Papers

Despite the exhaustive understanding gathered around non-interacting topological states of matter, there is no single method capable of systematically delivering simple, numerically efficient topological invariants that is applicable to all crystalline and non-crystalline systems alike. Here we revisit the spectral localizer operator, constructed from the Hamiltonian and position operators, and show how it can be treated it as an auxiliary zero-dimensional Hamiltonian whose topology encodes the higher dimensional phases of the parent Hamiltonian. Its classification reduces every topological invariant to a matrix signature or the sign of a Pfaffian for an appropriate localizer, both of which are simple to interpret and efficient to compute in real space. We validate this approach by deriving simple real-space invariants for weak and rotationally invariant crystalline phases that were previously beyond the grasp of the spectral localizer formalism, atomic limits that escape scattering invariants, phases that evade symmetry-based indicator methods, as well as phases that had no previously known invariant. Our work provides a systematic way to construct any non-interacting topological invariant for a crystalline or non-crystalline systems, opening avenues to classify and predict the topology of previously unexplored classes of materials.

We study the high-dimensional asymptotics of Harish-Chandra-Itzykson-Zuber (HCIZ) integrals in the extensive-rank regime. The limit of these integrals is governed by a one-dimensional boundary-value hydrodynamical problem originally derived by Matytsin (1994) and rigorously proved by Guionnet and Zeitouni (2002). Despite its wide-ranging applications, explicit solutions to this problem are known only in a few specific cases. In this work, we introduce an efficient numerical scheme based on a particle discretization and prove its convergence to the continuous boundary-value problem for generic boundary densities. We validate our approach against known analytical solutions and apply it to generic densities, uncovering interesting dynamical phenomena. The high-dimensional limit of HCIZ integrals appears in various contexts, from the large deviations of random matrix spectra to the limiting free energy of disordered systems, high-dimensional statistics, and machine learning. As such, our contribution opens the way towards the numerical exploration of a wide range of high-dimensional models that were previously intractable.

Modern Hopfield networks provide a unifying framework for associative memory, transformer attention, diffusion-based generative models, and biological attractor dynamics, linking these systems through a common energy-based dynamics in which states are updated toward weighted combinations of stored patterns. Across these settings, network dynamics is determined by the organization of the energy landscape and the bifurcations of its fixed points. Despite their central role, a general theory of these bifurcations has remained unavailable beyond specific architectures and idealized pattern ensembles. Here we derive stability and bifurcation criteria for the fixed points of general convex-dual Modern Hopfield networks for a general statistics of the stored patterns. Applying this framework to random, block-correlated, and infinitely hierarchical pattern ensembles, we show how memory correlations systematically organize the emergence of hierarchical attractors through successive bifurcations. We further demonstrate that these predictions quantitatively describe retrieval bifurcations in MHNs storing patterns sampled from MNIST, and recapitulate the hierarchical organization of hematopoietic cell identities. Our results establish a general bifurcation theory for Modern Hopfield networks and identify the organization of fixed points as a unifying principle underlying their computational and biological behavior.

Foundation models for ground states in spin-1/2 systems are a promising method for problems ranging from quantum chemistry to identifying new phase diagrams. Nearly all such models are currently pure-states that condition on the Hamiltonian's parameters, whose Monte Carlo samples give energy estimates according to the variational principle. In this contribution, we show that this representation is topologically obstructed. For any gapped Hamiltonian family whose ground-state bundle is non-trivial, every continuous normalized state-vector model has zero fidelity with the ground state at some parameter value in the Hamiltonian family. For that value, the energy is at least one spectral gap, $\Delta$, with an $O(\Delta)$ gap in an open-neighbourhood of that point. We show that this is a sufficient no-go also in the case of degenerate ground-state manifolds, time dynamics, and periodic systems with mixed space-time topology, demonstrating these obstructions on one- and two-qubit systems. We discuss how this causes a spike in the fidelity susceptibility, giving a numerical signature of a phase-transition where there is none. We then show that operator-valued models canonically avoid these obstructions and preserve topological information, implying a structural necessity in representation for foundation neural quantum states.

