Publications

Order by: publication date number of citations

• Ground state of the staggered Heisenberg-Γ honeycomb model in a magnetic field

  Mojtaba Ahmadi-Yazdi, Mohammad-Hossein Zare, Hamid Mosadeq, Farhad Fazileh

  SciPost Phys. Core 7, 030 (2024) · published 15 May 2024 | Toggle abstract · pdf

  We study the ground state properties of the S=1/2 staggered Heisenberg-$\Gamma$ honeycomb model under a magnetic field based on analytical and numerical methods. Our calculations show that the conventional zigzag and stripy phases are favored because of the staggered Heisenberg interaction away from the pure $\Gamma$ limit. In our classical analysis, we find that the field induces a series of competing magnetic phases with relatively large unit cells in the region sandwiched between the two magnetic phases with long-range ordering. In the quantum treatment, these large magnetic unit cells are destabilized by strong quantum fluctuations that result in the stabilization of a gapless quantum spin liquid behavior. In a honeycomb $\Gamma$ magnet, we disclose an intermediate-field gapless quantum spin liquid phase driven by a tilted field away from the out-of-plane direction only for a narrow region between the low-field zigzag and high-field fully polarized phases.

• The quantum cartpole: A benchmark environment for non-linear reinforcement learning

  Kai Meinerz, Simon Trebst, Mark Rudner, Evert van Nieuwenburg

  SciPost Phys. Core 7, 026 (2024) · published 7 May 2024 | Toggle abstract · pdf

  Feedback-based control is the de-facto standard when it comes to controlling classical stochastic systems and processes. However, standard feedback-based control methods are challenged by quantum systems due to measurement induced backaction and partial observability. Here we remedy this by using weak quantum measurements and model-free reinforcement learning agents to perform quantum control. By comparing control algorithms with and without state estimators to stabilize a quantum particle in an unstable state near a local potential energy maximum, we show how a trade-off between state estimation and controllability arises. For the scenario where the classical analogue is highly nonlinear, the reinforcement learned controller has an advantage over the standard controller. Additionally, we demonstrate the feasibility of using transfer learning to develop a quantum control agent trained via reinforcement learning on a classical surrogate of the quantum control problem. Finally, we present results showing how the reinforcement learning control strategy differs from the classical controller in the non-linear scenarios.

• Physics-informed neural networks viewpoint for solving the Dyson-Schwinger equations of quantum electrodynamics

  Rodrigo Carmo Terin

  SciPost Phys. Core 8, 054 (2025) · published 20 August 2025 | Toggle abstract · pdf

  Physics-informed neural networks (PINNs) are employed to solve the Dyson-Schwinger equations of quantum electrodynamics (QED) in Euclidean space, with a focus on the non-perturbative generation of the fermion's dynamical mass function in the Landau gauge. By inserting the integral equation directly into the loss function, our PINN framework enables a single neural network to learn a continuous and differentiable representation of the mass function over a spectrum of momenta. Also, we benchmark our approach against a traditional numerical algorithm showing the main differences among them. Our novel strategy, which is expected to be extended to other quantum field theories, is the first step towards forefront applications of machine learning in high-level theoretical physics.

• Rapid measurements and phase transition detections made simple by AC-GANs

  Jiewei Ding, Ho-Kin Tang, Wing Chi Yu

  SciPost Phys. Core 7, 035 (2024) · published 11 June 2024 | Toggle abstract · pdf

  In recent years, significant attention has been paid to using end-to-end neural networks for analyzing Monte Carlo data. However, the exploration of non-end-to-end generative adversarial neural networks remains limited. Here, we investigate classical many-body systems using generative adversarial neural networks. We employ the conditional generative adversarial network with an auxiliary classifier (AC-GAN) and integrate self-attention layers into the generator. This modification enables the network learn the distribution of the two-dimensional (2D) XY model's spin configurations and the physical quantities of interest. Utilizing the symmetry of the systems, we discover that AC-GAN can be trained with a very small raw dataset. This approach allows us to obtain reliable measurements for models typically demanding large samples, such as the large-sized 2D XY and the 3D constrained Heisenberg models. Moreover, we demonstrate the capability of AC-GAN to identify the phase transition points by quantifying the distribution changes in the spin configurations of the systems.

• Exact solution for two $\delta$-interacting bosons on a ring in the presence of a $\delta$-barrier: Asymmetric Bethe Ansatz for spatially odd states

  Maxim Olshanii, Mathias Albert, Gianni Aupetit-Diallo, Patrizia Vignolo, Steven G. Jackson

  SciPost Phys. Core 8, 083 (2025) · published 12 November 2025 | Toggle abstract · pdf

  In this article, we apply the recently proposed Asymmetric Bethe Ansatz method to the problem of two one-dimensional, short-range-interacting bosons on a ring in the presence of a $\delta$-function barrier. Only half of the Hilbert space—namely, the two-body states that are odd under point inversion about the position of the barrier—is accessible to this method. The other half is presumably non-integrable. We consider benchmarking the recently proposed $1/g$ expansion about the hard-core boson point [D. Sen, Int. J. Mod. Phys. A 14, 1789 (1999); A. G. Volosniev et al., Nat. Commun. 5, 5300 (2014)] as one application of our results. Additionally, we find that when the $\delta$-barrier is converted to a $\delta$-well with strength equal to that of the particle-particle interaction, the system exhibits the spectrum of its non-interacting counterpart while its eigenstates display features of a strongly interacting system. We discuss this phenomenon in the "Summary and Future Research" section of our paper.

