Papers

Neural networks trained by gradient descent on a smooth cost function can nevertheless learn in steps: the cost holds on long plateaus and then drops abruptly. Meanwhile, training losses instead follow smooth power laws. Variants of both behaviors occur in architectures with very different microscopic structures, which is the signature of a few relevant collective variables. We show that a symmetry fixes what those variables are: a network layer is a sum over interchangeable units, so relabeling the units leaves it unchanged; given smoothness and the condition that a unit's gradient vanish at the origin, symmetry then enforces a universal leading form for the expansion about the near-zero weights present at the start of training, the quadratic $\Tr[WW^{\top}A(x)]$, in which every architectural detail is confined to a single “structure matrix" $A(x)$ that we compute for each architecture. Perceptrons, attention layers, mixtures of experts, and convolutions become one model at different $A$. Its training dynamics then close on the “order parameter" $M=WW^{\top}$ and, whenever the data matrices share an eigenbasis, reduce to a Lotka–Volterra equation whose modes switch on one after another. The smaller the initial weights, the further apart the switch-on times, and the plateaus appear as a singular limit of a smooth flow; when many modes are unresolved the same events merge into a power law in training time whose exponent the theory predicts. We confirm both numerically across training methods and architectures.

In theoretical physics, finding connections between problems that appear in distinct contexts is an important way to leapfrog progresses, often by illuminating deep aspects that may otherwise seem obscure. In this paper, we consider in 2d CFTs the phenomenon of forbidden singularities in auto-correlation functions – a key signature of eigenstate thermalization. We show that they correspond to phase transitions in the context of eigenstates. The connection is made explicit by utilizing the AGT correspondence, which relates eigenstate auto-correlations to the Nekrasov partition functions describing an instanton gas of the $\mathcal{N}=2$ SUSY gauge theories. We show that by taking the counter-part of the heavy-light limit, two phases emerge for the instanton gas. They are dominated by configurations represented by string-like Young tableaux with distinct structures and thermodynamic properties, which bare resemblance to the confined and the deconfined phases. A phase transition occurs as instantons proliferate from one side, in a manner that mimics the Lee-Yang theory. We work out the critical fugacity and find it corresponding exactly to the forbidden singularity.

Many traditional models of electoral dynamics emphasize opinion change, social influence and persuasion mechanisms. In contrast, contemporary polarized societies often exhibit relatively stable partisan blocs coexisting with an electorally relevant mobile population. We propose a minimal dynamical model in which electoral competition is governed not by direct persuasion between opposing blocs, but by the dynamics and allocation of a mobile electoral interface. The electorate is partitioned into two stable partisan blocs and a mobile fraction, while direct transitions between the polarized blocs are strongly suppressed. Mobile voters are allocated between the competing blocs through a Fermi-like probabilistic rule governed by rejection asymmetries. Analytical calculations show that the electoral susceptibility to political shocks is proportional to the stationary size of the mobile electoral interface, identifying electoral mobility as the key quantity controlling the macroscopic response of polarized systems to external perturbations. The model naturally predicts a continuum of mobility regimes, ranging from frozen polarization to highly responsive electoral states characterized by a broad mobile interface. These results suggest that, in strongly polarized elections, aggregate electoral changes may be governed primarily by fluctuations at the mobile electoral interface rather than by large-scale ideological conversion.

As phase transitions in isolated quantum systems remain elusive, here we show how a thermodynamic-like phase transition, falling into the Lee-Yang paradigm, can arise in systems displaying eigenstate thermalization. Specifically, we show that in holographic conformal field theories, the eigenstate expectation of the auto-correlation function can be mapped to the partition function ${\cal Z}_{{gauge}}(z)$ of a virtual interacting instanton gas, with the conformal mapping of the imaginary time: $z=1-e^{-\tau}$ and the central charge $c$ mimicking the instanton fugacity and volume, respectively. We find that akin to the Lee-Yang paradigm, for $c\to\infty$ a pair of complex conjugate zeros of ${\cal Z}_{{gauge}}(z)$ move to the real axis located at the famous forbidden singularity. Passing through the singularity the system transits from the low- to high-fugacity phase, accompanied by dramatic changes in scaling behaviors of the free energy and dominant microscopic configurations. Our findings indicate that physics of phase transitions from eigenstate thermalization is very rich.

