In this paper, we systematically investigate the connection between linearizable objects and forward simulation. We prove that the sets of linearizable objects satisfying wait-freedom (resp., lock-freedom or obstruction-freedom) form a bounded join-semilattice under the forward simulation relation, and that the sets of linearizable objects without liveness constraints form a bounded lattice under the same relation. Thus, forward simulation is not only a proof technique for linearizability but also induces an algebraic hierarchy of linearizable objects. As part of our lattice result, we propose an equivalent characterization of linearizability by reducing checking linearizability w.r.t. sequential specification $Spec$ into checking forward simulation w.r.t. a wait-free universal construction $\mathcal{U}_{Spec}^{WF}$. We also propose an object $\mathcal{U}_{Spec}^s$, which simplifies $\mathcal{U}_{Spec}^{WF}$ and is more suitable for verification. We prove that the Herlihy-Wing queue is simulated by $\mathcal{U}_{Queue}^s$ with $Queue$ the sequential specification of the queue. Thus, our object $\mathcal{U}_{Spec}^s$ can be used in the verification of linearizability. To demonstrate the forward simulation relation between concrete linearizable objects, we prove that the time-stamped queue simulates the Herlihy-Wing queue, while the Herlihy-Wing queue cannot simulate the time-stamped queue. All these three proofs have been machine-verified by Isabelle/HOL.

Linear differential equations and recurrences reveal many properties about their solutions. Therefore, these equations are well-suited for representing solutions and computing with special functions. We identify a large class of existing algorithms that compute such representations as a linear relation between the iterates of an elementary operator known as a \emph{pseudo-linear map}. Algorithms of this form have been designed and used for solving various computational problems, in different contexts, including effective closure properties for linear differential or recurrence equations, the computation of a differential equation satisfied by an algebraic function, and many others. We propose a unified approach for establishing precise degree bounds on the solutions of all these problems. This approach relies on a common structure shared by all the specific instances of the class. For each problem, the obtained bound is tight. It either improves or recovers the previous best known bound that was derived by ad hoc methods.

Phase transitions are omnipresent in modern condensed matter physics and its applications. In solids, first-order phase transformations typically occur by nucleation and growth under non-equilibrium conditions. Under constant external conditions, $\textit{e.g.}$, constant annealing temperature and pressure, the nucleation and growth dynamics are often thought of as spatially and temporally independent. Here, $\textit{in-situ}$ Bragg X-ray photon correlation spectroscopy (XPCS) reveals nanoscale spatial and dynamical heterogeneity in the perovskite-to-brownmillerite topotactic phase transformation in La$_{0.7}$Sr$_{0.3}$CoO$_3$ thin films annealed under constant reducing conditions over a time span of multiple hours. Specifically, a timescale associated with domain growth remains stable, with a corresponding domain wall speed of $v_d = 6 \pm 0.5 \times10^{-4}$~nm/s ($2 \pm 0.2$~nm/h), while a slower timescale, associated with temperature-driven de-pinning of domains, leads to accelerating dynamics with timescales following an aging power law with exponent -2.2$\pm$0.5. This experiment demonstrates that Bragg XPCS is a powerful tool to study nanoscale dynamics in structural phase transformations, with the ability to extract quantitative average values related to nano-domain motion $\textit{in-situ}$. The results are relevant for phase engineering of phase-change devices, as they show that nanoscale dynamics, linked to domain and domain-wall motion, can continuously evolve and speed up with time, even hours after the initiation of the phase transformation, with potential repercussions on electrical performance.

Boson stars have been extensively studied in classical gravity, but their quantum properties remain comparatively unexplored. In this paper, we compute the quantum scalar fields and stress tensor in boson star spacetimes within the framework of semiclassical gravity. Divergences are regularized using Pauli-Villars fields, and accurate numerical results are obtained through spectral methods. Employing coherent states enables a direct comparison between the classical part of the stress tensor and the quantum fluctuation. Our results indicate that strong spacetime curvature is the primary source of large quantum effects. The renormalized quantum energy density is mostly positive but the radial pressure is negative, suggesting that classical boson star solutions require modification once quantum effects are included. Moreover, in regimes of large curvature, the quantum fluctuations can constitute a significant fraction of the total stress tensor. The methods developed here can be generalized to other compact objects and used to study their response to quantum corrections.

