This paper proposes a simple yet effective jailbreak attack named FlipAttack against black-box LLMs. First, from the autoregressive nature, we reveal that LLMs tend to understand the text from left to right and find that they struggle to comprehend the text when noise is added to the left side. Motivated by these insights, we propose to disguise the harmful prompt by constructing left-side noise merely based on the prompt itself, then generalize this idea to 4 flipping modes. Second, we verify the strong ability of LLMs to perform the text-flipping task, and then develop 4 variants to guide LLMs to denoise, understand, and execute harmful behaviors accurately. These designs keep FlipAttack universal, stealthy, and simple, allowing it to jailbreak black-box LLMs within only 1 query. Experiments on 8 LLMs demonstrate the superiority of FlipAttack. Remarkably, it achieves $\sim$98\% attack success rate on GPT-4o, and $\sim$98\% bypass rate against 5 guardrail models on average. The codes are available at GitHub\footnote{https://github.com/yueliu1999/FlipAttack}.

We introduce the Binless Multidimensional Thermodynamic Integration (BMTI) method for nonparametric, robust, and data-efficient density estimation. BMTI estimates the logarithm of the density by initially computing log-density differences between neighbouring data points. Subsequently, such differences are integrated, weighted by their associated uncertainties, using a maximum-likelihood formulation. This procedure can be seen as an extension to a multidimensional setting of the thermodynamic integration, a technique developed in statistical physics. The method leverages the manifold hypothesis, estimating quantities within the intrinsic data manifold without defining an explicit coordinate map. It does not rely on any binning or space partitioning, but rather on the construction of a neighbourhood graph based on an adaptive bandwidth selection procedure. BMTI mitigates the limitations commonly associated with traditional nonparametric density estimators, effectively reconstructing smooth profiles even in high-dimensional embedding spaces. The method is tested on a variety of complex synthetic high-dimensional datasets, where it is shown to outperform traditional estimators, and is benchmarked on realistic datasets from the chemical physics literature.

Polaritons are a superposition of light and matter, that combine Strong Interferences (of light) with Weak Interactions (of excitons), making them WISI (Weakly-Interacting, Strongly-Interfering) particles. Their condensation is the main highlight of a field which occupies a unique position at the intersection of quantum optics, condensed matter physics and nonlinear dynamics of driven, dissipative systems. This chapter surveys selected theoretical concepts of polariton condensates' formation, coherence and dynamics, with an emphasis on their distinctions from their atomic counterparts and on points of ongoing controversy. We argue that linear and non-interacting effects are undervalued in polariton physics, and that a significant part of the phenomenology -- including bosonic correlations and coherence buildup -- can often be understood without invoking strong interactions or genuine quantum effects.

Ray tracing (RT) is a 3D graphics technique that offers highly realistic visuals. It is becoming prominent and accessible as GPU vendors have integrated dedicated ray tracing acceleration hardware. However, tracing millions of rays through 3D scenes consisting of high numbers of triangles in real time is challenging and requires expensive hardware. The main bottleneck in RT workloads is the expensive Bounding Volume Hierarchy (BVH) traversal task, which is a large tree structure that encodes the 3D scene. BVH traversal is a memory-bound problem, as the GPU threads spend most of their time reading tree node data from memory. In this work, we attack the memory latency bottleneck of ray tracing through prefetching. We propose a novel hardware prefetcher, named Tree Traversal Prefetcher (TTP), for ray tracing. The main idea is to leverage the existing tree traversal stack in the RT units for highly accurate prefetching. In particular, TTP prefetches nodes using the addresses already available on the hardware traversal stacks of each thread. For DFS (Depth-first search) based traversal, prefetches are generated when nodes are being popped consecutively from the traversal stack, potentially corresponding to upward traversal through the tree. We evaluate TTP on a cycle-level simulator, Vulkan-sim 2.0, and show that it achieves 1.48x speedup on average (up to 1.89x) compared to the baseline, with nearly negligible hardware overhead. TTP achieves 98.92% average L1 accuracy, which is the ratio of the prefetched blocks being actually referenced by demand loads. The coverage, computed as the ratio of L1 miss reduction over baseline L1 misses, is 31.54%, correlating well with the achieved speedup.

