Particle Monte Carlo methods for Lattice Field Theory

2025-11-19T07:31:46Z

High-dimensional multimodal sampling problems from lattice field theory (LFT) have become important benchmarks for machine learning assisted sampling methods. We show that GPU-accelerated particle methods, Sequential Monte Carlo (SMC) and nested sampling, provide a strong classical baseline that matches or outperforms state-of-the-art neural samplers in sample quality and wall-clock time on standard scalar field theory benchmarks, while also estimating the partition function. Using only a single data-driven covariance for tuning, these methods achieve competitive performance without problem-specific structure, raising the bar for when learned proposals justify their training cost.

Isotensor $πππ$ scattering with a $ρ$ resonant subsystem from QCD

2025-11-18T17:25:53Z

This work presents a lattice quantum chromodynamics (QCD) determination of $πππ$ scattering amplitudes for the isospin-2 channel with angular momentum and parity $J^{P} = 1^+$. The calculation is performed using unphysically heavy light-quark masses, corresponding to a pion mass of $m_π \approx 400$~MeV, for which the $ρ$ meson manifests as a narrow resonance. The analysis employs a previously developed formalism that non-perturbatively relates the finite-volume spectra of two- and three-pion systems to their infinite-volume scattering amplitudes. We combine earlier lattice QCD results for $ππ$ systems with isospins 1 and 2 with new determinations of the isospin-2 three-pion finite-volume spectrum, obtained for three cubic, periodic lattice volumes with box length ranging from $\approx 4/m_π$ to $6/m_π$. These combined data constrain the low-lying partial waves of the infinite-volume three-pion K matrix, from which we solve a set of coupled integral equations to extract the corresponding scattering amplitudes.

Matching Lagrangian and Hamiltonian Simulations in (2+1)-dimensional U(1) Gauge Theory

2025-11-18T16:22:34Z

At finite lattice spacing, Lagrangian and Hamiltonian predictions differ due to discretization effects. In the Hamiltonian limit, i.e. at vanishing temporal lattice spacing $a_t$, the path integral approach in the Lagrangian formalism reproduces the results of the Hamiltonian theory. In this work, we numerically calculate the Hamiltonian limit of a U$(1)$ gauge theory in $(2+1)$ dimensions. This is achieved by Monte Carlo simulations in the Lagrangian formalism with lattices that are anisotropic in the time direction. For each ensemble, we determine the ratio between the temporal and spatial scale with the static quark potential and extrapolate to $a_t \to 0$. Our results are compared with the data from Hamiltonian simulations at small volumes, showing agreement within $<2σ$. These results can be used to match the two formalisms.

The Hidden Symmetries of Yang-Mills Theory in (3+1)-dimensions

2025-11-18T15:38:19Z

We show that classical, non-supersymmetric Yang-Mills theories coupled to spin-1/2 and spin-0 elementary matter fields, in (3+1)-dimensional Minkowski space-time, possess exact structures that resemble integrability, with an infinite number of conserved charges in involution. Such structures live in the space of non-abelian electric and magnetic charges, and are based on flat connections in generalized loop spaces, presenting an R-matrix, and Sklyanin relation. We present two novel symmetries of Yang-Mills theories. The first one corresponds to global transformations generated by the infinity of those conserved charges under the Poisson brackets. The gauge and matter fields, as well as Wilson lines and fluxes, have interesting transformation laws under such a global symmetry. The second one corresponds to symmetries of the integral Yang-Mills equations, which lead to the conserved charges. They generate an infinite-dimensional group, where the elements are holonomies of connections on the loop space of functions from the circle S^1 to the space-time. Our approach certainly applies to the Standard Model of the Fundamental Interactions. The conserved charges are gauge invariant, and so, in the case of QCD, they are color singlets and perhaps are not confined. Therefore, the hadrons may carry such charges. Our results open up the way for the construction of non-perturbative methods for Yang-Mills theories.

