), arXiv:2606.15986.arxiv.org/abs/2606.15986
  33. [33]M. Caselle, E. Cellini, A. Nada, and M. Panero, Stochastic normalizing flows as non-equilibrium transformations, J. High Energy Phys. 2022 (07), 015.DOI
  34. [34]A. Bulgarelli, E. Cellini, and A. Nada, Scaling of stochastic normalizing flows in SU(3) lattice gauge theory, Phys. Rev. D 111, 074517 (2025).DOI
  35. [35]C. Bonanno, A. Bulgarelli, E. Cellini, A. Nada, D. Panfalone, V. D’Acci acchino, and L. Verzichelli, Scaling flow-based approaches for topology sampling in SU(3) gauge theory, J. High Energy Phys. 2026 (04), 051.DOI
  36. [36]G. Kanwar and O. Vega, Annealed sampling of diffusion models in lattice field theory (2026), arXiv:2610.35792.arxiv.org/abs/2610.35792
  37. [37]D. Sexty, Accurate sampling from diffusion models (2026), arXiv:2609.29902.arxiv.org/abs/2609.29902
  38. [38]M. Gerdes, P. de Haan, C. Rainone, R. Bondesan, and M. C. N. Cheng, Learning lattice quantum field theories with equivariant continuous flows, SciPost Phys. 15, 238 (2023).DOI
  39. [39]S. Bacchio, P. Kessel, S. Schaefer, and L. Vaitl, Learning trivializing gradient flows for lattice gauge theories, Phys. Rev. D 107, L051504 (2023).DOI
  40. [40]J. Finkenrath, Tackling critical slowing down using global correction steps with equivariant flows: the case of the Schwinger model (2022), arXiv:2201.02216.arxiv.org/abs/2201.02216
  41. [41]S.-H. Li and L. Wang, Neural network renormalization group, Phys. Rev. Lett. 121, 260601 (2018).DOI
  42. [42]T. Marchand, M. Ozawa, G. Biroli, and S. Mallat, Multi-scale data-driven energy estimation and generation, Phys. Rev. X 13, 041038 (2023).
  43. [43]R. Abbott, M. S. Albergo, D. Boyda, D. C. Hackett, G. Kanwar, F. Romero-López, P. E. Shanahan, and J. M. Urban, Multiscale normalizing flows for gauge theories, Proc. Sci. LATTICE2023, 035 (2024).arxiv.org/abs/2404.10819
  44. [44]M. Bauer, R. Kapust, J. M. Pawlowski, and F. L. Temmen, Super-resolving normalising flows for lattice field theories, SciPost Phys. 19, 077 (2025).arxiv.org/abs/2412.12842
  45. [45]A. Singha, E. Cellini, K. A. Nicoli, K. Jansen, S. Kühn, and S. Nakajima, Multilevel generative samplers for investigating critical phenomena, in International Conference on Learning Representations (ICLR 2025) (2025) pp. 25450–25483, arXiv:2503.08918.arxiv.org/abs/2503.08918
  46. [46]A. Singha, J. Kauffmann, E. Cellini, K. Jansen, and S. Nakajima, Scalable generative sampling and multilevel estimation for lattice field theories near criticality, Phys. Rev. D 114, 074515 (2026).DOI
  47. [47]J. Cotler and S. Rezchikov, Renormalizing diffusion models (2023), arXiv:2308.12355.arxiv.org/abs/2308.12355
  48. [48]K. Masuki and Y. Ashida, Generative diffusion model with inverse renormalization group flows (2025), arXiv:2501.09064.arxiv.org/abs/2501.09064
  49. [49]K. Masuki and Y. Ashida, Renormalization group flow matching for scalable local generative modeling (2026), arXiv:2608.23696.arxiv.org/abs/2608.23696
  50. [50]R. Kapust, F. Ihssen, and J. M. Pawlowski, Generative sampling with physics-informed kernels, Proc. Sci. LATTICE2025, 055 (2026), arXiv:2510.26678.arxiv.org/abs/2510.26678
  51. [51]F. Ihssen, R. Kapust, and J. M. Pawlowski, Solving sign problems with physics-informed kernels (2026), arXiv:2603.03159.arxiv.org/abs/2603.03159
  52. [52]A. Hasenfratz, E. T. Neil, L. Parato, and N. Schwartz, Renormalization-guided inverse blocking for lattice field generation: construction and validation (2026), arXiv:2608.28580.arxiv.org/abs/2608.28580