We study the use of graph neural networks (GNNs) for finding approximate ground states of Ising models. Efficiently finding these ground states is of broad significance because many combinatorial optimization problems can be formulated as an Ising model with the appropriate choice of couplings and fields. Exactly solving these problems is hard but there are many good heuristic methods. A lineage of these heuristics build from mean-field approximations: one approach uses the leading eigenvector of an appropriately defined matrix, another is the min-sum algorithm, also known as the energetic cavity method. Without modification, GNNs perform worse than both of these methods. We consider small modifications to the GNN to incorporate these heuristics and find that this considerably improves performance. While the modified approach is competitive against other deep-learning approaches, we still find that simulated annealing is reliably at least as good as deep learning methods for the same computational cost.

We develop a dynamical mean-field theory for nonreciprocal associative networks that store an extensive number of dynamical attractors, from limit cycles to strange attractors. Using a path integral calculation under quenched disorder, we derive self-consistent dynamical mean-field equations for pattern overlaps, autocorrelations and response functions. Memory retrieval capacity is governed by the spectral structure of the coupling matrices encoding stored patterns. When their eigenvalues are coherently aligned, retarded self-interactions and quenched noise feed back destructively: at zero eigenphase (fixed point attractors) the classical equilibrium capacity bound is recovered, while for limit cycles retrieval collapses far below it. In contrast, for uniformly distributed eigenphases, retarded self-interactions and much of the quenched noise cancels, reducing the dynamics to an effective single-spin process and amplifying capacity substantially. We validate the theory against microscopic simulations for limit-cycle and chaotic attractors, identifying eigenvalue decoherence as the mechanism enabling enhanced storage of dynamical memories.

Critical universality does not determine the microscopic conditional structure of a probability measure. We study a bicolored cluster recursion constrained to remain critical at every generation, with no equilibrium spin measure or fixed coupling imposed. In both two and three dimensions, the resulting history-dependent sequence develops a common Ising/Fortuin–Kasteleyn (FK) compatibility structure: the second-shell dependence of a one-site conditional law is strongly suppressed, nearest-neighbor effective couplings move progressively toward one another near the critical Ising value, and cluster and interface observables organize around the corresponding FK geometry. An exact cluster-coloring factorization singles out $q=2$ as the point where the residual connectivity weight disappears from the two-color spin marginal. Thus equilibrium-compatible conditional structure can emerge along a trajectory that remains critical throughout.

We study finite-rank multiplicative deformations of rotationally invariant non-Hermitian random matrices. More precisely, we consider models of the form $\mathbf{A}(\mathbf{I}+\mathbf{T})$, where $\mathbf{A}$ is a large rotationally invariant non-Hermitian random matrix, $\mathbf{T}$ is a finite-rank normal perturbation, and $\mathbf{I}$ denotes the identity matrix. We characterize the emergence of outlier eigenvalues, their fluctuations, and the associated eigenvector overlaps. Our results provide a multiplicative non-Hermitian counterpart to the classical Baik–Ben Arous–Péché framework.

The triangular Ising antiferromagnet, with its residual entropy density and critical correlations at absolute zero, is the archetype of a two-dimensional spin liquid. Its ground state is partitioned into topological sectors connected by global spin flip events that wrap the lattice boundaries, and become statistically irrelevant in the thermodynamic limit where magnetic fluctuations are restricted to the dominant sector. Thus far, these properties have been mainly investigated from a theoretical perspective, and one might wonder to what extent they can be transposed to real materials. Here, we present experimental observations obtained in an artificial, macroscopic realisation of the seminal triangular Ising antiferromagnet that consists of a lattice of millimeter-sized NdFeB cylinders put into motion by a mechanical shaker. Specifically, we demonstrate that the very-low-energy physics and the true ground state of this model can be reached experimentally. Besides, we are able to probe, in space and time, the magnetic properties within manually preselected pockets of the ground-state manifold that emulate, to a good approximation, the behaviour in distinct topological sectors. Our approach opens new avenues for naked-eye visualisation and hand manipulation of many-body phenomena associated to frustrated magnetism and models of statistical physics.

Frictional relaxation after a sudden velocity change underlies the phenomenological laws that describe the stability of steady sliding and of seismic faults. To connect this relaxation with measurable statistics of microscopic contacts, we introduce a minimal model incorporating logarithmic aging of contact forces and a power-law contact-size distribution reflecting fractal interface roughness, and derive the relaxation function analytically. Depending on the power-law exponent $\alpha$, the intermediate-time response exhibits a plateau, logarithmic decay, or power-law decay, in contrast to the exponential decay assumed in the phenomenological laws. The theory provides experimentally testable links between microscopic contact statistics and macroscopic frictional relaxation.