Phase diagram and eigenvalue dynamics of stochastic gradient descent in multilayer neural networks

2025-11-18T14:22:03Z

Hyperparameter tuning is one of the essential steps to guarantee the convergence of machine learning models. We argue that intuition about the optimal choice of hyperparameters for stochastic gradient descent can be obtained by studying a neural network's phase diagram, in which each phase is characterised by distinctive dynamics of the singular values of weight matrices. Taking inspiration from disordered systems, we start from the observation that the loss landscape of a multilayer neural network with mean squared error can be interpreted as a disordered system in feature space, where the learnt features are mapped to soft spin degrees of freedom, the initial variance of the weight matrices is interpreted as the strength of the disorder, and temperature is given by the ratio of the learning rate and the batch size. As the model is trained, three phases can be identified, in which the dynamics of weight matrices is qualitatively different. Employing a Langevin equation for stochastic gradient descent, previously derived using Dyson Brownian motion, we demonstrate that the three dynamical regimes can be classified effectively, providing practical guidance for the choice of hyperparameters of the optimiser.

Visualization of Entanglement Geometry by Structural Optimization of Tree Tensor Network

2025-11-18T11:14:08Z

In tensor-network analysis of quantum many-body systems, it is of crucial importance to employ a tensor network with a spatial structure suitable for representing the state of interest. In the previous work [Hikihara et al., Phys. Rev. Research 5, 013031 (2023)], we proposed a structural optimization algorithm for tree-tensor networks. In this paper, we apply the algorithm to the Rainbow-chain model, which has a product state of singlet pairs between spins separated by various distances as an approximate ground state. We then demonstrate that the algorithm can successfully visualize the spatial pattern of spin-singlet pairs in the ground state.

Cluster phenomena using few-body and Lattice QCD theories

2025-11-18T05:31:13Z

With the advancement of first-principles calculations for baryon-baryon interactions, it becomes possible to obtain reliable hyperon-nucleon potentials by lattice QCD simulations with the HAL QCD method. High-precision few-body methods, such as the Gaussian Expansion Method (GEM), are applicable to solve quantum few-body systems up to four- and five-body systems. By combining the HAL QCD potentials with the GEM, one can predict the level structure of novel hypernuclei prior to experimental observation. In this review, we utilize the lattice QCD $NΞ$ potential obtained by the HAL QCD method to investigate the few-body systems $NNΞ$ and $NNNΞ$. Our analysis indicates that the lightest bound $Ξ$ hypernucleus is the $NNNΞ$ system. To extract detailed information on the isospin and spin components of the $NΞ$ interaction, we perform a four-body calculation for the $ααNΞ$ system with the total isospin $T = 0$ and $T = 1$. We demonstrate that the level structure of this system is sensitive to the isospin and spin dependencies of the $NΞ$ interaction. Furthermore, we propose experimental investigations to produce the $NNNΞ$ and $ααNΞ$ systems via the $(K^-, K^+)$ and $(K^-, K^0)$ reactions on $^4$He and $^{10}$B targets, respectively.

A Machine Learning study of the two-dimensional antiferromagnetic Ising model with nearest and next-to-nearest interactions on the triangular lattice

2025-11-18T03:11:18Z

We study the phase transitions of the two-dimensional antiferromagnetic Ising model with nearest $J_1$ and next-to-nearest $J_2$ interactions on the triangular lattice for $J_2/J_1 = 0.1, 0.5$ and 1.0. The method of supervised neural networks (NN) is employed for the investigation. While supervised NN is used, no real spin configurations are needed for the training. In addition, two kinds of configurations having their spins be arranged in a staggered pattern are considered as the training set. Remarkably, with this unconventional training strategy, not only the critical temperatures of the studied $J_2/J_1$ are computed accurately by the resulting NN, but also the nature of the investigated phase transitions are determined correctly. Specifically, the phase transitions associated with $J_2/J_1 = 0.1, 0.5$ and 1.0 are first order. These conclusions are consistent with the known results obtained by other methods. Since the training strategy is simple, the NN calculations is highly efficient. It remains to examine whether the unconventional training approach considered in this study can be used to investigate other models with untypical phase transitions or with nontrivial ground state configurations.