Inverse problems are ubiquitous in physics, chemistry, and engineering, arising when reconstructing hidden quantities from indirect measurements. A key example is the analytic continuation of imaginary-time correlation functions (iTCFs) to the real-frequency domain. This process requires an inverse Laplace transform, which is inherently ill-posed and highly sensitive to small input variations. Recent neural network (NN)-based methods have shown promising results by learning mappings from imaginary-time to real-frequency spectra, often outperforming traditional techniques such as maximum entropy. However, because the problem is ill-posed, many spectra fit the same iTCF. Regression-based approaches output a single solution, which approximates an average over the true solution space, and therefore fail to capture the full distribution of plausible power spectra. To address this issue, we introduce a diffusion-based framework for analytic continuation that learns the distribution of spectra consistent with a given iTCF. It offers two key advantages. First, it quantifies uncertainty directly from the learned distribution. Second, by analyzing the spread and structure of this distribution, we can quantitatively assess the intrinsic hardness of each inversion problem. We measure this hardness with a new metric, the uncertainty pseudo-volume. Applying the framework to an iTCF from a path-integral molecular dynamics simulation of liquid parahydrogen, we obtain the self-diffusion coefficient with an error bar and flag a secondary high-frequency peak as a possible spurious artifact. In contrast to previous attempts at uncertainty quantification, our generative approach rests on a concrete probabilistic basis, providing a more theoretically grounded measure of confidence in the reconstructed power spectra.

Using patch-matrix-product methods, we study two reversible three-state interaction-round-a-face cellular automata introduced by Klobas and Prosen [J. Phys. A 55, 094003 (2022)] - the species-preserving and species-flipping rules - and a coherent quantum deformation interpolating between them. Within an explicit translationally invariant ansatz, the two classical rules share a one-parameter family of quasi-local conservation laws with an auxiliary-dimension-three realization. Every regular member is also the first centered logarithmic derivative of an analytic auxiliary-dimension-two family passing through the identity observable. On the closed span of this family and the three elementary local charges, the only nonzero Euler velocities are $\pm\sqrt{3/23}$. Consequently, $\sqrt{3/23}$ is a rigorous lower bound on the maximal Euler speed in any larger conserved sector. For the species-flipping rule, this value agrees with the reported extrapolated value $0.361$ of Klobas and Prosen, strongly supporting completeness of the known sound-active charges. The species-preserving rule admits, in addition, a staggered generating family whose logarithmic derivatives form an infinite tower of strictly local charges. In the quantum deformation, the same algebraic structures yield exact low-bond-dimension invariant states (scar candidates), exponentially long-lived quasi-local quasimodes, and diagonal quasi-local charges that produce a nonzero Mazur bound after projection away from the three elementary local charges.

What a finite learning device has recorded and what will hold value for it on future tasks are not the same quantity. We develop a typed accounting for finite-state learning devices that separates four components: a training-side fit functional $\Phi_{\mathrm{fit}}$, the record-correlation stock $J_{D}=I(M;D)$, an update-side search ledger $\sigma_{M}$, and an operational capital value $V(M;T,b)$. This value is the work gap between an informed protocol class and a blind class obtained by deleting the memory-read port and re-optimizing from scratch. (I) Separation: for every $n$, there is a device family on which record correlation and world correlation grow by $n\ln 2$ while the capital gain is exactly zero. In the $\mathrm{flat}^{*}$ regime, data-free updates never increase $V$. (II) Capitalization ledger: an exact $\mathrm{flat}^{*}$ extraction identity and a universal ledger identity give, for (F5$'$)-stable $M$-local updates under a no-discarded-record-correlation condition (f), the bound $\eta_{\mathrm{cap}}\le 1$ for the capitalization efficiency $\eta_{\mathrm{cap}}=\Delta V/(k T\,\sigma_{M})$, together with necessary and sufficient conditions for equality. (III) Value retention: for the retention gap $L_{\mathrm{gen}}$ and retention ratio $\rho_{\mathrm{gen}}$ (the former carries no sign constraint; the latter is defined for positive training-side value and is not confined to $[0,1]$) we give a two-layer alignment domain: an exact exchange rate between value and the side-information-adjusted record fit $I(M';D\mid Y)$ without any record-side-information independence assumption, and a raw record-stock exchange rate under a joint side-information neutrality condition $(M,D)\perp Y$, whose boundary is marked by an explicit one-time-pad witness. These are statements about finite-device value retention under task-distribution shift, not a theory of statistical generalization.