Heterogeneous operators often organize response through hidden barriers, wells, and weakly communicating compartments that are not faithfully described by Euclidean distance or raw graph connectivity alone. We study this structure through the localization landscape, obtained from a single source solve, and the associated effective potential $W_{\mathrm{eff}} = 1/u$. Starting from both discrete Schrödinger-type operators and heterogeneous reaction-diffusion operators, we show how $W_{\mathrm{eff}}$ induces effective wells, basin partitions, and Agmon-type neighborhoods that serve as an operator-aware geometry for localization and transport. This geometry explains the spatial support of low-energy eigenmodes, produces adaptive receptive fields on a fixed graph, sharpens locality in attention-like interactions, and yields bottleneck-sensitive modulation of diffusion. We also include a finite-element reaction-diffusion example with strongly compartmentalized coefficients, demonstrating that the same framework extends beyond simple grid operators to heterogeneous continuum systems. The results support a precise conclusion: localization-landscape geometry is a strong inductive bias when relevance is controlled by confinement, barriers, and low-energy accessibility, but its benefits are task-dependent rather than universal.

Fast ionic transport is a defining feature of many solid electrolytes, yet its microscopic origin is not fully understood. In the absence of microscopic insights, the development of next-generation solid-state batteries remains largely empirical. Most existing measurements access either the low-frequency transport response or the high-frequency bound polarization, yet the intermediate mesoscopic frequency regime is where ionic transport emerges. By varying the $\mathrm{AgI}$ concentration ($x$) and performing time-domain terahertz spectroscopy (TDTS) in a prototypical glassy electrolyte $\left(\mathrm{AgI}\right)_{x}\left(\mathrm{AgPO_3}\right)_{(1-x)}$, we reveal this intermediate frequency regime and identify a crossover from bound-current-dominated conduction to conductivity arising from short-range dispersive ionic transport. We find that bound polarization associated with the bond-bending motion of the $\mathrm{P{-}O^- -Ag^+}$ motif is present across compositions but is insufficient to produce ionic transport on its own. Transport emerges only when this polarization is embedded in a sufficiently soft $\mathrm{AgPO_3}$ glassy matrix and accompanied by a high carrier density. These ingredients together take the system from a vibrationally bound response to short-range dispersive motion.

Metal-insulator transitions in correlated materials are often accompanied by a change of crystal structure. They are commonly described within a coherent lattice approximation, which combines a correlated treatment of the electrons with a lattice represented by one or a few classical distortion coordinates fixed by minimizing an energy. The structural degrees of freedom then carry no entropy of their own, even though the partition of the transition entropy between electrons and lattice is often what decides the transition temperature. Here we develop a stochastic semiclassical extension of dynamical mean-field theory in which correlated electrons and fluctuating lattice distortions are evolved together, extending previous formulations from linear to nonlinear electron-phonon interactions and from a single to several coupled lattice modes. The resulting non-conservative Langevin equations have deterministic forces, damping, and correlated noise generated self-consistently by the interacting electronic subsystem. Applying the approach to a minimal two-orbital model containing the two symmetry-distinct distortions of the monoclinic phase of vanadium dioxide, we find that the two distortions melt at well separated temperatures, giving an insulating, an intermediate metallic, and a high-symmetry phase. The two transitions have different origins: the melting of the dimerization is driven by the coupling to the correlated electrons and is captured already within the coherent lattice approximation, whereas the restoration of the undistorted structure requires the entropy of the fluctuating lattice. Since the electronic subsystem is treated within nonequilibrium dynamical mean-field theory, the framework can be extended in future work to photoexcited systems with nonthermal electronic distributions.