Lattice QCD Benchmark of Proton Helicity and Flavor-Dependent Unpolarized Transverse Momentum-Dependent Parton Distribution Functions at Physical Quark Masses

2025-11-17T14:04:26Z

We present the first lattice QCD calculations of the isovector helicity transverse momentum-dependent parton distribution function (TMDPDF) and the flavor-dependent unpolarized TMDPDFs for up and down quarks in the proton. Our computations utilize domain-wall fermion discretization with physical quark masses. Employing Coulomb-gauge-fixed quark correlation functions within the large-momentum effective theory framework, we access nonperturbative transverse quark separations $b_T$ up to approximately 1 fm, corresponding to transverse momenta as low as 200 MeV. Based on the quasi-TMD factorization theorem, we construct renormalization-group-invariant ratios that are equal to the corresponding light-cone TMDPDF ratios. At moderate $x$, our results reveal that the isovector helicity and unpolarized TMDPDFs exhibit nearly identical transverse structure up to a normalization factor, and the unpolarized distributions display only mild flavor dependence. These findings not only support key trends observed in recent global analyses but also provide robust, nonperturbative constraints that can distinguish between different parameterizations. This work establishes a first-principles benchmark for TMDPDFs, offering valuable input for ongoing and future experimental efforts to map the proton's three-dimensional structure.

Exact Chiral Symmetries of 3+1D Hamiltonian Lattice Fermions

2025-11-16T06:41:50Z

We construct Hamiltonian models on a 3+1d cubic lattice for a single Weyl fermion and for a single Weyl doublet protected by exact (as opposed to emergent) chiral symmetries. In the former, we find a not-on-site, non-compact chiral symmetry which can be viewed as a Hamiltonian analog of the Ginsparg-Wilson symmetry in Euclidean lattice models of Weyl fermions. In the latter, we combine an on-site $U(1)$ symmetry with a not-on-site $U(1)$ symmetry, which together generate the $SU(2)$ flavor symmetry of the doublet at low energies, while in the UV they generate an algebra known in integrability as the Onsager algebra. This latter model is in fact the celebrated magnetic Weyl semimetal which is known to have a chiral anomaly from the action of $U(1)$ and crystalline translation, that gives rise to an anomalous Hall response - however reinterpreted in our language, it has two exact $U(1)$ symmetries that gives rise to the global $SU(2)$ anomaly which protects the gaplessness even when crystalline translations are broken. We also construct an exact symmetry-protected single Dirac cone in 2+1d with the $U(1) \rtimes T$ parity anomaly. Our constructions evade both old and recently-proven no-go theorems by using not-on-siteness in a crucial way, showing our results are sharp.

A universal framework for the quantum simulation of Yang-Mills theory

2025-11-15T23:03:42Z

We provide a universal framework for the quantum simulation of SU(N) Yang--Mills theories on fault-tolerant digital quantum computers adopting the orbifold lattice formulation. As warm-up examples, we also consider simple models, including scalar field theory and the Yang--Mills matrix model, to illustrate the universality of our formulation, which shows up in the fact that the truncated Hamiltonian can be expressed in the same simple form for any N, any dimension, and any lattice size, in stark contrast to the popular approach based on the Kogut--Susskind formulation. In all these cases, the truncated Hamiltonian can be programmed on a quantum computer using only standard tools well-established in the field of quantum computation. As a concrete application of this universal framework, we consider Hamiltonian time evolution by Suzuki--Trotter decomposition. This turns out to be a straightforward task due to the simplicity of the truncated Hamiltonian. We also provide a simple circuit structure that contains only CNOT and one-qubit gates, independent of the details of the theory investigated.