The classical Einstein-Langevin theory of Brownian motion assumes a memoryless thermal bath, establishing a universal fractal dimension of $d_v = 3/2$ for the velocity fluctuations of a particle. In this Letter, we demonstrate experimentally and theoretically that fluid-inertial memory effects fundamentally redefine the fractal scaling of these fluctuations. In analyzing highly resolved measurements of Brownian microspheres in liquids, we show that the non-Markovian hydrodynamic thermal noise establishes a distinct velocity fractal dimension of $d_v = 7/4$. Coupled with theoretical analysis of non-equilibrium short-time dynamics and the initial scaling of the velocity autocorrelation function, this result establishes the non-equilibrium universality class of Brownian motion in fluid media possessing a finite non-vanishing density.

Generative models are capable to address difficult problems with non-unique solutions like bandwidth extension and gap filling, removing highly non-linear artifacts from codecs, clipping and distortion, as opposed to removing linear additive components like noise and reverb. While large offline processing models have shown impressive results, these tasks have not been solved with real-time capable models with low latency and compute. We propose a few-step flow matching model using Data Prediction Mean Flows in combination with suitable novel low-latency architecture to make flow matching models an attractive choice under theses constraints. Compared to state-of-the-art, our proposed mean flow model uses 120x less compute and introduces no algorithmic latency other than the STFT, while achieving similar audio quality.

Entanglement and interference are among the most fundamental properties of quantum mechanics. In this work, we investigate the role and power of interference in the context of detecting entanglement. We do so from a computational complexity lens by proving that unentanglement gives no additional power to stoquastic Merlin-Arthur verification. For every polynomial number of provers $k=k(n)$, \[ \text{StoqMa}(k)=\text{StoqMa} . \] Conceptually, the proof separates the role of entanglement from the role of interference: once destructive interference is ruled out by stoquasticity, the product-state constraint can be absorbed into a polynomially larger one-witness stoquastic verification. The main analytic ingredient is a positive, value-based de Finetti theorem for separately symmetric extensions. If $M$ is an entrywise nonnegative positive semidefinite contraction on $A_1\otimes\cdots\otimes A_k$, then the nonnegative product value of $M$ is approximated to additive error $ε$ by the largest eigenvalue of \[ Π_R^{<k} (M_{A_{1,1}\cdots A_{k-1,1}A_k}\otimes I) Π_R^{<k}, \qquad R=O\!\left(\frac{k^2\sum_i\log\dim A_i}{ε^3}\right), \] where $Π_R^{<k}$ is the operator on $A_1^{\otimes R} \otimes \cdots \otimes A_{k-1}^{\otimes R} \otimes A_k$ projecting to the subspace $\mathrm{Sym}^R(A_1) \otimes \cdots \otimes \mathrm{Sym}^{R}(A_{k-1}) \otimes A_k$. The spectral relaxation is then realized as an actual one-witness stoquastic verifier. After replacing the uniform permutation averages in the symmetric projectors by inverse-polynomially close dyadic inverse-invariant averages. Consequently, \[ \text{StoqMa}(k)=\text{StoqMa}\subseteq\text{AM}\cap\text{PP}\subseteq\text{PSPACE} . \] The positive de Finetti theorem is isolated as a standalone technique and may be useful in other nonnegative tensor-optimization and stoquastic-verification settings.

Softmax and related normalized response functions are widely used in choice theory, machine learning, and cognitive science. In non-Boolean event structures with overlapping contexts, however, local normalization does not automatically yield a global probability weight. We show that imposing single-valuedness on shared atoms -- equivalently, no-disturbance or consistent connectedness -- collapses generalized softmax rules to coordinate parametrizations of the strictly positive part of the admissible-weight polytope. Any strictly positive admissible weight can be represented in this way, while boundary weights arise as limits. Exotic weights that exceed classical or quantum bounds are therefore properties of the event structure and the chosen weight, not of the normalizing link. The resulting hierarchy separates local normalization, cross-context gluing, Cauchy--Gleason linearity, and physical or cognitive realizability.