Correlation functions $\langle O_1(x) O_2(0) \rangle$ reveal scaling dimensions through spatial decay. We instead consider static susceptibility, the change in $\langle O_1(x) \rangle$ from perturbing the Hamiltonian by $O_2(0)$, which we term causation for short. In a conformal field theory (CFT), dimensional analysis predicts decay of $|x|^{-2\Delta}$ for correlation and $|x|^{-2\Delta+1}$ for causation. Yet we find causation can decay up to fifteen additional orders in $x$ through a general mechanism, which we trace to time-derivative fields being unable to contribute to static response. In higher-dimensional CFTs, this mechanism ensures leading causation arises from primaries, even when descendants dominate correlation, which we leverage with DMRG to identify a previously unresolved corner primary of $\Delta \approx 8.8$ and a heavy magnetic line defect primary of $\Delta \approx 4.6$ in the $(2+1)$D critical Ising model. Moreover, the same mechanism governs edge-mode localization in $(1+1)$D gapless symmetry-protected topological phases, explaining previously observed anomalously small edge-mode splittings and guiding our construction of spin chains with splittings as small as $1/L^{18}$ and $1/L^{25}$.

The coupled dynamics of ions and water molecules in their first hydration shell impact a variety of processes including ion diffusion, selective ion transport in water-filled nanopores, and the kinetics of ion-pairing, ion adsorption, and metal-ligand binding reactions. In this work, we study these coupled dynamics for alkali metals (Li, Na, K, Rb, Cs), alkaline Earth metals (Mg, Ca, Sr, Ba), and chloride through the lens of their dependence on ion isotopic mass. Results are validated against previous measurements of the isotopic mass-dependence of ion diffusion coefficients in water and previous ab initio calculations of ion high-frequency dynamics in water. We find that the vibrational power spectra of ions in water consistently exhibit either two or three peaks, i.e., ions have several rattling frequencies within their solvations shells as previously reported for a subset of the species examined here. These frequencies have different sensititivies to isotopic mass that may serve as signatures of ion solvation processes (such as the tendency of ions to orient their first-shell water molecules) and that also may relate to Hofmeister-like effects including the relative affinity of different metals for ribonucleic acid (RNA).

We study a spin-$\frac12$ impurity coupled to the boundary of a strongly correlated spin-$s$ Takhtajan–Babujian chain, an integrable model whose low-energy physics is described by a perturbed $SU(2)_{2s}$ Wess–Zumino–Witten conformal field theory. While boundary conformal field theory determines the low-energy universality class of the weak-coupling regime, exact Bethe Ansatz methods reveal a sequence of boundary quantum phase transitions in which impurity-bound states emerge and reorganize the Hilbert space into multiple excitation towers built on distinct boundary configurations. This tower restructuring provides the organizing principle for a rich boundary phase diagram extending beyond the conventional Kondo regime. Weak antiferromagnetic coupling realizes the overscreened $2s$-channel Kondo universality class, whereas stronger couplings generate localized boundary modes and qualitatively new screening mechanisms. To describe the resulting thermodynamics, we develop a generalized thermodynamic Bethe Ansatz framework that captures the multi-tower structure across all regimes. The impurity entropy reproduces the boundary conformal field theory prediction in the overscreened Kondo regime but develops pronounced nonmonotonic temperature dependence once boundary-bound states appear, in quantitative agreement with large-scale finite-temperature matrix-product-operator simulations. Complementary dynamical calculations reveal sharp threshold features in the impurity spectral function that directly track the underlying tower structure. Together, boundary conformal field theory, exact Bethe Ansatz, generalized thermodynamic Bethe Ansatz, and tensor-network simulations provide a unified description of impurity screening, boundary-bound-state formation, and excitation-tower reconstruction in a correlated spin-$s$ chain.

