-
Data-driven models can accurately describe and predict the dynamical properties of glass-forming liquids from structural data. Accurate predictions, however, do not guarantee an understanding of the underlying physical phenomena and the key factors that control them. In this paper, we illustrate the merits and limitations of linear regression models of glassy dynamics built on high-dimensional structural descriptors. By analyzing data for a two-dimensional glass model, we show that several descriptors commonly used in glass-transition studies display multicollinearity, which hinders the interpretability of linear models. Ridge regression suppresses some of the shortcomings of multicollinearity, but its solutions are not concise enough to be physically interpretable. Only by using dimensional reduction techniques we do eventually obtain linear models that strike a balance between prediction accuracy and interpretability. Our analysis points to a key role of local packing and composition fluctuations in the glass model under study.
"Defects, Corrugation and Temperature Govern Rarefied-Air Drag on Graphene Coatings"
Samuel Cajahuaringa, Davide Bidoggia, Maria Peressi, Antimo Marrazzo
Physics of Fluids 38, 051706 (2026)
In rarefied atmospheric environments, where continuum fluid dynamics breaks down, aerodynamic drag is governed by gas-surface momentum exchange, making surface structure and chemistry key design knobs. Using molecular dynamics simulations, we show that coating the $α$-Al2O3(0001) surface with graphene markedly reduces the tangential momentum accommodation coefficient (TMAC) of N2, shifting scattering toward more specular reflection and thereby lowering drag; we further benchmark this response against graphite. The reduction strengthens up to 900 K. While structural defects can increase TMAC via defect-induced corrugation and local atomic and electronic rearrangements, graphene retains its performance at experimentally relevant defect densities.
"Interaction-induced phases in the half-filled Bernevig-Hughes-Zhang model in one dimension"
Roberta Favata, Davide Piccioni, Alberto Parola, Federico Becca
Phys. Rev. B 111, 155105 (2025)
We explore the ground-state properties of a one-dimensional model with two orbitals per site, where, in addition to atomic energies $\pm M$, intra- and inter-orbital hoppings, the intra-orbital Hubbard ($U$) and nearest-neighbor density-density ($V$) repulsions are included. Our results are primarily based on a Jastrow-Slater wave function and variational Monte Carlo methods, but also corroborated by density-matrix renormalization group calculations. In the non-interacting limit, when varying $M>0$, a gapless point separates a trivial phase from a topological one. The inclusion of a finite Hubbard-$U$ repulsion does not give rise to any phase transition within the topological region, inducing a smooth crossover into a Haldane (spin gapped) insulator; notably, the string-order parameter, which characterizes the latter phase, is already finite in the non-interacting limit. Most importantly, at finite values of $U$, the transition between the trivial and topological states is not direct, since an emergent insulator, which shows evidence of sustaining gapless spin excitations, intrudes between them. A small $V$ interaction further stabilizes the intermediate insulator, while a sufficiently large value of this nearest-neighbor repulsion gives rise to two different charge-density wave insulators, one fully gapped and another still supporting gapless spin excitations. Our results demonstrate the richness of multi-orbital Hubbard models, in the presence of a topologically non-trivial band structure, and serve as a basis for future investigations on similar two-dimensional models.
"Insulating and metallic phases in the one-dimensional Hubbard-Su-Schrieffer-Heeger model: Insights from a backflow-inspired variational wave function"
Davide Piccioni, Francesco Ferrari, Michele Fabrizio, Federico Becca
Phys. Rev. B 111, 045125 (2025)
The interplay between electron-electron and electron-phonon interactions is studied in a one-dimensional lattice model, by means of a variational Monte Carlo method based on generalized Jastrow-Slater wave functions. Here, the fermionic part is constructed by a pair-product state, which explicitly depends on the phonon configuration, thus including the electron-phonon coupling in a backflow-inspired way. We report the results for the Hubbard model in presence of the Su-Schrieffer-Heeger coupling to optical phonons, both at half-filling and upon hole doping. At half-filling, the ground state is either a translationally invariant Mott insulator, with gapless spin excitations, or a Peierls insulator, which breaks translations and has fully gapped excitations. Away from half-filling, the charge gap closes in both Mott and Peierls insulators, turning the former into a conventional Luttinger liquid (gapless in all excitation channels). The latter, instead, retains a finite spin gap that closes only above a threshold value of the doping. Even though consistent with the general theory of interacting electrons in one dimension, the existence of such a phase (with gapless charge but gapped spin excitations) has never been demonstrated in a model with repulsive interaction and with only two Fermi points. Since the spin-gapped metal represents the one-dimensional counterpart of a superconductor, our results furnish evidence that a true off-diagonal long-range order may exist in the two-dimensional case.