We propose a new method to estimate structural parameters in multi-way networks while controlling for rich structures of fixed effects. The method is based on a series of classification tasks and is agnostic to both the number and structure of fixed effects. In contrast to full maximum likelihood approaches, our estimator does not suffer from the incidental parameter problem. For sparsely connected networks, it is also computationally faster than PPML. We provide empirical evidence that our estimator yields more reliable confidence intervals than PPML and its bias-correction strategies. These improvements hold even under model misspecification and are more pronounced in sparse settings. While PPML remains competitive in dense, low-dimensional data, our approach offers a robust alternative for multi-way models that scales efficiently with sparsity. The method is applied to study the causal effect of a policy reform on spatial accessibility to health care in France.

The 1.3 mm ground-based very long baseline interferometry (VLBI) array Event Horizon Telescope (EHT), is limited by Earth's diameter, restricting its black hole shadow imaging to only M87* and Sgr A*. Extending baselines to the Moon would achieve ~0.7 uas angular resolution at 230 GHz, enabling shadow detection for a much larger sample of supermassive black holes (SMBHs). The concept is motivated by space VLBI missions and lunar exploration, including the ongoing Lunar Orbit VLBI EXperiment (LOVEX) aboard QueQiao-2 (Chang'E-7) and the planned International Lunar Research Station (ILRS). We assess shadow detectability for 31 SMBHs with predicted large angular sizes, exploring different telescope location and antenna size. Assuming a telescope at the lunar antipode, we simulate the Moon-Earth (u,v) coverage and show that sources with direction near the Moon's orbital plane yield projected baselines spanning from short to long, enabling sampling of the first visibility null - a key shadow signature. Using a geometric ring model, we identify six shadow-detectable candidates for Moon-Earth VLBI. Among these, M104, NGC 5077, and NGC 1052 are detectable with a 5 m lunar-based telescope; PGC 049940 requires 10 m; NGC 524 requires 20 m; and NGC 5252 requires 40 m. Furthermore, if space telescopes fill the baseline coverage gaps between Moon and Earth, the n=2 photon ring region is detectable for Sgr A*, M87* with a 10 m lunar-based telescope, and 12 candidates are detectable for the n=1 photon ring region using a lunar-based telescope of up to 40 m. These results provide a clear scientific and technical motivation for lunar-based telescopes in future black hole shadow studies.

The rapid scaling of deep neural networks comes at the cost of unsustainable power consumption. While optical neural networks offer an alternative, their capabilities remain constrained by the lack of efficient optical nonlinearities. To address this, we propose an optical neural computing architecture by embedding quantum emitters in inverse-designed nanophotonic structures. Due to their saturability, quantum emitters exhibit exceptionally strong nonlinearity compared with conventional materials. Using physics-aware training, we numerically demonstrate that the proposed architecture can solve complex tasks, including nonlinear classification and reinforcement learning, within all-optical neural networks. To enable fair comparison across different platforms, we introduce a framework that quantitatively links nonlinearity to a network's expressive power. Analysis shows that our quantum activation operates at $\text{nW}/μ\text{m}^2$ intensity, which is seven orders of magnitude below the nonlinearity threshold of conventional optical materials. Looking ahead to large language models, we estimate the nonlinearity-limited optical power, which scales sublinearly with model size. Our results indicate that quantum nanophotonics may provide a route toward sustainable AI inference.

In this work we study the mean-field description of Consensus-Based Optimization (CBO), a derivative-free particle optimization method. Such a description is provided by a non-local SDE of McKean-Vlasov type, whose fields lack of global Lipschitz continuity. We propose a novel approach to prove the well-posedness of the mean-field CBO equation based on a truncation argument. The latter is performed through the introduction of a cut-off function, defined on the space of probability measures, acting on the fields. This procedure allows us to study the well-posedness problem in the classical framework of Sznitman. Through this argument, we recover the established result on the existence of strong solutions, and we extend the class of solutions for which pathwise uniqueness holds.