Neural network quantum states (NQS) for representing spin-$\frac{1}{2}$ systems are typically built from Multi-Layer Perceptrons (MLPs) with binary visible variables, in which the input spins are first combined through linear affine maps before more expressive nonlinear transformations are applied. The standard way for representing multi-categorical systems, namely spin-1 and spin-2, or the q-state quantum Potts systems with more than two local degrees of freedom, is a unary or one-hot encoded MLP architecture. Following a few similar studies, I observe that while the one-hot construction is the mathematically faithful encoding for systems with more than two local states such as spin-$S$ models with $S > 1/2$ or $q$-state Potts models, its parameter count grows with the number of local states, and the resulting optimization landscape can hinder convergence. I find numerically that allowing $\log \Psi$ to be a nonlinear function of the raw multi-valued spin variable is a natural categorical generalization: it preserves the labelling freedom of the local basis and reproduces the one-hot model with strictly fewer parameters, often with improved trainability. I first benchmark this idea on shallow Restricted Boltzmann Machines (RBMs), equipping them with several distinct nonlinear connections, for spin-1, 2, and 3 Heisenberg chains. I then turn to Kolmogorov-Arnold Networks (KANs), where each edge carries a learnable univariate nonlinearity, and show that they provide a strictly more expressive realization of the same principle. Finally, I demonstrate that this framework captures the critical behaviour of the quantum Potts Hamiltonian, recovering its phase transition.

Quantum entanglement growth is commonly summarized by a single entropy, obscuring where correlations reside in the exponentially large Schmidt spectrum and how they form. Here, we introduce Schmidt-scale concentration and dominant Schmidt scale, two coordinates that locate the probability maximum across logarithmic windows in ordered Schmidt-rank space. Applied to quenches of a random-field spin chain, these coordinates distinguish rapid transport of the dominant scale to higher Schmidt rank at weak disorder from strongly suppressed transport despite continued logarithmic entropy growth at strong disorder. The disorder-averaged dynamics exhibit an ordered hierarchy: entropy production peaks first, spectral roughness and exact nonlocal magic peak next, and dominant-Schmidt-scale transport becomes typical only after a substantial delay. Moreover, a solvable head–tail model and controlled numerical experiments reveal the physical origin of this hierarchy: the spectral path determines the order of events, local dynamics on active exchange bonds set their early timing, and intra-subsystem many-body dressing further delays dominant-Schmidt-scale transport. These results establish the Schmidt-scale coordinates as powerful dynamical probes for uncovering fine-grained entanglement structures distinguishing entanglement production, entanglement-spectrum reorganization, and dominant-Schmidt-scale transport beyond entropy alone.

Physical reservoir computing (PRC) with neuromorphic networks offers a promising approach to brain-inspired information processing, exploiting emergent nonlinear dynamics of physical neural networks as a computational resource. This study demonstrates fully autonomous closed-loop prediction of the Mackey–Glass (MG) chaotic time series using a simulated neuromorphic nanowire network as the physical reservoir. Two strategies are evaluated: the virtual node (VN) method, which expands the feature space by temporal multiplexing of reservoir states, and a non-VN approach that uses all physical node readouts directly without temporal multiplexing. Results are reported for two values of the MG time delay parameter, $\tau = 18$ and $\tau = 21$, the latter representing a more complex chaotic regime not previously evaluated for this class of physical reservoir. Over a short prediction horizon of $T = 100$ timesteps, the VN approach achieves autonomous prediction accuracies of $90.4$% and $89.7$% at $\tau = 18$ and $\tau = 21$, respectively, while the non-VN approach achieves $81.5$% and $76.2$%. Long-horizon analysis over $T = 500$ timesteps shows that both approaches reproduce the qualitative attractor structure and dominant spectral content of the true MG signal, with trajectories remaining bounded throughout. These results suggest that the intrinsic dynamics of neuromorphic nanowire networks are sufficient to support meaningful autonomous chaotic time series prediction without virtual node augmentation, and that performance may improve further as physical network sizes scale to the millions of nodes achievable in hardware. As this study uses simulated networks, extrapolation to physically fabricated large-scale arrays remains to be validated experimentally.

We numerically implement full quantum circuits of optimal tapered Quantum Phase Estimation (in the form of an implementable approximate bandwidth-limited DPSS taper) for the purpose of minimum eigenvalue computation, where the Hamiltonian time-evolution unitary is implemented using standard Trotterization, and the taper state is constructed using an optimized bandwidth-limited taper. We illustrate, on representative quantum Hamiltonians, with tractable exact classical numerical quantum circuit simulations, how tapered QPE performs with respect to optimal phase sampling rate, absolute error, Shannon entropy of the optimal phase distribution, when the tQPE algorithm parameters are changed. Those algorithm parameters are Trotter error, finite sampling, imperfect initial states, total evolution time, and digitization error (number of phase qubits). The reported numerical experiments include up to $\ell=10$ phase qubits of precision with $m=3$ additional phase-register error suppression qubits, on a $4$-qubit Heisenberg quantum magnet model Hamiltonian, using up to 12th-order Trotterization. We numerically show i) the steady-state optimal phase sampling rate is determined by the initial state overlap with the ground-state, ii) with respect to phase sampling probability, or overall eigenvalue sampling error rate, there is no strong evolution time dependence for tQPE unless there is high Trotter error, iii) the approximated time-evolution unitaries in tQPE, like in standard QPE, can result in substantial ringing, which leads to non-physical eigenvalue estimates.