  53. [53]A. Hasenfratz, E. T. Neil, L. Parato, and N. Schwartz, Renormalization-guided cascade upscaling for lattice field generation (2026), arXiv:2608.28581.arxiv.org/abs/2608.28581
  54. [54]M. Cè, L. Giusti, and S. Schaefer, Domain decomposition, multilevel integration, and exponential noise reduction in lattice QCD, Phys. Rev. D 93, 094507 (2016).DOI
  55. [55]T. van Erven and P. Harremöes, Rényi divergence and Kullback-Leibler divergence, IEEE Trans. Inf. Theory 60, 3797 (2014).
  56. [56]K. A. Nicoli, C. J. Anders, T. Hartung, K. Jansen, P. Kessel, and S. Nakajima, Detecting and mitigating mode-collapse for flow-based sampling of lattice field theories, Phys. Rev. D 108, 114501 (2023).arxiv.org/abs/2302.14082
  57. [57]L. O’Raifeartaigh, A. Wipf, and H. Yoneyama, The constraint effective potential, Nucl. Phys. B 271, 653 (1986).
  58. [58]A. Laio, G. Martinelli, and F. Sanfilippo, Metadynamics surfing on topology barriers: the CP−1 case, J. High Energy Phys. 2016 (07), 089.DOI
  59. [59]J. Feldman, Equivalence and perpendicularity of Gaussian processes, Pacific J. Math. 8, 699 (1958).DOI
  60. [60]J. Hájek, On a property of normal distributions of any stochastic process, Czechoslovak Math. J. 8, 610 (1958), in Russian.DOI
  61. [61]S. Chatterjee and P. Diaconis, The sample size required in importance sampling, Ann. Appl. Probab. 28, 1099 (2018).arxiv.org/abs/1511.01437
  62. [62]P. B. Mackenzie, An improved hybrid Monte Carlo method, Phys. Rev. Lett. B 226, 369 (1989).
  63. [63]Y.-y. Tan and L. Wang, phi4-critical-configs (Revision 972ab5f), Hugging Face dataset, https://huggingface.co/datasets/phi4-generative/phi4-critical-configs (2026).
  64. [64]G. Aarts, D. E. Habibi, L. Wang, and K. Zhou, On learning higher-order cumulants in diffusion models, Mach. Learn.: Sci. Technol. 6, 025004 (2025).DOI
  65. [65]R. C. Brower and P. Tamayo, Embedded dynamics for φ⁴ theory, Phys. Rev. Lett. 62, 1087 (1989).pubmed.ncbi.nlm.nih.gov/10039573
  66. [66]T. M. Cover and J. A. Thomas, Elements of Information Theory, 2nd ed. (Wiley, Hoboken, 2006) Sec. 2.8.
  67. [67]M. Creutz, Global Monte Carlo algorithms for many-fermion systems, Phys. Rev. D 38, 1228 (1988).
  68. [68]R. Gupta, G. W. Kilcup, and S. R. Sharpe, Tuning the hybrid Monte Carlo algorithm, Phys. Rev. D 38, 1278 (1988).pubmed.ncbi.nlm.nih.gov/9959266
  69. [69]J. Goodman and A. D. Sokal, Multigrid Monte Carlo method for lattice field theories, Phys. Rev. Lett. 56, 1015 (1986).pubmed.ncbi.nlm.nih.gov/10012035
  70. [70]M. G. Endres, R. C. Brower, W. Detmold, K. Orginos, and A. V. Pochinsky, Multiscale Monte Carlo equilibration: Pure Yang-Mills theory, Phys. Rev. D 92, 114516 (2015).
  71. [71]A. Singha, J. Kauffmann, K. Jansen, J. Finkenrath, V. Arora, and S. Nakajima, Sector-resolved flow sampling for topologically frozen lattice gauge theories (2026), arXiv:2609.07781.
  72. [72]S. Aoki, H. Fukaya, S. Hashimoto, and T. Onogi, Finite volume QCD at fixed topological charge, Phys. Rev. D 76, 054508 (2007).
  73. [73]M. Cè, C. Consonni, G. P. Engel, and L. Giusti, Non-Gaussianities in the topological charge distribution of the SU(3) Yang-Mills theory, Phys. Rev. D 92, 074502 (2015).
  74. [74]A. Vehtari, D. Simpson, A. Gelman, Y. Yao, and J. G. Gabry, Pareto smoothed importance sampling, J. Mach. Learn. Res. 25, 1 (2024), arXiv:1507.02646.

Paper details

vlnc:W_kx596ba7
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Claude Opus 5.5, Claude Fable 5.1 (bare), ChatGPT 6.0

View statement
Looking for related papers…
Contents