Owing to the Chapman-Kolmogorov equation for Markovian dynamics,any equilibrium trajectory of a Brownian particle in a solvent fluid can be viewed as the superposition of an uncountable number of non-equilibrium states. This property permits the unraveling of fine details of fluid-particle interactions at microscales defined by its non-equilibrium properties from the analysis of a single Brownian trajectory and to connect them to the hydrodynamics of the solvent fluid, simply considering the lower-order (second) moments of particle position in trapped conditions. In this way, the acceleration due to thermal-hydrodynamic fluctuational forces is isolated from the other factors and the short-time displacement statistics is completely determined by the correlation properties of the fluctuational thermal-hydrodynamic force. This approach not only confirms the $t^{5/2}$-law obtained by Boynewicz et al. (2026), related to fluid inertial effects, but indicates that this scaling may be superseded by a $t^4$-scaling at very short times once the correlated nature of the stochastic forcings is taken into account. The latter result is related to the regularity properties of particle velocity realizations.

This note introduces FRESH (Fusion of Recent Evidence and Subject Histories), a method for incorporating population-level summary results -- published clinical trials, registry summaries, prior natural-history studies, and peer-reviewed indirect comparisons -- into predictive models trained on patient-level data. This method provides a principled means of combining both patient-level and aggregate-level data types into a unified data-efficient model for clinical decision making. FRESH assumes access to a generative model trained on patient-level data sources (e.g. clinical trial or real-world data). The method produces patient-level predictions from a re-calibrated model that matches a set of specified aggregate statistics for a target population. This can be understood as a patient-level recapitulation of the aggregate source -- with the key property that the recalibration is a minimal perturbation of the original joint distribution in a specific information-geometric sense. The resulting samples can be analyzed directly or combined into a post-training procedure to update the original generative model. This approach enables several applications where rigorously incorporating patient-level data with summary information is valuable, including (i) contextualizing single-arm trial results with respect to recent standard-of-care, (ii) clinical-trial simulations for design and probability-of-technical-success estimation, and (iii) comparative-effectiveness analyses of on-market therapies.

Let $G$ be a finite group acting on a finite set $X$. This group action splits $X$ into disjoint orbits. The Burnside process is a Markov chain on $X$ which has a uniform stationary distribution when the chain is projected to orbits. We initiate the study of the Burnside process on Catalan structures. We consider two special cases: the first where the state space is the set of parking functions of length $n$ and $G = S_n$ is the symmetric group on $[n]$, such that $G$ acts by permuting coordinates, and the second where the state space is the set of labeled Dyck paths of length $2n$ and $G = S_n$ acts by permuting labels. The resulting Burnside processes give novel algorithms for sampling, respectively, an increasing parking function and a Dyck path approximately uniformly at random. Our main result shows that both processes are rapidly mixing, with mixing times upper bounded by $O(n \log n)$. As an application, we show how our Burnside process can be used to sample triangulations of an $(n+2)$-gon approximately uniformly at random.

We introduce a Stern--Gerlach (SG) spin-kinetic control parameter for magnetic reconnection. The fully projected branch parameter, $Ξ_0=<Z>/r_L$ compares the SG cross-sheet displacement accumulated during a diffusion-region transit with the relativistic Larmor radius. For an ensemble or partially participating population the relevant effective parameter is $Ξ_{\rm Eff}=P_{eff}Ξ_0$, where $P_{eff}$ represents the surviving branch weight or effective spin/moment projection. Evaluating $Ξ_{\rm Eff}$ across representative space and astrophysical environments reveals a robust hierarchy: SG transport is negligible in the magnetotail, solar corona, active galactic nuclei (AGN)/blazar jets, and pulsar-wind nebulae, but becomes transitional to strong in magnetar current sheets and extreme near magnetar surfaces. We further show, using electron--positron particle-in-cell simulations, that the SG force sorts particles by magnetic-moment projection into opposite sides of a Harris current sheet without measurably changing the global reconnection rate in the tested regime. This identifies magnetars as the clearest natural target for strong-field spin-kinetic reconnection ($Ξ_{\rm eff}\gg 1$) near the surface; transitional in the outer magnetosphere), while SG transport is safely negligible ($Ξ_{\rm eff}\ll 1$) in all heliophysical and jet environments considered, and provides a falsifiable framework for assessing where SG physics is relevant.