We solve the Fokker-Planck equation for a Brownian particle in a tilted cosine potential and derive the intermediate scattering function (ISF), which captures the full spatio-temporal dynamics of the system. The model consists of a single overdamped Brownian particle in one dimension. We derive a generalized ISF comprising two wave vectors to describe correlations in the periodic potential. Exploiting the periodicity via Bloch's theorem, we formulate the problem within a spectral-theoretical framework and numerically compute the corresponding eigenfunctions and eigenvalues, from which we obtain the ISF and the probability density. Using time-dependent perturbation theory, we expand the ISF and derive low-order moments, including the mean-square displacement, time-dependent diffusivity, skewness, and the non-Gaussian parameter. Our analytical results are validated by Brownian-dynamics simulations and analyzed focussing on different regimes of the tilting force. The results are compared to a harmonic approximation and the deterministic limit.

We introduce a framework to classify quantum phases, phase transitions, and non-equilibrium dynamics of interfaces separating ordered bulk domains in 2D quantum magnets - equivalently, confining strings in dual lattice gauge theories - based on effective 1D Hamiltonians governing geometric fluctuations. Building on a "holographic" approach from [Phys. Rev. Lett. 129, 120601 (2022)], here reinterpreted as an exact bosonization, we uncover a rich quantum phase structure, with a variety of stiff and rough interface phases described by gapped and gapless 1D ground states, respectively, all distinguishable through the statistics of 2D wave-function snapshots. Our framework allows us to predict distinct spatiotemporal scaling laws for non-equilibrium curvature-driven interface dynamics across parameter space, which can be readily probed in existing experiments. We finally show that our approach enables the unprecedented experimental opportunity of directly measuring charge full counting statistics and symmetry-resolved properties of an encoded 1D system, as we explicitly demonstrate by numerically simulating a neutral-atom array experiment.

The min-plus process is a stochastic coagulation-annihilation-type process on the binary tree, of interest in mathematics, physics, and computer science as a tractable instance of max-type recursive distributional equations. We carry out large Monte Carlo simulations at effective tree depths up to $N=60$ that provide finite-depth corroboration of the Beta(2,1) stretched-exponential limit for its root value $X_{N}$ at $p=1/2$, on the asymmetric $\sqrt{N}$ side of the random-homogeneous-systems classification recently introduced by Chen, Duquesne, and Shi and by Morfe. Off criticality, our simulations confirm the sub-critical closed form $\mathbb{P}(X_{\infty}=1)=(1-2p)/(1-p)$ within Monte Carlo error and document a super-critical mean growth exceeding the elementary $(2p)^{N}$ lower bound at the depths we reach. For a Bernoulli($q$)-initial-condition variant, we identify an elementary closed-form identity at $p=1/2$ that pins down the order parameter $\mathbb{P}(X_{N}=0)=q$ exactly, locates the absorbing-state phase transition at $p_{c}=1/2$ in the operator-mixing probability rather than in the initial-zero density, and shows that the conditional law on positives deforms substantially with $q$. Our simulations use a level-wise recursion and an FFT-based precomputed leaf table which reduce the effective simulation depth while preserving the recursive tree law and may be useful for the simulation of related recursive equations on large trees.