$CPT$-Symmetric Kähler-Dirac Fermions

2025-11-14T18:39:56Z

Kähler-Dirac (KD) spinors have generated excitement in the lattice gauge theory community, as a way to (i) deal with the ``fermion doubling" problems that plague ordinary (Dirac, Majorana, or Weyl) spinors when discretized on a lattice, and (ii) help explain the structure of the standard model. But if one naively quantizes this theory in Lorentzian signature, problems arise: half the KD fields have the ``wrong sign" Lagrangian, and give rise to negative norm states. Here we propose a new resolution/interpretation: the KD field actually lives on a two-sheeted spacetime, with the sheets related by $PT$ symmetry or, alternatively, by $i\leftrightarrow-i$. And, to avoid any unphysical interactions between the two sheets, the KD field obeys a reality condition (which we call the ``KD-Majorana condition"), which forces every particle on one sheet to be accompanied by a mirror (anti-)particle on the other sheet. We discuss how the standard model fits in this framework, how the fermion (kinetic and Yukawa) terms simplify, and how it may relate to the CPT-symmetric universe model.

Window observables for benchmarking parton distribution functions

2025-11-14T15:43:40Z

Global analysis of collider and fixed-target experimental data and calculations from lattice quantum chromodynamics (QCD) are used to gain complementary information on the structure of hadrons. We propose novel ``window observables'' that allow for higher precision cross-validation between the different approaches, a critical step for studies that wish to combine the datasets. Global analyses are limited by the kinematic regions accessible to experiment, particularly in a range of Bjorken-$x$, and lattice QCD calculations also have limitations requiring extrapolations to obtain the parton distributions. We provide two different ``window observables'' that can be defined within a region of $x$ where extrapolations and interpolations in global analyses remain reliable and where lattice QCD results retain sensitivity and precision.

On the origin of the $Nc^1$ scaling in the confined but chirally symmetric phase at high T

2025-11-14T13:50:34Z

There is lattice evidence that the QCD matter above the chiral restoration temperature Tch and below the deconfinement temperature Td, called stringy fluid, is characterized by an approximate chiral spin symmetry, which is a symmetry of confinement in QCD with light quarks. The energy density, pressure and entropy density in the stringy fluid scale as Nc^1, which is in contrast to the Nc^0 scaling in the hadron gas and to the Nc^2 scaling in the quark-gluon plasma. Here we clarify the origin of the Nc^1 scaling. We employ a solvable field-theoretical large $N_c$ chirally symmetric and confining model. In vacuum the confining potential induces a spontaneous breaking of chiral symmetry. The mesons are spatially localized states of quarks and antiquarks. Still in the confining regime the system undergoes the chiral restoration phase transition at $T_{ch}$ because of Paili blocking of the quark levels required for the existence of the quark condensate, by the thermal excitation of quarks and antiquarks. The same Paili blocking leads to a delocalization of the color singlet low-spin meson-like states that become infinitely large in the chiral limit. Consequently the stringy fluid represents a very dense medium of the overlapping huge color-singlet low-spin quark-antiquark systems. The Bethe-Salpeter equation that determines the rest-frame excitation energies of the color-singlet quark-antiquark system is Nc-independent both in vacuum and in the medium in the confining regime. The excitation energy of the quark-antiquark color-singlet systems scales as Nc^0, i.e. as meson mass in vacuum. The Nc^1 scaling of the energy density in the stringy fluid is provided by the fluctuations of the color-singlet quark-antiquark systems.

Confinement-Higgs and deconfinement-Higgs transitions in four-dimensional SU(2) LGT at finite temperature

2025-11-14T12:37:14Z

We re-examine by numerical simulation the phase structure of the (3+1)-dimensional SU(2) lattice gauge theory (LGT) with gauge fields coupled to Higgs fields at finite temperature. Concretely, we explore two different order parameters which are able to distinguish the three phases of the theory: (i) the Fredenhagen-Marcu operator used to discriminate between deconfinement and confinement/Higgs phases and (ii) the Greensite-Matsuyama overlap operator proposed recently to distinguish confinement and Higgs phases.