We investigate dynamical dark energy models that admit non-phantom behaviour at early times, including thawing, scaling--thawing, and effective fluid extensions. Using current cosmological observations, we find that late-time background parameters remain stable across all models. Time-dependent parametrizations such as CPL show a $\sim2σ$ preference for phantom evolution at low redshift. Imposing non-phantom scaling dynamics at early times leads to strong lower bounds on the potential steepness, $λ\gtrsim 20$--$30$, constraining the early dark energy density to below the percent level at matter--radiation equality. Consequently, early scaling behaviour does not alleviate the Hubble tension and is penalised by Bayesian model selection. Our results indicate that while late-time dynamics can mildly improve the fit, early non-phantom scaling is strongly disfavoured by current data.

This paper develops and analyzes three families of estimators that continuously interpolate between classical quantiles and the sample mean. The construction begins with a smoothed version of the $L_1$ loss, indexed by a location parameter $z$ and a smoothing parameter $h \ge 0$, whose minimizer $\hat q(z,h)$ yields a unified $M$-estimation framework. Depending on how $(z, h)$ is specified, this framework generates three distinct classes of estimators: fixed-parameter smoothed quantile estimators, plug -- in estimators of fixed quantiles, and a new continuum of mean -- estimating procedures. For all three families we establish consistency and asymptotic normality via a uniform asymptotic equicontinuity argument. The limiting variances admit closed forms, allowing a transparent comparison of efficiency across families and smoothing levels. A geometric decomposition of the parameter space shows that, for fixed quantile level $τ$, admissible pairs $(z, h)$ lie on straight lines along which the estimator targets the same population quantile while its asymptotic variance evolves. The theoretical analysis reveals two efficiency regimes. Under light-tailed distributions (e.g., Gaussian), smoothing yields a monotone variance reduction. Under heavy-tailed distributions (e.g., Laplace), a finite smoothing parameter $h^{*}(τ) > 0$ strictly improves efficiency for quantile estimation. Numerical experiments -- based on simulated data and real financial returns -- validate these conclusions and show that, both asymptotically and in finite samples, the mean-estimating family does not improve upon the sample mean.

Fractional Chern insulators (FCI) are exotic phases of matter realized at partial filling of a Chern band that host fractionally charged anyon excitations. Recent numerical studies in several microscopic models reveal that increasing the bandwidth in an FCI can drive a direct transition into a charge-2e superconductor rather than a conventional Fermi liquid. Motivated by this surprising observation, we propose a theoretical framework that captures the intertwinement between superconductivity and fractionalization in a lattice setting. Leveraging the duality between three field-theoretic descriptions of the Jain topological order, we find that bandwidth tuning can drive a single parent FCI at $ν= 2/3$ into five different superconductors, some of which are intrinsically non-Abelian and support Majorana zero modes. Our results reveal a rich landscape of exotic superconductors with no normal state Fermi surface and predict novel higher-charge superconductors coexisting with neutral non-Abelian topological order at more general filling fractions $ν= p/(2p+1)$.

Scaling superconducting quantum processors is fundamentally limited by the escalating complexity of cryogenic wiring and the detrimental effects of microwave crosstalk and Purcell decay. This paper proposes a novel architecture based on frequency-multiplexed millimeter-wave superconducting qubits, integrating an on-chip cryogenic nonreciprocal space-time-periodic Josephson frequency multiplier as a universal control bus. The bus replaces multiple high-frequency XY drive lines with a single low-frequency input tone, which is parametrically converted into a comb of high-order harmonics, each resonantly addressing a distinct qubit. The nonreciprocal nature of the bus provides intrinsic isolation that suppresses Purcell decay and reduces coherent crosstalk by more than $98\%$ compared to a conventional reciprocal shared drive line. Full error-budget analysis demonstrates that the architecture can maintain gate errors below the fault-tolerance threshold for arrays exceeding 25 qubits, converting a crosstalk-dominated error budget into one primarily limited by intrinsic material coherence. Theoretical modeling based on a non-Markovian master equation further indicates that the engineered environment enables information backflow, offering a pathway to enhanced coherence. This integrated, frequency-multiplexed, and nonreciprocal control bus offers a compelling route toward dramatic I/O simplification, improved noise resilience, and scalable high-coherence superconducting quantum processors.