The output state of a 2D geometrically local shallow random quantum circuit does not have long range correlations due to its lightcone structure. But this changes if one measures a subset of the qubits: long-range entanglement can be induced by the measurement process, leading to conditional correlations between distant qubits. In this paper we investigate the structure of conditional dependence in these circuits and its consequences for quantum advantage. For a tripartition $ABC$ of the qubits, we consider the ensemble of post-measurement states on $A$ that is conditioned on a specific measurement outcome on $B$ and ranges over all possible measurement outcomes on $C$. For circuit depths exceeding a constant critical value $d^*$, we conjecture that this ensemble is well approximated by a certain generalization of the Haar ensemble, called the Scrooge ensemble~[Jozsa et al., \href{https://doi.org/10.1103/PhysRevA.49.668}{Phys. Rev. A 49, 668 (1994)}]; we also provide supporting numerical and analytical evidence. Our conjecture describes a precise sense in which the state retains its lightcone structure on the remaining unmeasured qubits, but also develops some globally random features arising from the measurement. A consequence is that $n$-qubit shallow random quantum circuits in two dimensions are classically efficiently simulable in the presence of a tiny depolarizing noise rate $\Omega(\log(n)/n)$.

Microbial populations generally evolve in fluctuating environments under time-varying conditions. These are often described by binary switching models, sometimes seen as coarse-grained feast-famine cycles, in which resource availability switches abruptly between abundant and scarce conditions. However, experimental studies suggest that feast-famine environments actually exhibit more complex temporal dynamics. Here, we study how two strains, one growing slightly slower than the other, compete for the same resources in fluctuating environments comprising a finite number of intermediate states, each having its own carrying capacity. Environmental switching between these states and their carrying capacities represents gradual changes in nutrient availability. This class of multi-state stochastic switching models can be interpreted as a coarse-grained description of feast–famine cycles and allows us to investigate strain competition under the gradual recovery and depletion of resources. By computational and analytical means, we characterise the population dynamics in these multi-state fluctuating environments. In particular, we study how the switching rates and distribution of carrying capacities affect the population-size statistics, fixation probability, and mean fixation time. By comparing these results with their counterparts in binary environments, we clarify how the frequency and amplitude of environmental fluctuations influence population dynamics in coarse-grained feast-famine cycles.

Efficient evaluation of stationary parametric sensitivities over broad parameter ranges is important for identifying influential training data, fitting force fields, and predicting material responses, but standard pointwise approaches require repeated relaxation and sampling. Here we introduce Shortcuts to Parameter Sweeps (STPS), an engineered control strategy that uses an auxiliary control to transport the probability density along a prescribed family of instantaneous stationary states during a finite-time parameter sweep. This enables the continuous response curve over the full parameter interval to be estimated from a single controlled sweep using covariance-based response relations. STPS applies to both equilibrium and nonequilibrium steady-state systems, including those with unknown stationary distributions, and can be implemented directly using stationary samples in high-dimensional settings. Numerical tests on single-particle and interacting many-body systems show that STPS yields response curves in close agreement with reference results. These findings establish STPS as an efficient, sample-based framework for continuous sensitivity analysis in stochastic simulations.

The statistics of current fluctuations has long been a central object of study in non-equilibrium physics. Most existing work has focused on one-time statistics, while their multi-time generalisation remains comparatively less explored. We address this gap by extending the fluctuating hydrodynamics framework of Macroscopic Fluctuation Theory (MFT) to study multi-time statistics in the non-stationary state of a diffusive system on an infinite line. For the simplest cases of a non-interacting lattice gas and hard-core Brownian point particles, we present explicit solution of the MFT leading to multi-time large-deviation statistics. For generic systems, the MFT is solved perturbatively, yielding explicit results for two-time correlations. These reveal that the connection between current fluctuations and fractional Brownian motion, previously observed for flat initial conditions, does not persist for step initial conditions. We independently verify these hydrodynamic results by solving the corresponding microscopic dynamics for the non-interacting gas and for the symmetric simple exclusion process. Additional confirmation comes from numerical simulations.