Nanothermodynamics describes the process where large systems subdivide into equilibrium distributions of small subsystems. A key ingredient is Hill's subdivision potential (E) that ensures adherence to the 1st and 2nd laws of thermodynamics in systems of any size. In this review and reassessment, it is emphasized that nanothermodynamics gives new insight into many measurements, theories, and simulations. Measurements establishing the need for nanothermodynamics show thermodynamic heterogeneity from multiple effective temperatures (T_i) inside most types of materials. One theoretical result that requires E=0 is the stable solution of Ising's original model for finite chains of interacting spins, a solution Ising could not have found 40 years before Hill's work. Another result is a novel solution to Gibbs' paradox that makes the entropy of the semiclassical ideal gas exactly extensive. Molecular dynamics simulations reveal how a standard fluctuation relation is modified when local degrees of freedom fluctuate faster than their coupling to the heat bath, consistent with the measured thermodynamic heterogeneity. Simulations of a Creutz-like model, comprised of Ising spins coupled to an explicit heat bath of Einstein oscillators, are used to study the 2nd law. It is found that maximizing the total entropy (S_t) requires an intrinsically irreversible step, providing a counterexample to the usual claim that statistical mechanics emerges from reversible dynamics. Furthermore, fluctuations of this model are best described by Einstein's reversal of Boltzmann's relation and the 2nd-law, not by recent fluctuation theorems.

Metastability in open quantum systems is usually inferred from spectral separation in the Liouvillian, which governs unconditional, ensemble-averaged dynamics. We show that this diagnosis is incomplete at the level of individual quantum trajectories: conditioned realizations of the same unconditional dynamics can bypass, transiently access, or operationally preserve a metastable memory, depending on the monitored channel and the observed record. We demonstrate these mechanisms in a driven-dissipative nonlinear oscillator realizing a quantum associative memory, by comparing spectra, target fidelities, and phase-space distributions. The non-Hermitian dynamics obtained by post-selecting on the absence of detected events supports metastable retrieval, but with a distinct long-time fate: normalization selects the least-decaying mode of the non-Hermitian spectrum rather than the addressed memory branch. In contrast, stochastic jump trajectories can preserve retrieval over extended times when the post-measurement update remains compatible with the coherent memory structure. Thus trajectory-level metastability is not determined by the averaged generator alone; it requires compatibility between the monitored channel, the measurement record, and the metastable manifold.

Diffusion sampling can be viewed as imaginary-time annealing of probability densities. From a forward/backward Euclidean Schrödinger pair, we show that fixing the noising drift forces the reverse score term, and that the same logarithmic force is the one-sided imaginary-time counterdiabatic connection of a supersymmetric Score Hamiltonian. The correspondence turns score-learning error into a Hamiltonian perturbation and yields a schedule principle: if the first $r$ terminal derivatives vanish, Morita–Nishimori boundary cancellation suppresses the residual Hellinger error from $T^{-2}$ to $T^{-2r-2}$ until score or sampling floors are reached. Reverse-SDE benchmarks, including two-dimensional learned-score tests, confirm the predicted improvement.

Simple synthesis of graphitic compounds, their high conductivity, and integrability with other materials motivate the effort toward graphite-based thermoelectric generators. At the same time, semimetallic nature of graphite and graphene results in nearly-zero thermopower due to electron-hole compensation. Here, we observe large transverse thermopower in chemically derived multilayer graphene films with strong fluctuations of thickness and carrier density at low magnetic fields $B$. Using the scanning laser-induced heating of macroscopic film, we find that transverse (Nernst) thermoelectric voltage becomes comparable to the longitudinal thermoelectric voltage at the metal-doped graphene contact at $B^*\approx4$ mT and ambient conditions. Estimates of transverse thermopower $S_{xy}$ based on the known laser-induced temperature show that it is as large as $\sim 10$ $\mu$V/K at $B^*$, and raises in a sub-linear fashion to 250 $\mu$V/K at $B\approx315$ mT, the maximum field we reach with centimeter-sized permanent magnet. Extra increase in thermoelectric signal is achieved upon voltage measurement at Hall probes when the dc field lines are co-directional with local Nernst current. Our results show the promise of large-scale multilayer graphene for thermoelectricity generation.