Altermagnets (AMs) have recently emerged as a distinct magnetic class bridging central features of ferromagnets (FMs) and antiferromagnets (AFMs), offering new opportunities for spin-based electronics. While they possess zero net magnetization like collinear AFMs, they simultaneously exhibit momentum-dependent spin splitting long thought exclusive to FMs. Despite intense theoretical interest, experimentally accessible materials hosting both altermagnetism and nontrivial band topology remain scarce. EuAgAs, crystallizing in space group $P6_3/mmc$, was previously identified via density functional theory (DFT) as a bulk altermagnetic Dirac semimetal. Contrary to these predictions, our neutron diffraction experiments reveal that the bulk ground state adopts a $\mathbf{q} = (0,0,\tfrac{1}{2})$ AFM structure with an in-plane $\uparrow\uparrow\downarrow\downarrow$ spin sequence. Systematic DFT calculations, however, uncover a remarkable near-degeneracy among competing magnetic orders: the FM and AM configurations lie only $0.11$ and $0.40~\text{meV/f.u.}$ above the AFM ground state, respectively. We further show that while a simple Heisenberg model favors a spin-spiral ground state, the inclusion of non-Heisenberg biquadratic coupling stabilizes the observed commensurate AFM phase. This near-degeneracy renders the magnetic state highly tunable, with DFT predicting a transition to the altermagnetic phase under hydrostatic pressure at approximately $14 \text{ GPa}$, establishing EuAgAs as a controllable platform for accessing topological altermagnetism.

We theoretically investigate acoustic spin resonance in a spatially homogeneous spinor polariton condensate. A longitudinal acoustic wave generates a time-periodic strain-induced effective magnetic field acting on the condensate pseudospin. When this field is transverse to the static in-plane linear-polarization splitting, it resonantly drives polarization oscillations. We show that spin-dependent interactions shift the resonance and produce nonlinear line shapes, while gain, reservoir dynamics, and spin relaxation make the response dissipative and history-dependent, producing amplitude hysteresis. In the presence of lifetime anisotropy, the condensate can develop a bifurcated stationary state with finite circular polarization, and a resonant acoustic drive can switch between the corresponding out-of-plane branches. A Zeeman splitting provides an additional conservative knob for tuning the resonance frequency. Our results identify coherent acoustic driving as a route to resonant, nonlinear, and switchable control of polariton pseudospin dynamics.

This is the first part of a series of papers proving the nonlinear stability of a one-parameter family of continuously self-similar $C^{1,α}$ naked singularity solutions, with $0<α\ll1$, to the spherically symmetric Einstein-scalar field equations. The stability holds for initial perturbations lying in a small open neighborhood of the data generating these naked singularity solutions, measured in a localized Hölder topology. These continuously self-similar naked singularity spacetimes were previously constructed by Christodoulou [D. Christodoulou, Examples of naked singularity formation in the gravitational collapse of a scalar field, Ann. of Math. 140 (1994), 607--653], who also proved their instability to black hole formation under sufficiently rough perturbations [D. Christodoulou, The instability of naked singularities in the gravitational collapse of a scalar field, Ann. of Math. 149 (1999), 183--217], thereby verifying weak cosmic censorship within a rough functional framework. In complete contrast, in this paper, we obtain the first nonlinear stability of these naked singularity spacetimes under general perturbations of the same regularity as the background. We rely on the linearized stability result established in the companion paper [J. Singh and W. Zheng, Nonlinear stability of continuously self-similar naked singularities for the Einstein--scalar field equations II: linearized stability]. Our result underscores the decisive role of the functional framework in formulating the Weak Cosmic Censorship conjecture.