The non-equilibrium properties of open quantum systems are determined by the microscopic laws governing energy exchange with their environment. In particular, an enhancement of the performance of quantum heat engines has been predicted by speeding up the dynamics through control of the system–bath interaction. However, direct microscopic control of heat transfer between the machine and the reservoir has remained elusive so far. Here, we experimentally demonstrate such control in a quantum Otto engine realized with ultracold Cs-133 atoms coupled to an atomic reservoir of ultracold Rb-87 atoms. Heat exchange between the two is mediated by inelastic s-wave collisions whose energy-dependent scattering cross sections lead to an asymmetric equilibration dynamics in the isochoric heating and cooling strokes. By tuning the kinetic temperature of the atomic reservoir, we modify the associated microscopic scattering rates, and thereby the heat transfer law, giving control over the time allocation within the engine cycle through control over the microscopic, multi-exponential relaxation dynamics. This enables power output optimization at fixed efficiency. Our results establish microscopic control of system-reservoir interactions as a tool for manipulating heat flow at the nanoscale and engineering the finite-time performance of quantum thermal machines.

We investigate stationary states of a collisionless, gravitationally stratified plasma atmosphere composed of electrons, protons, and alpha particles by extending Pannekoek–Rosseland theory to multispecies and multi-temperature plasmas. Starting from Liouville's theorem, we derive the self-consistent ambipolar electric field from kinetic equilibrium and charge neutrality. For a single-temperature atmosphere, we obtain analytical expressions for the ambipolar field, show its dependence on alpha-particle abundance, and determine the relative stratification of the three species. A first-order analytical approximation to the electrostatic potential accurately reproduces the numerical solution. We then generalize the formalism to multi-temperature plasmas generated by stochastic boundary heating, representing the stationary distribution as a superposition of Maxwellian populations. Gravitational filtering produces non-exponential density profiles and increasing temperatures with altitude, while preserving the relative species stratification, with alpha particles most strongly stratified and protons least. The ambipolar electric field contains a dominant gravitational contribution, corresponding to the generalized Pannekoek–Rosseland field, and a thermoelectric contribution arising from species-dependent temperature gradients, which accounts for its non-monotonic structure. These results provide a framework for studying the combined effects of plasma composition and stochastic heating in gravitationally stratified astrophysical plasmas.

Given a particular heat engine cycle, what is the optimal working medium that results in the highest efficiency? While one might jump to the conclusion that it must surely be the ideal gas, the situation is actually more intricate. Starting with a general Helmholtz potential that depends polynomially on molar volume and temperature we derive exact expressions for the ideal Stirling, Otto, and Brayton cycles. We find that for the thermodynamic systems described by our ansatz for the Helmholtz potential the maximal efficiency is achieved, if the working medium is described by a fundamental relation linear in temperature. This includes the ideal gas, but also classical harmonic oscillators and phenomenological models of the rubber band.

The melting of stripe-forming systems involves changes in connectivity that are not fully captured by conventional structural and orientational descriptors. We investigate a two-dimensional model with competing interactions whose low-temperature phase consists of one-particle-thick stripes. Molecular dynamics simulations along seven heating isochores are analyzed using thermodynamic, orientational, dynamical, and graph-based observables. Heating produces a multistage reconstruction in which the loss of stripe alignment and the reorganization of filament connectivity occur over distinct temperature ranges. Density controls whether the disordered filaments fragment into finite polymer-like clusters or remain joined in a dynamically fluid, system-spanning network. The topological observables distinguish these outcomes, which are not resolved by the thermodynamic and orientational responses alone. Thus, the same ordered stripe microphase can melt into topologically distinct fluids selected by density.

Magnetic interactions provide a versatile and powerful tool for controlling and organizing active matter, where individual units continuously consume energy to drive autonomous motion. These interactions arise naturally in biological systems, such as magnetotactic bacteria, and can be engineered into synthetic platforms, including colloidal microswimmers, magnetic nanoparticles, and macroscopic granular robots. This review focuses on active, self-propelled particles that carry an intrinsic magnetic dipole moment, powered by their own energy consumption rather than driven by external fields; here, the dipole moment mediates interactions and self-organization, not propulsion. We survey experimental and theoretical studies across all length scales, showing how dipolar interactions shape single-particle dynamics, collective behavior, and self-organization. We discuss models incorporating pairwise dipolar forces and confinement, and examine emergent phenomena such as chaining, swarming, and tunable pattern formation. We close by outlining challenges and opportunities in the design, control, and application of magnetic active systems, from programmable materials and biomedical actuation to nonequilibrium physics.