I examine whether early but temporary access to low-barrier hospitality employment affects refugees' labor market integration. I exploit within-region, within-year variation by combining the quasi-exogenous allocation of refugees to Austrian regions with seasonality in hospitality, where 25% of refugees first find work. Labor market access during high seasonal demand raises early employment probability by 3 percentage points (9% of the mean). Employment gains fade after one year, but treated refugees accumulate significantly higher three-year earnings, with no differences in medium-term wages or job quality. However, early hospitality work increases segregation into refugee-typical industries and firms with fewer Austrian coworkers.

In 1977, Chung, Chung and Liu generalized the definition of the Ramsey number. They introduced the $s$-chromatic Ramsey number as follows. Let $1\leq s< t$ be integers and let $A_{1}, A_{2}, \dots, A_{c}$ be subsets with size $s$ of $[t]$, where $c= {t\choose s}$. For given graphs $G_{1}, G_{2}, \dots, G_{c}$, the {\it $s$-chromatic Ramsey number} $r^{s, t}(G_{1}, G_{2}, \dots, G_{c})$, is the minimum positive integer $N$ such that every $t$-coloring of $E(K_{N})$ yields a copy of $G_{i}$ whose edges are colored by colors in the color set $A_{i}$ for some $i\in [c]$. The {\it star-critical $s$-chromatic Ramsey number} $r_{*}^{s, t}(G_{1}, G_{2}, \dots, G_{c})$, is the minimum integer $\ell$ such that every $t$-coloring of the edges in $K_{N}- E(K_{1, N- 1- \ell})$ yields a copy of $G_{i}$ whose edges are colored by colors in the color set $A_{i}$ for some $i\in [c]$, where $N= r^{s, t}(G_{1}, G_{2}, \dots, G_{c})$. If $G_{1}= G_{2}= \dots= G_{c}= G$, then we simplify them to $r^{s, t}(G)$ (also called the {\it weakened Ramsey number}) and $r^{s, t}_{*}(G)$, respectively. In this paper, we determine all the values of $r^{s, t}(K_{1, m})$ and $r_{*}^{s, t}(K_{1, m})$, and part of the value of $r^{s, t}(K_{1, m_{1}}, K_{1, m_{2}}, \dots, K_{1, m_{c}})$.

We use the Caley-Hamilton theorem to derive analytical solutions for the three-flavor neutrino propagation amplitude in a constant-density medium and their derivatives with respect to the mixing parameters. This approach avoids the diagonalization of the Hamiltonian and exploits precomputed matrix invariants to separate the dependence of oscillation probabilities on neutrino energy and propagation baseline. The results are implemented in the CHIC software, which provides simple, fast and efficient computation of oscillation probabilities and their derivatives. Finally, we demonstrate the value of probability gradients for neutrino data analyses and introduce a complementary visualization, the oscillograds, to probe underlying features of neutrino mixing.

We describe a pilot study to explore a new generation of fast and low noise CMOS image sensors for time domain astronomy, using two remote telescopes with a baseline of 1635 km. The experiment involves direct imaging with novel qCMOS image sensor technology that combines fast readout with sub-electron readout noise. Moreover, synchronized observations from two remote telescope sites will be used to explore new approaches for measuring Solar System bodies, precision stellar photometry, and speckle imaging. A fast-track installation of an ORCA-Quest2 camera at the Calar Alto Observatory 1.23m telescope has demonstrated the potential of the qCMOS technology for time domain astronomy. Numerical simulations suggest that owing to sub-electron readout noise, qCMOS sensors outperform classical CCDs for high-cadence imaging on 1m-class telescopes. The small penalty for post-readout binning, that is almost insignificant in comparison to higher readout noise detectors, opens interesting applications for scene-dependent data processing in direct imaging, and potentially even for spectroscopy.

In this work we introduce an effective approach to quantize the electromagnetic response of plasmonic metallic nanostructures. Their shape is arbitrary and they feature a realistic description of the frequency-dependent metal dielectric function that is based on experimental data. The derived quantum modes correctly reproduce the linear response macroscopic polarization of the nanoparticle upon external drive according to classical macroscopic Maxwell equations in the quasistatic limit. We further investigate the coupling of these modes to a quantum-chemical molecular description. The presented methodology paves the way for accurate modeling of plexcitonic system, where strong plasmon-molecule coupling and/or strong-driving fields call for a quantized description of the plasmonic response.