"Intertwined superconductivity and orbital selectivity in a three-orbital Hubbard model for the iron pnictides"
Vito Marino, Alberto Scazzola, Federico Becca, Massimo Capone, Luca F. Tocchio
Phys. Rev. Lett. 134, 196502 (2025)
We study a three-orbital Hubbard-Kanamori model relevant for iron-based superconductors using variational wave functions explicitly including spatial correlations and electron pairing. We span the nonmagnetic sector from filling $n=4$, which is representative of undoped iron-based superconductors, to $n=3$, where a Mott insulating state with each orbital at half filling is found. In the strong-coupling regime, when the electron density is increased, we find a spontaneous differentiation between the occupation of $d_{xz}$ and $d_{yz}$ orbitals, leading to an orbital-selective state with a nematic character that becomes stronger at increasing density. One of these orbitals stays half-filled for all densities while the other one hosts (together with the $d_{xy}$ orbital) the excess of electron density. Most importantly, in this regime long-range pairing correlations appear in the orbital with the largest occupation. Our results highlight a strong link between orbital-selective correlations, nematicity, and superconductivity, which requires the presence of a significant Hund's coupling.
"Monopole excitations in the $U(1)$ Dirac spin liquid on the triangular lattice"
Sasank Budaraju, Alberto Parola, Yasir Iqbal, Federico Becca, Didier Poilblanc
Phys. Rev. B 111, 125150 (2025)
The $U(1)$ Dirac spin liquid might realize an exotic phase of matter whose low-energy properties are described by quantum electrodynamics in $2+1$ dimensions, where gapless modes exists but spinons and gauge fields are strongly coupled. Its existence has been proposed in frustrated Heisenberg models in presence of frustrating super-exchange interactions, by the (Abrikosov) fermionic representation of the spin operators [X.-G. Wen, \href{https://doi.org/10.1103/PhysRevB.65.165113}{Phys. Rev. B {\bf 65}, 165113 (2002)}], supplemented by the Gutzwiller projection. Here, we construct charge-$Q$ monopole excitations in the Heisenberg model on the triangular lattice with nearest- ($J_1$) and next-neighbor ($J_2$) couplings. In the highly frustrated regime, singlet and triplet monopoles with $Q=1$ become gapless in the thermodynamic limit; in addition, the energies for generic $Q$ agree with field-theoretical predictions, obtained for a large number of gapless fermion modes. Finally, we consider localized gauge excitations, in which magnetic $π$-fluxes are concentrated in the triangular plaquettes (in analogy with $\mathbb{Z}_2$ visons), showing that these kind of states do not play a relevant role at low energies. All our findings lend support to a stable $U(1)$ Dirac spin liquid in the $J_1-J_2$ Heisenberg model on the triangular lattice.
"Transformer Wave Function for two dimensional frustrated magnets: emergence of a Spin-Liquid Phase in the Shastry-Sutherland Model"
Luciano Loris Viteritti, Riccardo Rende, Alberto Parola, Sebastian Goldt, Federico Becca
Phys. Rev. B 111, 134411 (2025)
Understanding quantum magnetism in two-dimensional systems represents a lively branch in modern condensed-matter physics. In the presence of competing super-exchange couplings, magnetic order is frustrated and can be suppressed down to zero temperature. Still, capturing the correct nature of the exact ground state is a highly complicated task, since energy gaps in the spectrum may be very small and states with different physical properties may have competing energies. Here, we introduce a variational Ansatz for two-dimensional frustrated magnets by leveraging the power of representation learning. The key idea is to use a particular deep neural network with real-valued parameters, a so-called Transformer, to map physical spin configurations into a high-dimensional feature space. Within this abstract space, the determination of the ground-state properties is simplified and requires only a shallow output layer with complex-valued parameters. We illustrate the efficacy of this variational Ansatz by studying the ground-state phase diagram of the Shastry-Sutherland model, which captures the low-temperature behavior of SrCu$_2$(BO$_3$)$_2$ with its intriguing properties. With highly accurate numerical simulations, we provide strong evidence for the stabilization of a spin-liquid between the plaquette and antiferromagnetic phases. In addition, a direct calculation of the triplet excitation at the $Γ$ point provides compelling evidence for a gapless spin liquid. Our findings underscore the potential of Neural-Network Quantum States as a valuable tool for probing uncharted phases of matter, and open up new possibilities for establishing the properties of many-body systems.
"Foundation Neural-Networks Quantum States as a Unified Ansatz for Multiple Hamiltonians"
Riccardo Rende, Luciano Loris Viteritti, Federico Becca, Antonello Scardicchio, Alessandro Laio et al.