We study a Riesz gas with interaction parameter $s \in (d-3,d)$ in an arbitrary spatial dimension $d$, confined to a half-space by a hard wall. We prove that the equilibrium measure exhibits a dichotomy according to the interaction range. For $s \in (d-2,d)$, corresponding to the weakly long-ranged regime, the equilibrium measure always retains a non-trivial bulk component and thus is never completely confined to the wall. In contrast, for $s \in (d-3,d-2]$, corresponding to the strongly long-ranged regime, a confinement transition occurs: there exists a critical wall position beyond which the equilibrium measure is supported entirely on the wall. Our theorem generalises recent results established for the Coulomb gas, corresponding to the case $s=d-2$.

Colloidal stochastic engines are often regarded as microscopic analogues of macroscopic pis ton cylinder heat engines. However, although they share some underlying physical principles, such systems remain far from being direct force generators from a practical perspective. Motivated by this limitation, the present study introduces a polymer-based stochastic engine that more closely mimics the operation of piston cylinder engines. In this setup, heat is converted into work through a cyclic process in which a polymer is pulled into and out of a narrow channel under varying tem peratures. The work performed by the engine can be directly obtained from the cyclic trajectory in the force position diagram, analogous to the pressure volume diagram in traditional heat engines. Despite its much higher number of degrees of freedom compared to colloidal engines, the polymer en gine nevertheless follows several characteristic features observed in such systems. Numerical results demonstrate consistency with universal low-dissipation bounds for e ciency, recovery of Carnot e ciency under regeneration, and low-dissipation scaling of work and power.

The quantum exchange fluctuation theorem relates the probabilities of observing heat transfer along and against the temperature gradient between thermal baths at different temperatures. We investigate how this relation generalizes when the baths exhibit non-Markovian dynamics. Using a microscopic collisional model, bath memory is generated through interactions between successive bath auxiliaries before each heat-exchange collision. We derive exchange fluctuation theorems for both direct bath-bath interactions and probe-mediated heat exchange in the steady-state regime. As an illustrative example, we consider heat baths with qubit auxiliaries and show that non-Markovian memory enhances the probability of heat-transfer events against the temperature gradient, modifying the predictions made by the conventional Jarzynski-Wójcik exchange fluctuation theorem. Our results establish a microscopic connection between environmental memory and non-equilibrium heat-exchange statistics.

Extracting a thermodynamic nucleation rate from finite-time quantum dynamics requires separating observable decay from estimator and finite-size validity. We develop a multilevel identification framework for real-time tensor-network simulations of false-vacuum decay in the one-dimensional quantum Ising chain. Across twelve parameter points, the same coherent two-kink amplitudes semi-quantitatively predict both infinite-chain survival and magnetization dynamics: the survival coefficient has a median lattice-to-theory ratio of 0.896, while the magnetization-area slope ratios span 0.767–0.931. By contrast, the microscopic nearest-neighbour bond response is coherence dominated: vacuum–pair coherence contributes 60.0–81.5% across seven points with matched bond-dimension control, while substantial late-window slope discrepancies remain that cannot be removed by a scalar normalization. The analysis establishes finite-time survival and magnetization benchmarks and identifies the additional finite-size and branch-validation requirements for a bulk thermodynamic rate interpretation. Within the two-kink model and under the adopted common normalization, the lattice-resolved WKB action gives a median fixed-prefactor rate discrepancy of 4.13% from the coherent-bubble spectral calculation. These results distinguish finite-time lattice–theory consistency from the additional observable and finite-size evidence required to identify a thermodynamic nucleation rate.