Over an arbitrary field, we conduct a comprehensive study of the polynomial identities and codimensions of two- and three-dimensional metabelian non-Lie Leibniz algebras. In addition, we compute the images of multihomogeneous polynomials on two-dimensional Leibniz algebras and, as a consequence, prove that the image of any multilinear polynomial evaluated on such algebras is always a vector space. Our analysis includes the three nontrivial isomorphism classes in dimension two and the ten isomorphism classes in dimension three, all of which are metabelian. In particular, we determine finite bases for their corresponding $T$-ideals and provide explicit bases for the associated relatively free graded algebras.

We present an original data analysis and a physical model that provide new insights into the origin of, and relationship between, two observables of the dusty, polarized Galaxy at intermediate and high latitudes: (i) the misalignment between HI filamentary structures and magnetic fields and (ii) the positive $TB$ correlation measured in Planck data suggesting parity violation in the interstellar medium (ISM). We confirm an observational link between the two effects and find that both are predominantly produced at large angular scales ($\geq 10^\circ$, multipoles $\ell \leq 20$) with a significantly stronger signal in the northern hemisphere. We propose a model in which filaments and magnetic fields appear misaligned in projection because they are sourced by cold and warm gas phases distributed in different proportions in the Solar neighborhood, from the wall of the Local Bubble to larger distances. These projection effects at large angular scales can produce coherent signatures that propagate to smaller scales in power spectra without invoking local, small-scale filament misalignment. Within this frame, HI filaments remain statistically aligned with the magnetic field in 3D, although with a projected scatter of tens of degrees that requires further investigation. The multiphase, geometrical model presented in this work is supported by Planck polarization data at 30 GHz, where synchrotron radiation dominates, and at 217 and 353 GHz, where dust emission dominates. Our analysis also incorporates starlight polarization measurements. The model introduced here suggests a new interpretation of two unexplained observables and emphasizes the role of the large-scale magnetized ISM in shaping polarized Galactic emission, which has important implications for both Galactic astrophysics and cosmological foreground characterization.

The outlines of a "Galois theory" for bimeromorphic geometry is here developed, via the study of model-theoretic definable binding groups in the theory CCM of compact complex spaces. As an application, a structure theorem about principal meromorphic bundles with algebraic structure group, and admitting no horizontal subvarieties, is deduced. Examples of algebraic groups arising as binding groups are provided, as is a characterisation of when they are linear. Using binding groups in CCM it is shown that, in contrast to the situation in differentially closed fields, there are many algebraic groups which admit nontrivial definable torsors over acl-closed sets in the theory DCCM of existentially closed differential CCM-structures. A self-contained exposition of the binding group theorem in totally transcendental theories, that emphasises the bitorsorial nature of the construction, is also included.

This article investigates the phenomenon of maximal rigidity in spatial processes, where perfect interpolation of the process is possible from partial information, specifically, from its restriction to a strict subdomain, often resulting in a trivial tail $σ$algebra. A classical example known since the 1930's is that a time series is fully determined by its values on the negative integers if its spectrum has a gap, or at least a sufficiently deep zero. We extend such results to higher dimensions and continuous settings by establishing a connection with the concept of uniqueness pairs, rooted in the uncertainty principle of harmonic analysis. We present several other manifestations of this principle, unify and strengthen seemingly unrelated results across different models: quasicrystals and stealthy processes are shown to be maximally rigid on cones, and discrete integer-valued processes are necessarily periodic when they have a simply connected spectrum. Finally, we identify a surprising class of continuous fields with seemingly standard behavior, such as linear variance and finite dependency range, that undergo a phase transition: they are perfectly interpolable on B(0, $ρ$) for $ρ$ ___ 2 $π$ but exhibit no rigidity for $ρ$ > 2.