Inclusive deep inelastic scattering factorization combines two features that are often treated separately: an asymptotic reconstruction of the current-current matrix element from hard and long-distance data, and an invariance under finite changes of collinear scheme or operator basis. We formulate these two features as a single proof object. The construction packages the leading-region analysis, overlap subtraction, Wilson-line reduction, finite scheme kernels and physical measurement into a typed, filtered structure on a compactified space of asymptotic regimes. Its central carrier is the balanced hard-collinear core over the interface algebra of finite scheme transformations. The hard QCD input is the construction of a scheme-balanced comparison map from this core to the collinear collar of the regime algebra. Once this comparison is an equivalence through the chosen power accuracy and the measurement descends to convolution, the standard DIS convolution formula follows formally and independently of the chosen scheme presentation. We separate this formal implication from the analytic QCD obligations needed to construct the collar equivalence, describe Collins-style subtraction as descent and Möbius inversion on the region poset, and give a finite check relating $\overline{\mathrm{MS}}$ and DIS presentations. The framework is intended as proof infrastructure rather than as a new calculation of DIS coefficient functions. It supplies diagnostics for missing regions, nonclosed operator sectors, nonbalanced measurements and failed collar equivalences, and it gives a typed interface for future proof-assistant and machine-learning implementations of factorization workflows.

Dependence among marginally constrained observations can break a finite-sample barrier. To formalize this phenomenon, we introduce the \emph{minimum list entropy coupling} $H(P\|Q_1,\dots,Q_m)$, the minimum conditional entropy $H(X|Y_1,\dots,Y_m)$ over all joint distributions with prescribed discrete marginals $X\sim P$ and $Y_i\sim Q_i$. Unlike classical formulations based on independent observations, our model allows $Y_1,\dots,Y_m$ to be arbitrarily dependent while keeping each marginal fixed. This enlarged coupling space reveals a sharp dichotomy: independent observations reduce residual uncertainty exponentially, whereas dependent observations can eliminate it exactly after finitely many samples. We characterize this zero-entropy regime through necessary and sufficient conditions and give concrete structural criteria under which it occurs. In particular, under mild support assumptions, zero entropy is achieved with $O(\log(1/P_{\min}))$ observations, where $P_{\min}$ is the minimum nonzero mass of $P$. We also develop a greedy algorithm with monotone approximation guarantees for computing $H(P\|Q_1,\dots,Q_m)$. Finally, we show that the same framework formalizes finite-sample limits in distribution-matching representation learning and randomness extraction, where zero entropy corresponds to exact recovery and exact extraction.

The first observation of a new meson state decaying into $B_c^+$ and a photon is presented using a dataset of $pp$ collisions at a centre-of-mass energy $\sqrt{s} = 13$ TeV collected by the ATLAS detector at the Large Hadron Collider, corresponding to an integrated luminosity of 140 fb$^{-1}$. The $B_c^+$ mesons are reconstructed from a three-muon final state from $B_c^+\to J/ψ(μ^+μ^-)μ^+ν_μX$ decays, along with photons reconstructed via conversions to electron-positron pairs in the detector material. The new state is observed with a significance that exceeds 8 standard deviations. The mass difference between the new meson state and the ground-state $B_c^+$ meson is measured to be $64.5 \pm 1.4\text{(stat.)}^{+1.0}_{-1.4}\text{(syst.)}$ MeV. This corresponds to a mass for the observed state of $6339.0 \pm 1.4\text{(stat.)}^{+1.0}_{-1.4}\text{(syst.)}\pm 0.3(m_{B_c^+})$ MeV, where the last uncertainty is due to the precision of the best knowledge of the $B_c^+$ meson mass. The low mass difference value matches the theory expectations for the lowest vector state of the $\bar{b}c$ quarkonium system, the $B_c^{*+}$ meson.

