2026-03-23 | | Total: 20
We develop a one-class, deep-learning framework to detect the phase transition and recover critical behavior of the 3D Ising model. A 3D convolutional neural network autoencoder (CAE) is trained on ground-state configurations only, without prior knowledge of the critical temperature, the Hamiltonian, or the order parameter. After training, the model is applied to Monte Carlo configurations across a wide temperature range and different lattice sizes. The mean-square reconstruction error is shown to be sensitive to the transition. Finite-size scaling of the peak location for the reconstruction error susceptibility yields the critical temperature $T_c=4.5128(58)$ and the correlation-length critical exponent $ν=0.63(27)$, consistent with results from the literature. Our results show that a one-class CAE, trained on zero-temperature configurations only, can recover nontrivial critical behavior of the 3D Ising model.
Chiral active matter, which breaks both parity symmetry and time-reversal symmetry, is ubiquitous in living systems. Here, we introduce a minimal two-dimensional chiral active lattice gas by incorporating stochastic, biased local rotations. At low temperatures, the system coarsens into condensates with chiral orientations and faceted, crystal-like shapes. The interfaces align at characteristic angles with respect to the lattice axes and exhibit edge currents that are persistent, unidirectional, and angle-dependent. To generalise these findings, we propose a continuum theory by adding an active chiral edge current term to Model B, which reveals the essential role of active chiral transport in the interfacial dynamics of phase separation. Edge currents with $n$-fold symmetry produce condensates whose shapes resemble regular $n$-sided polygons. In the thin-interface limit, we construct an effective interface potential governing edge currents, from which the steady-state condensate geometry can be obtained, both in the lattice model and the continuum description.
Thermodynamic uncertainty relations (TURs) impose a universal trade-off between current precision and entropy production in autonomous steady states, constraining in particular the power, efficiency, and constancy of heat engines. We demonstrate strong violations of the long-time TUR in a minimal autonomous heat engine composed of a discrete ratchet generating work against a constant bias and an underdamped harmonic oscillator acting as an internal stochastic control. In the regime of time-scale separation, the model becomes exactly solvable and yields a closed analytical expression for the TUR ratio, where the influence of the continuous degree of freedom is fully captured by the Fano factor of oscillator zero crossings. We show that increasingly deterministic internal control drives the TUR ratio arbitrarily close to zero while the engine operates near maximal current and efficiency. In an appropriate limit, the model reduces to the classical pendulum-clock system of Pietzonka, Phys. Rev. Lett. 128, 130606 (2022).
The counterintuitive Mpemba effect, wherein a hotter system cools faster, critically lacks a universal macroscopic theory. Here, starting from linear irreversible thermodynamics, we formulate a generalized Newton's cooling law for the system-reservoir temperature difference $ΔT$, given by $\mathrm{d}ΔT/\mathrm{d}t = -[γ_0 + \mathcal{M}Q(t)][ΔT - \mathcal{I}Q(t)]$, where $γ_0$ is the bare relaxation rate, and the cumulative heat exchange $Q(t)$ explicitly encodes initial-state memory. The coefficients $\mathcal{M}$ and $\mathcal{I}$, arising from the interplay between heat flux and structural evolution, govern diverse anomalous relaxation behaviors. Specifically, $\mathcal{M} > 0$ ($\mathcal{M} < 0$) induces the (inverse) Mpemba effect, while $\mathcal{I}$ imposes a non-vanishing asymptotic $ΔT$, predicting incomplete thermalization. Our findings capture the full spectrum of memory-dependent relaxation, bridging kinetic speedup with structural freezing in complex systems.
Thermal machines are physical systems that, when fueled by input energy, perform output tasks such as heat pumping or the production of work. Their performance is characterized with several, often competing quantities, such as power, efficiency, energy waste, and resilience to environmental noise. Multi-objective optimization provides a key tool to investigate the characterization of the best thermal machines operating in the irreversible linear-response regime. Here, we derive exact analytical parameterizations for the optimal (Pareto) fronts associated with any given choice of relative weights assigned to their mean extracted power $P$, efficiency $η$, entropy production $Σ$ and the amplitude of power fluctuations $σ^2_P$. The geometry of the front of endoreversible machines is universal: two-, three-, and four-objective trade-offs follow analytical formulae that do not depend on the value of any physical parameter of the machine. We show that such universal thermodynamic Pareto fronts also set quantitative fundamental limits for the performance of non-endoreversible machines. Furthermore, we demonstrate that our results apply to existing experimental data from different physical systems also beyond the linear regime, ranging from atomic to macroscopic scales, including single-atom engines, colloidal systems, macroscopic engines and power plants.