The period between z ~ 3-6, a key transformational phase in galaxy evolution preceding 'cosmic noon' (z ~ 1-3), is very poorly explored in terms of feedback from AGN acting through gas outflows. In this work, we study the properties of outflows in AGN (mostly X-ray-selected) from the GOODS-S field, exploiting JWST NIRSpec IFU observations as part of the GA-NIFS GTO survey. Together with its twin sub-sample from COSMOS reported in a previous GA-NIFS work, this constitutes the largest spatially resolved sample of AGN outflows at these redshifts to date, comprising 16 targets with outflows (out of a total of 19 AGN), and probes the unexplored regime of AGN at z ~ 3-6 with bolometric luminosities ~$10^{45-46}$ erg/s. We mapped the rest-optical ionised gas emission lines at sub-kpc scales and spectrally isolated the broad wings tracing fast outflows from the gas at rest in the host galaxies. The incidence of ionised outflows in the GOODS-S + COSMOS GA-NIFS sample is high (>75 %), among the highest at any redshift. We inferred outflow velocities between ~600-2000 km/s, maximum radii of <1-4 kpc, and ionised gas mass outflow rates of ~0.1-100 $M_\odot$/yr, which in some cases can exceed the host galaxy star formation rate (SFR). We find that the outflow properties inferred for the GOODS-S + COSMOS GA-NIFS AGN sample and their relations with $L_{\rm bol}$ and SFR generally align with those observed for other spatially resolved literature samples of AGN outflows across different redshifts and AGN luminosities. Nonetheless, after accounting for any luminosity bias, our analysis suggests a cosmic evolution of the outflow properties, with higher median mass outflow rates (and possibly also mass loading factors) at higher redshifts, especially at z>3, indicating that AGN outflows were stronger in the early Universe than at later times, and thus potentially more capable of affecting their host galaxy.

Liquid mixtures can separate into phases with distinct composition. This phenomenon has recently come back to prominence due to its role in complex biological liquids, such as the cytoplasm, which contain thousands of components. For simple two-component mixtures phase-separated states are global free energy minima. However, local free energy minima, i.e. metastable states, are known to play a dominant role in complex systems with many components. For example, Hopfield neural networks can retrieve information from partial cues via relaxation to metastable states. Under what conditions can phase separated states be metastable, and what are the implications for information processing in multicomponent liquids? In this work we develop the general thermodynamic formalism of metastable phase separation. We then apply this formalism to an illustrative toy example inspired by recent experiments, binary mixtures with high-order interactions. Finally, as core application of the formalism, we study metastability in Hopfield liquids, a class of multicomponent mixtures capable of storing information on the composition of phases. We show that these phases can be retrieved from partial cues via metastable phase separation. Spatial simulations of liquids with a large number of components match our analytical solution. Our work suggests that complex biological mixtures can perform information retrieval through metastable phase separation.

A fundamental property of Segre classes is their birational invariance. This invariance implies that the Segre class of a closed subscheme only depends on the integral closure of the defining ideal sheaf. In this paper, we show that, conversely, the Segre class of a closed subscheme encodes an integral dependence criterion for its defining ideal sheaf. As an application, we prove that Aluffi's Segre zeta function provides an integral dependence criterion for homogeneous ideals in polynomial rings.

In a circuit QED architecture, we experimentally demonstrate a hardware-efficient and qubit-state-dependent Single-Step Phase-Engineered (SSPE) pulse scheme for actively depopulating a readout cavity. The protocol appends a reset segment with tailored amplitude and phase to a standard square readout pulse. Within the linear-response regime, the optimal reset amplitude scales proportionally with the readout amplitude, while the optimal reset phase remains invariant, significantly simplifying the experimental calibration procedure. Time-resolved measurements of the cavity photon number dynamics demonstrate that the SSPE scheme significantly outperforms the CLEAR protocol in terms of reset speed. Crucially, this approach enables arbitrarily fast, overshoot-free depletion of the cavity photon population, with the ultimate reset rate constrained by the finite analog bandwidth of the measurement chain. Furthermore, a comprehensive evaluation of the QND nature demonstrates that the SSPE scheme introduces no additional non-QND measurement errors. It exhibits non-QNDness comparable to both the free-decay and CLEAR protocols, with residual errors predominantly governed by state switching induced by qubit relaxation during the readout process. Thses results establish the SSPE scheme as a practical and scalable approach for achieving rapid and smooth cavity reset in superconducting quantum circuits.