Deep Neural Networks (DNNs) are high-value intellectual property (IP), yet deploying them to edge environments exposes them to \textbf{unrestricted oracle access}, rendering them vulnerable to model extraction and inversion attacks. Existing defenses fail to address this practically: passive watermarking only offers post-hoc provenance, while active defenses impose prohibitive latency or require persistent access to sensitive training data. To bridge this gap, we propose \textit{LymphNode}, a novel post-hoc defense framework that acts as an intrinsic ``immune system" within the model. \textit{LymphNode} enforces a strict ``default-deny'' policy: it actively neutralizes model utility for unauthorized queries via \textbf{Generalized Sparse Universal Adversarial Perturbations (GSUAP)} injected into the feature space, effectively blocking gradient estimation and data inference. Utility is selectively restored only for authorized inputs carrying a stealthy feature-domain credential. Our framework is highly practical: it is \textbf{data-efficient}, establishing robust protection with fewer than 100 samples ($<1\%$ of training data), and \textbf{cross-dataset adaptable}, enabling protection using public surrogate datasets. \textit{LymphNode} thus provides a lightweight, immediately deployable defense for high-stakes scenarios where original training data is restricted or unavailable.

The study of optimal preemption policies for status update systems has been a recurring topic in the age of information (AoI) literature, where threshold-based structures have been shown to be optimal under a generate-at-will update generation model under certain assumptions. In this work, we study the effectiveness of threshold-based policies for a system with random update arrivals. In this regard, we introduce an analytical framework for evaluating the AoI of multi-threshold preemption policies and present interesting characteristics of the structure of the optimal preemption policy. We show the effectiveness of these threshold-based policies over the traditional probabilistic preemption policies and single-threshold policies, where we observe that significant gains in terms of AoI can be obtained by utilizing both the age of the packet and the age of the system when designing these preemption policies.

Given a polyhedron (planar, $3$-connected graph) $G$, we investigate its common neighbourhood graph con($G$). For cubic ($3$-regular) polyhedra, we show that the planarity of con($G$) depends on the number of odd faces of $G$, and on their adjacency. We then prove that for all other polyhedra, con($G$) is non-planar. We introduce a novel concept for polyhedra (and more generally, for plane graphs) $G$, namely the `facial common neighbourhood graph' facecon($G$). Its definition takes into account pairs of vertices with a common neighbour on the same face of $G$. It is a spanning subgraph of con($G$), that coincides with con($G$) for cubic polyhedra. It also generalises the reverse construction of the radial graph. As part of our investigation, we also prove a technical result of independent interest: if a maximal planar graph (triangulation of the sphere) has exactly two vertices of odd degree, then they are not adjacent. We also answer several questions in extremal graph theory. Fixing the number of vertices, we characterise the polyhedra $G$ such that con($G$) is planar and the number of edges in con($G$) is minimal/maximal. We address the same problem for facecon($G$), and prove that if it is maximal planar, then $G$ has no face of length greater than $6$. We notably characterise and explicitly construct all polyhedra $G$ of maximal face length $4$ such that facecon($G$) is maximal planar.

Probability integral transforms (PITs) and empirical $p$-values are widely used to assess the calibration of predictive distributions. While exact PIT values are uniformly distributed under correct model specification, practical implementations rely on empirical estimates constructed from finite samples. We show that this estimation step fundamentally alters the statistical structure of the problem. In particular, common-sample and rolling-window implementations introduce dependence and variance distortions that invalidate classical one-sample uniformity tests. When empirical percentiles are conditioned on a shared reference sample, the resulting statistics converge towards a two-sample Kolmogorov--Smirnov regime, while rolling windows induce autocorrelation and variance suppression. Our findings indicate that treating empirical percentiles as independent uniform draws can distort statistical inference and that backtesting procedures based on PITs require revised calibration methods accounting for the underlying two-stage sampling structure.