Cellular metabolic networks exhibit scale-free topologies with power-law degree distributions across diverse organisms. Although such topologies are often linked to mutational robustness and evolutionary advantage, their role in metabolic dynamics remains unclear. Using dynamical mean-field theory, we derive an exact solution for an intracellular catalytic reaction model on dense random networks with arbitrary degree distributions. We show that the metabolic-starvation transition observed under nutrient-poor conditions for homogeneous degree distributions disappears when the out-degree distribution is scale-free. We also show that the power-law distribution of biomolecular abundances observed in real cells reflects the power-law in-degree distribution of the underlying catalytic reaction network. Large-scale numerical simulations validate these predictions. Our results provide a theoretical framework linking network topology and metabolic dynamics, and identify a dynamical advantage of scale-free topology under nutrient limitation.
A novel paradigm for sorting is introduced, based upon resetting. Using simple examples, we demonstrate that sorting is achieved by resetting the velocity component(s) or orientation of the particles, rather than position. The objects to be sorted are microparticles, modeled as suspended and spatially extended Brownian particles. This sorting-by-resetting scheme illustrates that stochastic resetting can create non-equilibrium conditions which enable tasks forbidden at thermodynamic equilibrium.
Self-interacting jump processes (SIJPs) describe systems with non-Markovian stochastic dynamics in which transition rates depend on empirical observables of the process, which gives rise to long-range memory and feedback. We derive the ``level-2.5'' large deviation (LD) principle governing the joint fluctuations of empirical occupation measure and the flux matrix for a broad class of SIJPs with general functional dependence on an empirical observable. The derivation is based on an exponential tilting construction and reveals a separation between a faster timescale of the microscopic dynamics and a slower timescale of the memory-driven evolution of transition rates, which is expressed through an exponentially discounted LD rate functional. Using this variational framework, we derive kinetic and thermodynamic uncertainty relations that extend classical Markovian bounds to non-Markovian systems, and illustrate their performance with simple examples.
We study dynamics in classical spin ice following a magnetic field quench to close to the Kasteleyn transition, using Monte Carlo simulations and dynamic scaling theory to characterize the relaxation of the magnetization and the density of magnetic monopoles. We have previously argued that this dynamics can be described in terms of seeding and growth of strings of flipped spins, and our results here demonstrate that a solvable stochastic model based on independent strings correctly describes the relaxation as well as the distribution of string lengths within the critical scaling regime near the transition. We also show how generalized scaling forms capture the behavior over a broader range of monopole densities and provide a clear understanding of the breakdown of the scaling picture further from the critical point.
We study a so-called semi-ergodic brickwork dual-unitary circuits where, in the infinite volume limit, the two-point correlation functions of single-site operators exhibit ergodic behavior along one light ray and non-ergodic behavior along the other light ray. Here, however, we study intermediate and long-time dynamics of a system in a finite, large volume. Under such dynamics, the Heisenberg evolution of a single traceless single-site operator lies within a restricted subspace, and this time evolution can be mapped to a simpler problem of a single qutrit scattering with a bunch of qubits sequentially. Despite the model being non-integrable and free from any quenched disorder, the operator entanglement grows at most logarithmic in time, contrary to prior expectations. The auto-correlation function can be written in terms of a sum of products of $SO(3)$ matrices, allowing for a random matrix prediction for the auto-correlation function at late times. The operator size distribution also becomes bimodal at certain times, displaying intermediate behavior between chaotic and free systems.
Adaptive physical and biological systems continually process fluctuating information from their environments. When the environment is nonstationary, inference itself becomes a nonequilibrium process with thermodynamic cost. We analyse a minimal stochastic model which is an overdamped particle in an adaptive double well potential whose control parameter tracks a drifting Ornstein Uhlenbeck signal. Using stochastic energetics, we derive explicit expressions for entropy production, mutual information rate, and a time dependent learning efficiency. High precision Langevin simulations reveal transient peaks in learning efficiency during rapid environmental shifts, absent in steady state averages. These results identify transient adaptive regimes as moments of maximal information to energy conversion, highlighting that maximal thermodynamic learning performance arises transiently rather than in steady state. Throughout this work, the environment is treated as an externally driven stochastic signal rather than a thermodynamic subsystem under control, and its intrinsic entropy production is therefore excluded from the thermodynamic accounting.