We study energy minimization problems in quantum chemistry through the lens of computational algebraic geometry. We focus on minimizing the Rayleigh quotient of a Hamiltonian over a tensor train variety. The complex critical points of this problem approximate eigenstates of the quantum system, with the global minimum approximating the ground state. We call the number of critical points the Rayleigh-Ritz degree. We first study the Rayleigh-Ritz degree and introduce the Rayleigh-Ritz discriminant, which describes Hamiltonians that lead to a deficient number of critical points. We then specialize this framework to tensor train varieties: we identify instances when they are Segre products of projective spaces, report what we know about their defining ideals, and present a birational parametrization from products of Grassmannians. We use homotopy continuation to compute all critical points of this optimization problem over various tensor train and determinantal varieties. Finally, we use these results to benchmark state-of-the-art methods, the Alternating Linear Scheme and Density Matrix Renormalization Group.

The goal of the paper is to introduce a convergence à la Gromov-Hausdorff for Lorentzian spaces, building on $ε$-nets consisting of causal diamonds and relying only on the time separation function. This yields a geometric notion of convergence, which can be applied to synthetic Lorentzian spaces (Lorentzian pre-length spaces) or smooth spacetimes. Among the main results, we prove a Lorentzian counterpart of the celebrated Gromov's pre-compactness theorem for metric spaces, where controlled covers by balls are replaced by controlled covers by diamonds. This yields a geometric pre-compactness result for classes of globally hyperbolic spacetimes, satisfying a uniform doubling property on Cauchy hypersurfaces and a suitable control on the causality, and a curvature-driven pre-compactness result. The final part of the paper establishes several applications: we show that Chruściel-Grant approximations are an instance of the Lorentzian Gromov-Hausdorff convergence here introduced, we prove that timelike sectional curvature bounds are stable under such a convergence, we introduce timelike blow-up tangents and discuss connections with the main conjecture of causal set theory.

A persistence diagram is a finite multiset of birth-death pairs representing the lifetimes of topological features across a filtration. Existing functional and kernel representations of persistence diagrams are typically constructed extrinsically through embeddings into auxiliary spaces. For filtrations with finite indexing sets, the associated virtual persistence diagram group obtained by Grothendieck completion of the persistence diagram monoid is a finitely generated lattice. We define a phase map sending each persistence interval to a circular coordinate and a character map aggregating the phases of intervals in a virtual persistence diagram. We introduce heat damping on characters of virtual persistence diagram groups to suppress the unstable frequencies. We derive Lipschitz bounds for the resulting kernels and apply them in a synthetic segmentation experiment.

The ``Pre-train, then fine-tune'' paradigm has revolutionized Natural Language Processing (NLP). In this context, transferable backdoors pose a severe threat to the Pre-trained Language Models (PLMs) supply chain, yet defensive research remains nascent, primarily relying on detecting anomalies in the output feature space. We identify a critical flaw that fine-tuning on downstream tasks inevitably modifies model parameters, shifting the output distribution and rendering pre-computed defense ineffective. To address this, we propose Patronus, a novel defense framework that shifts the defensive focus from output features to input-side invariance, exploiting the fact that adversarial triggers remain constant even as model weights change. To overcome the convergence challenges of discrete text optimization, Patronus introduces a multi-trigger contrastive search algorithm that effectively bridges gradient-based optimization with contrastive learning objectives. Furthermore, we employ a dual-stage mitigation strategy combining real-time input monitoring with model purification via adversarial training. Extensive experiments across 15 PLMs and nine tasks demonstrate that Patronus achieves $\geq98.3\%$ backdoor detection recall and reduces attack success rates to clean settings, significantly outperforming all state-of-the-art baselines in all settings. Code is available at https://github.com/zth855/Patronus.

We propose a hybrid variational quantum algorithm that has variational parameters used by both the quantum circuit and the subsequent classical optimization. Similar to the Variational Quantum Eigensolver (VQE), this algorithm applies a parameterized unitary operator to the qubit register. We generate this operator using diabatic state preparation. The quantum measurement results then inform the classical optimization procedure used by the Cascaded Variational Quantum Eigensolver (CVQE). We demonstrate the algorithm on a system of interacting electrons and show how it can be used on long-term error-corrected as well as short-term intermediate-scale quantum computers. Our simulations performed on IBM Brisbane produced energies well within chemical accuracy.