In situ adjoint training extracts parameter gradients directly from measurement, but has so far been limited to reciprocal or restricted systems. Here, we introduce the physical counterpart of the policy gradient theorem: a stochastic-adjoint gradient estimator that lifts these constraints by trading reciprocity for nondegenerate diffusion. As validation, we train a nonlinear resonator network, whose own dynamics supply the policy, against antagonistic temporal modulations with gradients from measured stochastic trajectories alone, without finite differences or a separate adjoint experiment.

Due to their intrinsic interplay between long-range order and quasi-disorder, quasicrystalline systems provide a rich platform for investigating novel quantum phenomena. Here we study the non-equilibrium dynamics following quantum quenches across the superfluid to Bose-glass transition using ultracold bosons in an optical quasicrystal. Fast quenches into the Bose glass regime induce a quantum walk in momentum space that spreads over increasingly higher momentum orders. Conversely, quenches into the superfluid regime initiate a real-space quantum walk of initially localized atoms, forming a light-cone-like structure bounded by Lieb-Robinson limits. Characteristic timescales reveal strong links to the underlying Hamiltonian and are governed primarily by quasi-disorder strength in the Bose glass phase, and by tunneling strength in the superfluid phase. Finally, we analyze slower quenches into deeper lattices and observe the decay of coherence across the phase transition.

Establishing a robust and physically interpretable link between static structure and heterogeneous relaxation dynamics remains a fundamental challenge in glass physics. Here, we introduce a weighted pair-entropy descriptor based on the conventional two-body excess entropy. For this, we multiply the integrand used to calculate the excess entropy by a weight function that is directly related to the length scale of the pair correlation function. This multiplication does not affect the contribution of the short range order to the local excess entropy, but allows to take into account the structure present on intermediate distances, i.e., the medium-range order. For a canonical two-dimensional Lennard-Jones glass former, the resulting descriptor exhibits a strong correlation with particle-level dynamical propensity at long times (multiples of the alpha-relaxation time), with a maximum structure–dynamics correlation reaching about 0.9, substantially outperforming the predictive power of the conventional local pair excess entropy. These results demonstrate that incorporating a physically motivated structural length scale into entropy-based descriptors markedly enhances their predictive power while preserving physical interpretability. Our findings provide a simple and general framework for investigating structure–dynamics correlations in glass-forming systems.

We investigate the slowing down of dynamics in a glass-forming mixture interacting via an inverse-power-law (IPL) potential using a combination of theory and large-scale molecular dynamics simulations. We measure the static pair-correlation function, configurational entropy, inherent-structure energy, and structural relaxation time. We employ a theoretical framework to calculate the structural relaxation time $\tau_{\alpha}$, which is found to be in very good agreement with the simulation results. The theory identifies a local structural order which defines the cooperativity of the relaxation and brings forth a fluctuation induced parameter $\psi ( T )$ and a crossover temperature $T_a$ that characterize the density and temperature dependence of the glassy dynamics. Furthermore, we determine a crossover temperature using independent dynamical and thermodynamic criteria and compare with the theoretically predicted crossover temperature $T_a$. Relaxation dynamics is shown to obey density-temperature scaling, similar to thermodynamic properties, in terms of a variable $\Gamma$ formed by an appropriate combination of density and temperature, characteristic of IPL interactions. Finally, we show that, when the excess thermodynamic and dynamic quantities obtained at different densities are plotted as functions of the reduced temperature $T/T_a$ (or $T_a/T$), the data collapse onto master curves with excellent agreement between theory and simulation. These scaling relations provide a unified description of the thermodynamics and dynamics in IPL systems, enabling the prediction of relaxation behavior over a wide range of densities from data at a single state point.