• Can one hear the full nonlinearity of a PDE from its small excitations?

  Maxim Olshanii, Danshyl Boodhoo

  SciPost Phys. Core 8, 055 (2025) · published 25 August 2025 | Toggle abstract · pdf

  In this article, we show how one can restore an unknown nonlinear partial differential equation of a sine-Gordon type from its linearization around an unknown stationary kink. The key idea is to regard the ground state of the linear problem as the translation-related Goldstone mode of the nonlinear PDE sought after.

• Learning selection cuts with gradients

  Mike Hance, Juan Robles

  SciPost Phys. Core 8, 079 (2025) · published 7 November 2025 | Toggle abstract · pdf

  Many analyses in high-energy physics rely on selection thresholds (cuts) applied to detector, particle, or event properties. Initial cut values can often be guessed from physical intuition, but cut optimization, especially for multiple features, is commonly performed by hand, or skipped entirely in favor of multivariate algorithms like BDTs or neural networks. We revisit this problem, and develop a cut optimization approach based on gradient descent. Cut thresholds are learned as parameters of a network with a simple architecture, and can be tuned to achieve a target signal efficiency through the use of custom loss functions. Contractive terms in the loss can be used to ensure a smooth evolution of cuts as functions of efficiency, particle kinematics, or event features. The method is used to classify events in a search for Supersymmetry, and the performance is compared with common classification tools. An implementation of this approach is available in a public code repository and python package.

• Gauge theory for topological waves in continuum fluids with odd viscosity

  Keisuke Fujii, Yuto Ashida

  SciPost Phys. Core 8, 058 (2025) · published 28 August 2025 | Toggle abstract · pdf

  We consider two-dimensional continuum fluids with odd viscosity under a chiral body force. The chiral body force makes the low-energy excitation spectrum of the fluids gapped, and the odd viscosity allows us to introduce the first Chern number of each energy band in the fluids. Employing a mapping between hydrodynamic variables and U(1) gauge field strengths, we derive a U(1) gauge theory for topologically nontrivial waves. The resulting U(1) gauge theory is given by the Maxwell-Chern-Simons theory with an additional term associated with odd viscosity. We then explicitly solve the equations of motion for the gauge fields in the presence of the boundary and find edge mode solutions. We finally confirm the bulk-boundary correspondence in the context of continuum systems.

• Freelance holography, part II: Moving boundary in gauge/gravity correspondence

  Aliasghar Parvizi, Mohammad M. Sheikh-Jabbari, Vahid Taghiloo

  SciPost Phys. Core 8, 075 (2025) · published 31 October 2025 | Toggle abstract · pdf

  We continue developing the freelance holography program, formulating gauge/gravity correspondence where the gravity side is formulated on a space bounded by a generic timelike codimension-one surface inside AdS and arbitrary boundary conditions are imposed on the gravity fields on the surface. Our analysis is performed within the Covariant Phase Space Formalism (CPSF). We discuss how a given boundary condition on the bulk fields on a generic boundary evolves as we move the boundary to another boundary inside AdS and work out how this evolution is encoded in deformations of the holographic effective boundary theory. Our analyses here extend the extensively studied $\text{T}\bar{\text{T}}$-deformation by relaxing the boundary conditions at asymptotic AdS or at the cutoff surface to be any arbitrary one (besides Dirichlet). We discuss some of the implications of our general freelance holography setting.

• Pair binding and Hund’s rule breaking in high-symmetry fullerenes

  Roman Rausch, Christoph Karrasch

  SciPost Phys. Core 9, 003 (2026) · published 20 January 2026 | Toggle abstract · pdf

  Highly-symmetric molecules often exhibit degenerate tight-binding states at the Fermi edge. This typically results in a magnetic ground state if small interactions are introduced in accordance with Hund's rule. In some cases, Hund's rule may be broken, which signals pair binding and goes hand-in-hand with an attractive pair-binding energy. We investigate pair binding and Hund's rule breaking for the Hubbard model on high-symmetry fullerenes C$_{20}$, C$_{20}$, C$_{40}$, and C$_{60}$ by using large-scale density-matrix renormalization group calculations. We exploit the SU(2) spin symmetry, the U(1) charge symmetry, and optionally the $\mathbb{Z}_N$ spatial rotation symmetry of the problem. For C$_{20}$, our results agree well with available exact-diagonalization data, but our approach is numerically much cheaper. We find a Mott transition at $U_c\sim2.2t$, which is much smaller than the previously reported value of $U_c\sim4.1t$ that was extrapolated from a few datapoints. We compute the pair-binding energy for arbitrary values of $U$ and observe that it remains overall repulsive. For larger fullerenes, we are not able to evaluate the pair binding energy with sufficient precision, but we can still investigate Hund's rule breaking. For C$_{28}$, we find that Hund's rule is fulfilled with a magnetic spin-2 ground state that transitions to a spin-1 state at $U_{c,1}\sim 5.4t$ before the eventual Mott transition to a spin singlet takes place at $U_{c,2}\sim 11.6t$. For C$_{40}$, Hund's rule is broken in the singlet ground state at half filling, but is restored if the system is doped with one electron. Hund's rule is also broken for C$_{60}$, and the doping with two or three electrons results in a minimum-spin state. Our results are consistent with an electronic mechanism of superconductivity for C$_{60}$ lattices. We speculate that the high geometric frustration of small fullerenes is detrimental to pair binding.