The phase behavior of a single type of colloid C suspended in near-critical solvents is known to be very rich. Motivated in part by recent experiments we consider a mixture of two colloidal types C1 and C2 in a binary solvent close to its demixing critical point. We extend a mean-field description of a lattice model, previously used to investigate systems with a single type of colloid in two dimensions, to the binary colloid case in three dimensions. The model treats the system as a full four-component mixture. For simplicity we choose C1 and C2 to be hard spheres with the same radius but with different affinities for one species, B, of the AB binary solvent. We show that intricate interplay between couplings of C1 and solvent, C2 and solvent as well as solvent-solvent interactions and hard sphere packing drive significant changes in the topology of the colloidal phase diagram when the relative volume fractions of the two different colloid types change. The behavior of the two lines of triple points is particularly interesting. Our results can provide some insight into the control of the self-assembly process for colloidal 'alloys' mediated by a near-critical solvent and therefore controlled by temperature in a reversible manner
We study the coupling mechanism of surface acoustic waves (SAW) with spin waves (SW) using micromagnetic analysis. The SAW magnetoacoustic excitation field is fully implemented, i.e., all strain and lattice-rotation terms are included. A realistic CoFeB film with a weak in-plane uniaxial anisotropy is considered. We investigate the conditions for efficient SAW--SW coupling, with particular emphasis on the case where the SAW propagates parallel to the external magnetic field, a configuration of special interest for magnonic applications. Remarkably, we find that the anisotropy orientation serves as a knob to tune the parallel resonant interaction. Overall, this work provides a unified and practical picture of SAW--SW coupling in thin magnetized films.
Over the last decades, two distinct approaches have been instrumental to our understanding of the computational complexity of statistical estimation. The statistical physics literature predicts algorithmic hardness through local stability and monotonicity properties of the Franz--Parisi (FP) potential \cite{franz1995recipes,franz1997phase}, while the mathematically rigorous literature characterizes hardness via the limitations of restricted algorithmic classes, most notably low-degree polynomial estimators \cite{hopkins2017efficient}. For many inference models, these two perspectives yield strikingly consistent predictions, giving rise to a long-standing open problem of establishing a precise mathematical relationship between them. In this work, we show that for estimation problems the power of low-degree polynomials is equivalent to the monotonicity of the annealed FP potential for a broad family of Gaussian additive models (GAMs) with signal-to-noise ratio $λ$. In particular, subject to a low-degree conjecture for GAMs, our results imply that the polynomial-time limits of these models are directly implied by the monotonicity of the annealed FP potential, in conceptual agreement with predictions from the physics literature dating back to the 1990s.
The proton-disordered molecular phase of water ice (ice-VII) and its ultrahigh-pressure non-molecular phase (ice-X) share identical macroscopic crystal symmetry (space group $Pn\bar{3}m$). This raises a fundamental thermodynamic question: are they distinct phases separated by a singularity, or are they adiabatically connected via a continuous crossover? To resolve this paradox, we investigate the finite-temperature phase diagram of high-pressure ices VII and X, as well as VIII, the proton-ordered phase that emerges at lower temperatures, using an effective classical spin-$1$ Blume-Capel model on the pyrochlore lattice. Through Monte Carlo simulations, we demonstrate that within this model, the transformation between the states corresponding to ice-VII and ice-X lacks a thermodynamic singularity, as characterized by non-divergent and non-coinciding peaks in the specific heat and susceptibility associated with the $S^z=0$ occupation. We attribute this continuous crossover behavior to the topological fragility of the hydrogen-bond network: the thermal proliferation of point-like monopole excitations (violations of the ice rules) induces Debye-Hückel screening of the emergent gauge field, destroying the topological Coulomb phase at any finite temperature. In contrast, the destruction of the proton-ordered ice-VIII phase involves spontaneous symmetry breaking and remains a first-order phase transition. Our findings provide a microscopic rationale that reconciles the macroscopic crystallographic symmetries of dense ice with its underlying topological properties.