Nature Communications 16, 7213 (2025)
Foundation models are highly versatile neural-network architectures capable of processing different data types, such as text and images, and generalizing across various tasks like classification and generation. Inspired by this success, we propose Foundation Neural-Network Quantum States (FNQS) as an integrated paradigm for studying quantum many-body systems. FNQS leverage key principles of foundation models to define variational wave functions based on a single, versatile architecture that processes multimodal inputs, including spin configurations and Hamiltonian physical couplings. Unlike specialized architectures tailored for individual Hamiltonians, FNQS can generalize to physical Hamiltonians beyond those encountered during training, offering a unified framework adaptable to various quantum systems and tasks. FNQS enable the efficient estimation of quantities that are traditionally challenging or computationally intensive to calculate using conventional methods, particularly disorder-averaged observables. Furthermore, the fidelity susceptibility can be easily obtained to uncover quantum phase transitions without prior knowledge of order parameters. These pretrained models can be efficiently fine-tuned for specific quantum systems. The architectures trained in this paper are publicly available at https://huggingface.co/nqs-models, along with examples for implementing these neural networks in NetKet.
-
We reassess the modeling of amorphous silica bilayers as a two-dimensional classical system whose particles interact with an effective pairwise potential. We show that it is possible to reparameterize the potential developed by Roy, Heyde, and Heuer to quantitatively match the structural details of the experimental samples. We then study the glassy dynamics of the reparameterized model at low temperatures. Using appropriate cage-relative correlation functions, which suppress the effect of Mermin-Wagner fluctuations, we highlight the presence of two well-defined Arrhenius regimes separated by a narrow crossover region, which we connect to the thermodynamic anomalies and the changes in the local structure. We find that the bond-orientational order grows steadily below the crossover temperature and is associated to transient crystalline domains of nanometric size. These findings raise fundamental questions about the nature of glass structure in two dimensions and provide guidelines to interpret the experimental data.
-
We clarify the relationship between freezing, melting, and the onset of glassy dynamics in a prototypical glass-forming mixture model. Our starting point is a precise operational definition of the onset of glassiness, as expressed by the emergence of inflections in time-dependent correlation functions. By scanning the temperature-composition phase diagram of the mixture, we find a disconnect between the onset of glassiness and freezing. Surprisingly, however, the onset temperature closely tracks the melting line, along which the excess entropy is approximately constant. At fixed composition, all characteristic temperatures display nonetheless similar pressure dependencies, which are very well predicted by the isomorph theory. While our results rule out a general connection between thermodynamic metastability and glassiness, they call for a reassessment of the role of crystalline precursors in glass-forming liquids.
"Roadmap on machine learning glassy dynamics"
Gerhard Jung, Rinske M. Alkemade, Victor Bapst, Daniele Coslovich, Laura Filion et al.
Nat. Rev. Phys. 7, 91 (2025)
Unraveling the connections between microscopic structure, emergent physical properties, and slow dynamics has long been a challenge when studying the glass transition. The absence of clear visible structural order in amorphous configurations complicates the identification of the key physical mechanisms underpinning slow dynamics. The difficulty in sampling equilibrated configurations at low temperatures hampers thorough numerical and theoretical investigations. This perspective article explores the potential of machine learning (ML) techniques to face these challenges, building on the algorithms that have revolutionized computer vision and image recognition. We present recent successful ML applications, as well as many open problems for the future, such as transferability and interpretability of ML approaches. We highlight new ideas and directions in which ML could provide breakthroughs to better understand the fundamental mechanisms at play in glass-forming liquids. To foster a collaborative community effort, this article also introduces the ``GlassBench" dataset, providing simulation data and benchmarks for both two-dimensional and three-dimensional glass-formers. We propose critical metrics to compare the performance of emerging ML methodologies, in line with benchmarking practices in image and text recognition. The goal of this roadmap is to provide guidelines for the development of ML techniques in systems displaying slow dynamics, while inspiring new directions to improve our theoretical understanding of glassy liquids.
"AiiDA-TrainsPot: Towards automated training of neural-network interatomic potentials"
Davide Bidoggia, Nataliia Manko, Maria Peressi, Antimo Marrazzo
Digital Discovery (2025)
Crafting neural-network interatomic potentials (NNIPs) remains a complex task, demanding specialized expertise in both machine learning and electronic-structure calculations. Here, we introduce AiiDA-TrainsPot, an automated, open-source, and user-friendly workflow that streamlines the creation of accurate NNIPs by orchestrating density-functional-theory calculations, data augmentation strategies, and classical molecular dynamics. Our active-learning strategy leverages on-the-fly calibration of committee disagreement against ab initio reference errors to ensure reliable uncertainty estimates. We use electronic-structure descriptors and dimensionality reduction to analyze the efficiency of this calibrated criterion, and show that it minimizes both false positives and false negatives when deciding what to compute from first principles. AiiDA-TrainsPot has a modular design that supports multiple NNIP backends, enabling both the training of NNIPs from scratch and the fine-tuning of foundation models. We demonstrate its capabilities through automated training campaigns targeting pristine and defective carbon allotropes, including amorphous carbon, as well as structural phase transitions in monolayer $\mathrm{W_xMo_{1-x}Te_2}$ alloys.
