Publications

Order by: publication date number of citations

• Mutual multilinearity of nonequilibrium network currents

  Sara Dal Cengio, Pedro E. Harunari, Vivien Lecomte, Matteo Polettini

  SciPost Phys. 19, 111 (2025) · published 27 October 2025 | Toggle abstract · pdf

  Continuous-time Markov chains have been successful in modelling systems across numerous fields, with currents being fundamental entities that describe the flows of energy, particles, individuals, chemical species, information, or other quantities. They apply to systems described by agents transitioning between vertices along the edges of a network (at some rate in each direction). It has recently been shown by the authors that, at stationarity, a hidden linearity exists between currents that flow along edges: if one controls the current of a specific "input" edge (by tuning transition rates along it), any other current is a linear-affine function of the input current [Phys. Rev. Lett. 133, 047401 (2024)]. In this paper, we extend this result to the situation where one controls the currents of several edges, and prove that other currents are in linear-affine relation with the input ones. Two proofs with distinct insights are provided: the first relies on Kirchhoff's current law and reduces the input set inductively through graph analysis, while the second utilizes the resolvent approach via a Laplace transform in time. We obtain explicit expressions for the current-to-current susceptibilities, which allow one to map current dependencies through the network. We also verify from our expression that Kirchhoff's current law is recovered as a limiting case of our mutual linearity. Last, we uncover that susceptibilities can be obtained from fluctuations when the reference system is originally at equilibrium.

• Noninvertible symmetry and topological holography for modulated SPT in one dimension

  Jintae Kim, Yizhi You, Jung Hoon Han

  SciPost Phys. 19, 110 (2025) · published 24 October 2025 | Toggle abstract · pdf

  We examine noninvertible symmetry (NIS) in one-dimensional (1D) symmetry-protected topological (SPT) phases protected by dipolar and exponential-charge symmetries, which are two key examples of modulated SPT (MSPT). To set the stage, we first study NIS in the $\mathbb{Z}_N × \mathbb{Z}_N$ cluster model, extending previous work on the $\mathbb{Z}_2 × \mathbb{Z}_2$ case. For each symmetry type (charge, dipole, exponential), we explicitly construct the noninvertible Kramers-Wannier (KW) and Kennedy-Tasaki (KT) transformations, revealing dual models with spontaneous symmetry breaking (SSB). The resulting symmetry group structure of the SSB model is rich enough that it allows the identification of other SSB models with the same symmetry. Using these alternative SSB models and KT duality, we generate novel MSPT phases distinct from those associated with the standard decorated domain wall picture, and confirm their distinctiveness by projective symmetry analyses at their interfaces. Additionally, we establish a topological-holographic correspondence by identifying the 2D bulk theories-two coupled layers of toric codes (charge), anisotropic dipolar toric codes (dipole), and exponentially modulated toric codes (exponential)-whose boundaries host the respective 1D MSPT phases.

  1 citation

• Precise quantum-geometric electronic properties from first principles

  José Luís Martins, Carlos Loia Reis, Ivo Souza

  SciPost Phys. 19, 109 (2025) · published 24 October 2025 | Toggle abstract · pdf

  The calculation of quantum-geometric properties of Bloch electrons – Berry curvature, quantum metric, orbital magnetic moment and effective mass – was implemented in a pseudopotential plane-wave code. The starting point was the first derivative of the periodic part of the wavefunction $\psi_{k}(r)$ with respect to wavevector $k$. This was evaluated with perturbation theory by solving a Sternheimer equation. Comparison of effective masses obtained from perturbation theory for silicon and gallium arsenide with carefully-converged numerical second derivatives of band energies confirmed the high precision of the method. Calculations of quantum-geometric quantities for gapped graphene were performed by adding a bespoke symmetry-breaking potential to first-principles graphene. As the two bands near the opened gap are reasonably isolated, the results could be compared with those obtained from an analytical two-band model, allowing to assess the strengths and limitations of such widely-used models. The final application was trigonal tellurium, where some quantum-geometric quantities flip sign with chirality.

• Accurate surrogate amplitudes with calibrated uncertainties

  Henning Bahl, Nina Elmer, Luigi Favaro, Manuel Haußmann, Tilman Plehn, Ramon Winterhalder

  SciPost Phys. Core 8, 073 (2025) · published 24 October 2025 | Toggle abstract · pdf

  Neural networks for LHC physics have to be accurate, reliable, and controlled. Using neural surrogates for the prediction of loop amplitudes as a use case, we first show how activation functions are systematically tested with Kolmogorov-Arnold Networks. Then, we train neural surrogates to simultaneously predict the target amplitude and an uncertainty for the prediction. We disentangle systematic uncertainties, learned by a well-defined likelihood loss, from statistical uncertainties, which require the introduction of Bayesian neural networks or repulsive ensembles. We test the coverage of the learned uncertainties using pull distributions to quantify the calibration of cutting-edge neural surrogates.

• A Lorentz-equivariant transformer for all of the LHC

  Johann Brehmer, Víctor Bresó, Pim de Haan, Tilman Plehn, Huilin Qu, Jonas Spinner, Jesse Thaler

  SciPost Phys. 19, 108 (2025) · published 23 October 2025 | Toggle abstract · pdf

  We show that the Lorentz-Equivariant Geometric Algebra Transformer (L-GATr) yields state-of-the-art performance for a wide range of machine learning tasks at the Large Hadron Collider. L-GATr represents data in a geometric algebra over space-time and is equivariant under Lorentz transformations. The underlying architecture is a versatile and scalable transformer, which is able to break symmetries if needed. We demonstrate the power of L-GATr for amplitude regression and jet classification, and then benchmark it as the first Lorentz-equivariant generative network. For all three LHC tasks, we find significant improvements over previous architectures.