Gravitational models with non-minimal couplings involving functions of the matter Lagrangian and curvature have become popular in recent decades. By coupling the matter Lagrangian directly to the gravitational Lagrangian, one hopes to construct theories that can explain dark energy or dark matter without introducing additional sources. When this matter Lagrangian describes a perfect fluid, some technicalities are involved in its variational formulation. We present a careful derivation of the gravitational field equations together with the complete set of fluid equations. The fluid's equations allow us to define thermodynamic quantities such as temperature, chemical potential and number density and thus allow us to understand the effects of the non-minimal couplings on these quantities. We demonstrate the non-equivalence of the Lagrangian formulations of Schutz and Brown for these types of models and provide a detailed interpretation of our results.

Motivated by recent observations of integer and fractional Chern insulators in rhombohedral graphene stacks aligned with hexagonal boron nitride (hBN), we propose and study a model in which the moiré potential is defined by the pattern of layer-shear strain fields produced by lattice relaxation in these heterostructures. Although these strain fields decrease exponentially with the number of layers, their imprints on electrons residing away from the contact layer are non-negligible. In the absence of a displacement field, lattice relaxation effects amplify the electronic differences among the two different stackings with hBN. These differences, although attenuated at the single-electron level, survive in the so-called moiré-distant regime and are further enhanced with the inclusion of electron interactions. We find that lattice relaxation plays a crucial role in flattening and isolating a valley-polarized Hartree-Fock electron band with $|C|=1$ Chern number. Our results challenge the conventional wisdom on moiré effects in these heterostructures by illustrating the intertwined effects of long-range Coulomb interactions and lattice relaxation, and opens the door to explore different regimes of twist angles and displacement fields for the search for topological states.

Very recently, Green and Sawhney obtained a quasipolynomial bound in the Furstenberg--Sárközy theorem for square differences by proving an ''arithmetic level-$d$'' inequality, thereby yielding a greatly improved density increment scheme. We adapt their method to general intersective polynomials $h\in\mathbb{Z}[x]$ and obtain an analogous quasipolynomial upper bound for the largest subset of $\{1,2,\dots,X\}$ whose difference set contains no nonzero element of the form $h(n)$ with $n\in \mathbb{Z}$. This is the best quantitative upper bound presently known for sets lacking intersective polynomial differences. In contrast to the square case, extending the method to general intersective polynomials requires performing a density increment iteration in which the underlying polynomial changes at each step; a key contribution of this paper is to show that the arithmetic level-$d$ inequality remains effective uniformly across all auxiliary polynomials arising in the iteration. We also develop smoothly weighted versions of the exponential sum estimates of Rice.

Atomic-scale modeling has advanced rapidly through integration of machine learning, yet a key bottleneck remains. Even with an accurate potential energy surface and a clear target material, we still lack a practical atomistic dynamics framework that can simulate how materials form under realistic synthesis and processing conditions. Many processing transformations are governed by rare events in non-idealized evolving environments, while direct molecular dynamics is limited by femtosecond timesteps and short accessible trajectories. Existing acceleration methods often require prior mechanistic knowledge, including reaction coordinates, collective variables, event tables, or pathway guesses, which is rarely available in real experiments. Here we present an E(3)-equivariant deep reinforcement learning framework that enables goal-directed pathway discovery without hand-crafted reaction coordinates. The framework introduces a complementary operating mode for atomistic simulation in which realistic, non-idealized environments can be addressed directly while retaining kinetic plausibility through barrier-aware rewards. As a challenging benchmark, we target silicon dry oxidation, where rare-event pathways in amorphous SiO2 are effectively inaccessible to conventional atomistic methods. We treat an O2 molecule as an agent that performs continuous rigid-body translations and rotations in a Si/a-SiO2 environment. The agent is trained with an episode-level objective that rewards verified O2 dissociation while preferring low effective activation barriers. We demonstrate that the learned policy discovers kinetically favorable O2 diffusion and dissociation pathways in a disordered Si/a-SiO2 environment, progressively improving success rate while reducing effective activation barriers over training. We also discuss how the approach can be generalized to other processing and synthesis problems.