We introduce the spectral density of geometric response, a measure of eigenstate sensitivity across an energy spectrum. Applied to many-body quantum systems, it reveals an exponentially sensitive integrability-to-chaos crossover, where eigenbasis deformations first accumulate in localized spectral regions before spreading throughout the spectrum. Physically motivated random-matrix ensembles reproduce this behaviour, whereas Gaussian ensembles do not, indicating universal features of the route to chaos that are absent from featureless random-matrix models.

We study unitary quantum dynamics in noisy Brownian models with global continuous symmetries, such as $U(1)$ and $SU(2)$, focusing on Rényi entanglement entropies and hydrodynamic and non-hydrodynamic correlators. By mapping the averaged late-time dynamics to the low-energy physics of effective replica Hamiltonians, we find that the evolution is controlled by the quantum geometry of their ground-state manifolds, which is directly related to the geometry of $k$-commutants—the symmetry algebra of $k$ replicas of the system. In interacting systems, these $k$-commutants are generically determined solely by the symmetries of the system, independent of microscopic details of the noisy evolution. This allows us to use the time-dependent variational principle (TDVP) to provide simple geometric explanations for the sub-ballistic Rényi entanglement growth and the anomalous decay of non-hydrodynamic correlators in interacting systems with continuous symmetries. We find this behavior to be intimately connected to singularities within the $k$-commutant manifolds, arising from frozen “void” states in the Hilbert space that exist due to continuous on-site symmetries. This also demystifies the important role of voids in the dynamics of these observables, previously identified in $U(1)$ symmetric systems. We compare these behaviors in interacting systems with Abelian and non-Abelian continuous symmetries and in free-fermion systems, which differ in the geometry of their $k$-commutants. Ultimately, this work provides a general geometric framework for systematically studying observables in noisy systems with continuous symmetries, including Haar-random circuits.

The mean first-passage time (MFPT) provides a universal temporal measure of transport, reaction, search, and switching processes in physical, chemical, and biological systems. Understanding how MFPTs respond to perturbations is therefore crucial for prediction and control, yet a systematic theory has been lacking. We establish a compact theoretical framework for linear and nonlinear MFPT response in continuous-time Markov processes. The key tool is an exact correspondence that maps the intrinsically transient response of MFPTs onto the steady-state response of an auxiliary system. This correspondence yields exact and universal response relations for MFPTs between arbitrary state pairs, expressed entirely in terms of unperturbed MFPTs and steady-state probabilities. We then obtain a factorized physical decomposition of the MFPT response into linear upstream, linear downstream, and nonlinear contributions. Further corollaries include response-curve inference rules, fundamental bounds on MFPT responses, analytical expressions for higher-order responses of MFPTs and steady-state probabilities, and multi-rate response formulas. Additionally, our result offers computational advantages in calculating both MFPTs and steady-state distributions. Finally, a biologically motivated folding network is analyzed, and a recently reported paradox on MFPT is clarified.

The non-equilibrium quantum thermodynamics of many-body open systems is notoriously rich, especially in the strong-coupling regime where strong system-bath correlations develop and interactions produce non-perturbative effects. Such systems can be described by `quantum impurity models' where the system and bath are treated on an equal footing as a single composite. Paradigmatic examples are the spin-boson and Anderson impurity models, in which system degrees of freedom interact with either bosonic or fermionic gapless baths. Here we study the quantum work distribution function (WDF) of such systems following a quench. Capturing the full continuum of many-body excitations in the WDF requires a non-perturbative solution of the underlying non-equilibrium quantum impurity problem. For this purpose, we extend the time-dependent numerical renormalization group (TDNRG) approach to the calculation of the WDF, which applies directly in the thermodynamic limit, can be used for arbitrary quench amplitudes at zero or finite temperature, and provides exponentially fine low-energy resolution. Our numerically-exact solution reveals power-law threshold behavior due to the Anderson orthogonality catastrophe when the work approaches its minimum value, with universal scaling collapse of the distribution below an emergent low-energy scale induced by system-bath correlations.