  2 citations

• Sequential adiabatic generation of chiral topological states

  Xie Chen, Michael Hermele, David T. Stephen

  SciPost Phys. 19, 107 (2025) · published 23 October 2025 | Toggle abstract · pdf

  In previous work, it was shown that non-trivial gapped states can be generated from a product state using a sequential quantum circuit. Explicit circuit constructions were given for a variety of gapped states at exactly solvable fixed points. In this paper, we show that a similar generation procedure can be established for chiral topological states as well, despite the fact that they lack a zero-correlation-length exactly solvable form. Instead of sequentially applying local unitary gates, we sequentially evolve the Hamiltonian by changing local terms in one subregion and then the next. The Hamiltonian remains gapped throughout the process, giving rise to an adiabatic evolution mapping the ground state from a product state to a chiral topological state. We demonstrate such a sequential adiabatic generation process for free fermion chiral states like the Chern Insulator and the $p+ip$ superconductor. Moreover, we show that coupling a quantum state to a discrete gauge group can be achieved through a sequential quantum circuit, thereby generating interacting chiral topological states from the free fermion ones.

• Simplified derivations for high-dimensional convex learning problems

  David G. Clark, Haim Sompolinsky

  SciPost Phys. Lect. Notes 105 (2025) · published 23 October 2025 | Toggle abstract · pdf

  Statistical-physics calculations in machine learning and theoretical neuroscience often involve lengthy derivations that obscure physical interpretation. Here, we give concise, non-replica derivations of several key results and highlight their underlying similarities. In particular, using a cavity approach, we analyze three high-dimensional learning problems: perceptron classification of points, perceptron classification of manifolds, and kernel ridge regression. These problems share a common structure—a bipartite system of interacting feature and datum variables—enabling a unified analysis. Furthermore, for perceptron-capacity problems, we identify a symmetry that allows derivation of correct capacities through a naïve method.

• An introduction to large deviations with applications in physics

  Ivan N. Burenev, Daniël W. H. Cloete, Vansh Kharbanda, Hugo Touchette

  SciPost Phys. Lect. Notes 104 (2025) · published 23 October 2025 | Toggle abstract · pdf

  These notes are based on the lectures that one of us (HT) gave at the Summer School on the "Theory of Large Deviations and Applications," held in July 2024 at Les Houches in France. They present the basic definitions and mathematical results that form the theory of large deviations, as well as many simple motivating examples of applications in statistical physics, which serve as a basis for the many other lectures given at the school that covered more specific applications in biophysics, random matrix theory, nonequilibrium systems, geophysics, and the simulation of rare events, among other topics. These notes extend the lectures, which can be accessed online, by presenting exercises and pointer references for further reading.

• TorchGDM: A GPU-accelerated Python toolkit for multi-scale electromagnetic scattering with automatic differentiation

  Sofia Ponomareva, Adelin Patoux, Clément Majorel, Antoine Azéma, Aurélien Cuche, Christian Girard, Arnaud Arbouet, Peter R. Wiecha

  SciPost Phys. Codebases 60 (2025) · published 22 October 2025 | Toggle abstract · pdf

  We present "torchGDM", a numerical framework for nano-optical simulations based on the Green's Dyadic Method (GDM). This toolkit combines a hybrid approach, allowing for both fully discretized nano-structures and structures approximated by sets of effective electric and magnetic dipoles. It supports simulations in three dimensions and for infinitely long, two-dimensional structures. This capability is particularly suited for multi-scale modeling, enabling accurate near-field calculations within or around a discretized structure embedded in a complex environment of scatterers represented by effective models. Importantly, torchGDM is entirely implemented in PyTorch, a well-optimized and GPU-enabled automatic differentiation framework. This allows for the efficient calculation of exact derivatives of any simulated observable with respect to various inputs, including positions, wavelengths or permittivity, but also intermediate parameters like Green's tensor components, which can be interesting for physics informed deep learning applications. We anticipate that this toolkit will be valuable for applications merging nano-photonics and machine learning, as well as for solving nano-photonic optimization and inverse problems, such as the global design and characterization of metasurfaces, where optical interactions between structures are critical.

  1 citation

• Codebase release 0.56 for TorchGDM

  Sofia Ponomareva, Adelin Patoux, Clément Majorel, Antoine Azéma, Aurélien Cuche, Christian Girard, Arnaud Arbouet, Peter R. Wiecha

  SciPost Phys. Codebases 60-r0.56 (2025) · published 22 October 2025 | Toggle abstract · src

  We present "torchGDM", a numerical framework for nano-optical simulations based on the Green's Dyadic Method (GDM). This toolkit combines a hybrid approach, allowing for both fully discretized nano-structures and structures approximated by sets of effective electric and magnetic dipoles. It supports simulations in three dimensions and for infinitely long, two-dimensional structures. This capability is particularly suited for multi-scale modeling, enabling accurate near-field calculations within or around a discretized structure embedded in a complex environment of scatterers represented by effective models. Importantly, torchGDM is entirely implemented in PyTorch, a well-optimized and GPU-enabled automatic differentiation framework. This allows for the efficient calculation of exact derivatives of any simulated observable with respect to various inputs, including positions, wavelengths or permittivity, but also intermediate parameters like Green's tensor components, which can be interesting for physics informed deep learning applications. We anticipate that this toolkit will be valuable for applications merging nano-photonics and machine learning, as well as for solving nano-photonic optimization and inverse problems, such as the global design and characterization of metasurfaces, where optical interactions between structures are critical.

  1 citation