Sorting is a fundamental operation across numerous computational domains. Traditionally, this process involves transferring data from main memory to a processing unit for sorting, followed by writing the sorted data back to memory. This conventional approach incurs substantial latency and energy overheads due to the extensive data movement between memory and processing components. To mitigate these overheads, this paper introduces novel architectures for executing sorting operations directly within the memory fabric, eliminating the need for off-chip data transfer. To our knowledge, this work represents the first exploration of in-memory sorting using 6T SRAM. The proposed architecture is designed to operate on data represented in the standard weighted binary radix format commonly used in digital systems. The proposed architecture achieves a significant 3.4x reduction in latency compared to memristor-based IMC sorting.

The wolf note is an acoustic instability that occurs in large bowed string instruments when a strong body resonance interacts with the vibrating string, producing amplitude modulation and loss of tonal control. Various wolf suppressors - tuned mass dampers attached to the string or to the instrument body - are used in practice to mitigate this effect. In this paper, we propose a mathematical model describing the coupled dynamics of a string and a two-dimensional body equipped with one or two wolf suppressors. Both string and body include elastic (second-order) and stiffness (fourth-order) contributions and can be excited either by plucking or bowing. Three performance indicators are introduced: The first one perceives the wolf-tone appearance, the second one quantifies the attenuation of the notes possibly caused by the wolf suppressor, and the third one measures the acoustic fidelity (in terms of spectrum) with respect to the original instrument. The proposed numerical tests give insights about optimal tuning and placement of one or two suppressors, achieving effective wolf-note suppression while preserving as much as possible the global tonal balance.

We study how the stationary dynamics of an oscillator chain is modified when coupled to a bath of run-and-tumble particles. First, assuming time-scale separation, we derive the induced Langevin chain dynamics with explicit expressions for the streaming term, friction coefficient, and noise amplitude. At high persistence of the run-and-tumble particle bath, the linear friction turns negative, creating an instability. Second, we find that this anti-damping is arrested at long times due to nonlinear effects, reminiscent of a Rayleigh oscillator. We conclude that a passive harmonic chain can be transformed by its coupling to active matter into a self-sustained fluctuating medium with many-body Rayleigh-like dynamics. That transfer of activity results in pulsations of the displacements, spatial oscillations, and the emergence of persistence in velocities along the chain.

We present fully kinetic two-dimensional, three-velocity-component (2D3V) PIConGPU simulations of a three-beam direct-drive interaction with a 15 $μ$m solid-density cryogenic hydrogen cylinder, establishing a predictive numerical baseline for the operational DRACO ($τ=30$ fs) and upcoming PENELOPE ($τ=150$ fs) laser facilities at HZDR. The simulations resolve charge-separation fields on the order of 3 TV/m and reveal a robust kinematic bifurcation of the accelerated population into a fast (1-5 MeV) ion beam and a slower bulk (1-100 keV) flow. We demonstrate analytically and numerically that the charge-separation front ($v_{hb}$) is an intrinsically non-quasi-neutral electrostatic double layer that lies outside the closure assumptions of radiation-hydrodynamic models. A simple $2v_{hb}$ reflection scaling derived directly from the front trajectory tracks the centroid of the constant-energy fast-ion band under the impulsive 30 fs driver and the time-varying upper edge of the swept fast-ion band under the sustained 150 fs driver, across both intensities ($a_{0}=12.7$ and 22.0), establishing this non-thermal mechanism as the dominant acceleration pathway. We then scan an external axial magnetic field from 0 T to 10 kT. Laboratory-achievable 20 T fields leave all macroscopic observables unchanged; fields at the kT scale progressively magnetise the MeV hot-electron population, quench the laser-driven charge-separation mechanism, suppress the fast-ion band, and more than double the net-inward compression time of the short-pulse driver-while extending the outer target envelope. A geometric equivalence argument maps these kT-scale results onto larger-diameter cryogenic hydrogen jets.