Overparameterized neural networks carry far more hidden units than a task nominally requires, raising the question of which neurons are essential and whether that distinction is legible in the representation itself, without labels or gradients. We cast neuron selection as the problem of coarse-graining the hidden layer by retaining a subset of its neurons, and score each putative selection by the mapping entropy (ME). This quantity measures the loss of discriminatory power inherent in discarding part of the network neurons, and the selection that minimises the ME is taken as particularly informative. This criterion is fully unsupervised, in that it depends only on hidden-activation statistics. In teacher-student networks, ME optimisation recovers the minimal teacher-consistent representation and retains extra units in proportion to the hidden layer's residual variability; in a non-linear Gaussian process task, it selects coherent functional-class mappings whose preferred class shifts across training. On this task and on translation-augmented MNIST, ME-selected subnetworks outperform random subsets of equal size, most clearly under strong compression - linking configurational distinguishability to predictive performance.

We study the stability of the solution corresponding to the polar phase of superfluid 3He in nematic aerogels with respect to perturbations induced by magnetic impurity scattering. We show that when the perturbation simultaneously breaks axial and time-reversal symmetry, the polar-distorted A phase is realized with the highest transition temperature. Within the proposed model of a p-wave scattering potential, which accounts for the effective spin-orbit interaction between the orbital angular momentum of a scattering quasiparticle and the magnetic moment texture of the scatterer, we solve the eigenvalue and eigenvector problem for the Gor'kov self-consistency equation. Finally, we discuss how our results relate to the available experimental data.

We discuss the possibility to train a thermodynamic system, whose micro-variables are fully determined through the Hamiltonian by the usual statistical mechanical rules, to perform tasks such as memorization and generalization. Training is achieved by the application of suitable external fields, playing the role of data. At variance with conventional machine learning, no other logical or algorithmic rules are introduced. The system is amenable, in principle, to exact analytical calculations. We specialize this general approach to a prototypical Ising system with annealed dichotomous couplings and study its learning ability. Results indicate excellent memorization and good generalization capacity already for small systems, and a tendency to improve with system size.

We investigate the interplay between geometric dilution and non-Hermitian disorder in the two-dimensional quantum site-percolation model. Non-Hermiticity is introduced through random imaginary on-site potentials, representing spatially uncorrelated gain and loss, while the hopping amplitudes remain reciprocal. By combining complex level-spacing statistics, participation entropy, and multifractal analysis, we characterize the localization properties of the eigenstates as functions of the disorder and the non-Hermiticity strength. Our finite-size scaling results show that non-Hermitian disorder shifts the quantum percolation threshold ($p_q$) toward larger occupation probabilities. Consequently, the fully delocalized phase is progressively suppressed and disappears at sufficiently strong disorder. This suppression is not a simple consequence of adding on-site disorder of a given strength, but is specifically enhanced by its imaginary character, as an equally strong real (Hermitian) on-site potential produces a weaker shift of $p_q$. Nevertheless, the intermediate region between the classical ($p_c$) and quantum percolation thresholds presents a genuine multifractal critical phase, while the localization-length exponent $\nu$ remains the same, relative to its Hermitian value. Altogether, our results demonstrate that random gain and loss enhance the multifractal regime while preserving the universality class of the quantum percolation transition.

A passive, reciprocal, linear network driven by forces or currents and read at points it drives cannot be trained below an error floor computed from the target beforehand. The usual parameter count for such networks does not charge for Maxwell-Betti reciprocity. We prove that when such an elastic, resistor or flow network has a symmetric response operator and p driven degrees of freedom are also read, its reachable response blocks lie in a subspace of codimension p(p-1)/2, whatever its size and topology, and the floor is the distance to it. On targets with a reachable symmetric part, a second-order optimiser ends within 0.1% of the floor in 142 of 144 simulated runs, and a contrastive rule with a bond-local update direction, within its step budget, in 22 of 24. Reciprocity charges a list of single drive-read tasks only for the pairs it instruments in both directions, and a rank test on the passive network says in advance whether, and on which bonds, odd couplings locally restore the lost directions, at a cost in non-reciprocity that the target bounds from below. Under an imposed-displacement drive, the one most physical-learning hardware uses, reciprocity survives as a weaker law, an exact balance between forward and reverse transmissions once each is weighted by the driving-point compliance at its input, and an inequality on their product, under which no symmetric positive-definite chain meets both targets of a published robotic metamaterial. All nineteen published layouts we tabulated pay nothing; a force-driven layout that asks one pair to respond differently in its two directions will.