We investigate the rheology of a sheared granular gas composed of hard spheres with a velocity-dependent restitution coefficient. Using kinetic theory, we derive the shear viscosity and show that it exhibits an S-shaped dependence on the shear rate when the restitution coefficient switches between two values depending on the collision velocity. As a result, a discontinuous change of viscosity emerges between low- and high-shear regimes, both characterized by Bagnold-type scaling. While the phenomenology resembles the Wyart-Cates scenario for dense suspensions, the present transition arises purely from kinetic effects without frictional contacts or jamming.
We propose and test logarithmic Krylov (logK) complexity, an operator growth measure akin to Krylov complexity defined through a replica approach, as a viable probe of early-time operator scrambling without false positives. In finite-dimensional quantum systems, such as the Lipkin--Meshkov--Glick (LMG) model and the mixed-field Ising model at the chaotic point, we provide numerical evidence that logK-complexity discriminates between genuine and saddle-dominated scrambling at early times, correctly avoiding the exponential contribution coming from the unstable saddle in the former case, and closely tracking the conventional Krylov complexity in the latter. In integrable quantum systems admitting infinite-dimensional Krylov subspaces, such as the SYK$_{2}$ model and the quantum inverted harmonic oscillator, we show that by modifying the Krylov spreading operator, obtained through generalizing the analytic continuation procedure in the replica trick, the logK complexity can be refined to capture the integrable properties of the theories. We supplement these analyses by extending the Krylov formalism in classical dynamical systems and defining classical versions of these operator growth measures, showing that the false positives arising from unstable saddles in classical phase space are non-existent.
Motivated by the interplay between 2D and 3D scaling signatures observed in unconventional layered superconductors, we present a systematic Monte Carlo study of the three-dimensional classical XY model with anisotropic in-plane $J_\parallel$ and inter-plane $J_\perp$ couplings. Our study includes very small values of the system anisotropy $Δ=J_\perp /J_\parallel$ not studied before, and focuses on characterizing the crossover from quasi-2D topological scaling to genuine 3D critical behavior. The numerical results for the critical temperature unambiguously reveal a logarithmic scaling with $Δ$, directly related to the topological scaling in the 2D limit. Despite the 3D nature of the layered XY criticality, topological scaling signatures survive up to system sizes comparable to the crossover length $\ell_J$, which diverges at small $Δ$ with a scaling behavior reminiscent of the Berezinskii-Kosterlitz-Thouless (BKT) transition. This shows that genuine 3D symmetry-breaking behavior emerges only at exceedingly large system sizes when the anisotropy is very strong. Our results indicate that new experimental evidence is required to clarify the extent to which the critical signatures observed in layered strongly correlated materials are shaped by their pronounced anisotropy.
We study the symmetry of the coupling between surface acoustic waves and ferromagnetic resonance in a thin magnetic film of CoFeB deposited on top of a piezoelectric Z-cut LiNbO3 substrate. We vary the orientation of the applied magnetic field with respect to the wavevector of the surface acoustic wave. Experiments indicate an unexpected 2-fold symmetry of the absorption of the SAW energy by the magnetic film. We discuss whether this symmetry can arise from the magnetoelastic torque of the longitudinal strain and the magnetic susceptibility of ferromagnetic resonance. We find that one origin of the 2-fold symmetry can be the weak in-plane uniaxial anisotropy present within the magnetic film. This phenomena adds to the previously identified other source of 2-fold symmetry but shall persist for ultrathin films when the dipolar interactions cease to contribute to the anisotropy of the slope of the spin wave dispersion relation.
We study the frequency non-reciprocity of the spin waves in symmetric CoFeB/Ru/CoFeB synthetic antiferromagnets stacks set in the scissors state by in-plane applied fields. Using a combination of Brillouin Light Scattering and propagating spin wave spectroscopy experiments, we show that the acoustical spin waves in synthetic antiferromagnets possess a unique feature if their wavevector is parallel to the applied field: the frequency non-reciprocity due to layer-to-layer dipolar interactions can be so large that the acoustical spin waves transfer energy in a unidirectional manner for a wide and bipolar interval of wavevectors. Analytical modeling and full micromagnetic calculations are conducted to account for the dispersion relations of the optical and acoustical spin waves for arbitrary field orientations. Our formalism provides a simple and direct method to understand and design devices harnessing unidirectional propagation of spin waves in synthetic antiferromagnets.