Peer Revenue Past citations are no guarantee of future returns

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

Mathematics

math Watch Mathematics (sign in to use)

Friday 8 May

Probability

Quantitative ergodicity for gene regulatory networks with transcriptional bursting

We study the long-term behavior of two piecewise-deterministic Markov processes used to model stochastic gene regulatory networks with bursting dynamics. Under regularity assumptions on the jump rate, we prove the existence and uniqueness of the stationary distribution for an arbitrary number of interacting genes and an arbitrary strength of interaction. Using coupling methods, we also provide explicit upper bounds for the convergence to equilibrium in terms of Wasserstein distances.
Read the full abstract

Numerical Analysis · Cross-listed from cs.LG

When Attention Beats Fourier: Multi-Scale Transformers for PDE Solving on Irregular Domains

We study the problem of architecture selection for deep learning models trained to solve partial differential equations (PDEs), asking when transformer-based architectures with learned attention outperform Fourier-domain neural operators. We introduce the Multi-Scale Attention Transformer (\msat{}), a deep learning architecture that encodes spatiotemporal solution histories as token sequences and trains end-to-end via a composite supervised objective with optional physics-informed regularization terms. We conduct a comprehensive empirical evaluation against nine baselines – including physics-informed neural networks (PINNs), neural operators (FNO, DeepONet, GNOT), and state-space models (Mamba-NO) – across five benchmark problems from the PINNacle suite, using identical train/test splits and reference data for all methods. \msat{} achieves state-of-the-art generalization on complex geometry problems ($L^2_\mathrm{rel} = 0.0101$ on Heat2D-CG, a $3.7\times$ improvement over FNO) at $34\,\mathrm{s}$ total inference vs.\ $120{,}812\,\mathrm{s}$ for Mamba-NO. Ablation studies over the physics regularization component reveal a precise inductive bias tradeoff: physics priors reduce test error on diffusion-dominated problems but degrade generalization on chaotic and recirculating-flow regimes, directly characterizing the prior misspecification boundary. Approximation error bounds as a function of domain boundary complexity $\kappa$ provide a theoretical basis for these empirical findings and a principled rule for architecture selection.
Read the full abstract

All authors Brandon Yee, Pairie Koh, Jack Rodriguez, Mihir Tekal

General Mathematics

Physical derivation of the coarea formula and an elementary proof via gradient flow

In this note, we derive an elementary version of the coarea formula by considering the mass of a solid body with density $g (x)$. Then we present an rigorous proof using the changing variable formula. To this end we construct the diffeomorphism $\Phi$ via the gradient flow and compute its Jacobian determinant via geometric method.
Read the full abstract

General Mathematics

Combinatorics of higher order degenerate $r$-DERANGED Bell Numbers

When one inserts a number of identical bars in between blocks of an ordered set partition, they obtain a barred preferential arrangement. In this study we define a new generalization of barred preferential arrangements, by considering barred preferential arrangements with no fixed blocks, and where $r$ fixed elements on each section of the barred preferential arrangement are singletons. We derive several combinatorial identities for the number of these barred preferential arrangements. We also provide some asymptotic results.
Read the full abstract

Optimization and Control · Cross-listed from cs.LG

Hierarchical Mixture-of-Experts with Two-Stage Optimization

Sparse Mixture-of-Experts (MoE) models scale capacity by routing each token to a small subset of experts. However, their routers exhibit a fundamental trade-off: strong load balancing can suppress expert specialization, while aggressive diversity often causes routing collapse. We propose Hi-MoE, a grouped MoE framework that decomposes routing control into two coupled levels: (i) inter-group balancing that enforces fair traffic across expert groups, and (ii) intra-group specialization that promotes complementary expert behaviors while preventing within-group collapse. Our analysis provides a principled explanation of how our hierarchical objectives reshape the router, thereby promoting stable specialization and mitigating collapse. We observe consistent improvements over recent sparse-routing and grouped-MoE baselines across NLP and vision benchmarks, and confirm robustness via scaling studies (model size, expert count) and targeted ablations. In large-scale pre-training on 58B tokens, Hi-MoE-7B achieves a 5.6% perplexity reduction and a 40% improvement in expert balance over OLMoE-7B across diverse evaluation domains.
Read the full abstract

Information Theory

Embodied Communication: Sensing-Induced Reliability Fields and Capacity Bounds

This paper introduces embodied communication, a new wireless communication modality in which information is imprinted onto environmental states and recovered by the receiver through sensing. No dedicated communication transmitter is activated, and no additional communication spectrum is occupied; instead, the sensed environment itself becomes the carrier of information. The key insight is that sensing must be reinterpreted for communication. Rather than asking how accurately an unknown physical state can be estimated, embodied communication asks how reliably two states can be distinguished. We formalize this idea through a multi-snapshot radio frequency (RF) sensing model and derive a sensing-induced reliability field that quantifies the distinguishability between physical states. This field turns embodied symbol design into a geometric packing problem shaped by the sensing resolution of the infrastructure. For this embodied channel, we characterize the finite-snapshot $\epsilon$-capacity through achievable designs and converses. We develop lattice-based codebooks, obtain a closed-form hexagonal design under a main-lobe approximation, and establish information-theoretic and geometric upper bounds. We further reveal an intrinsic sensing-duration tradeoff: more sensing snapshots improve reliability, but also lengthen each embodied symbol, leading to a finite optimal sensing time. These results expose a latent communication pathway in sensing-enabled infrastructure and show how the environment can be transformed from a passive backdrop into an active information carrier.
Read the full abstract

Logic

Bourbaki–Zorn Normal Forms for Maximality Arguments

We isolate a normal-form mechanism underlying Bourbaki–Witt fixed-point arguments and least-upper-bound versions of Zorn-type maximality principles. Given a progressive self-map on a partially ordered set, we define a Bourbaki tower as a well-ordered trajectory whose successor stages are generated by the map and whose limit stages are given by least upper bounds of earlier stages. We prove that least upper bounds for nonempty well-ordered subsets are sufficient to force a fixed point for every progressive self-map. Thus the fixed-point statement is obtained under a weaker completeness hypothesis than the usual chain-complete form of the Bourbaki–Witt theorem. The proof proceeds by constructing a largest Bourbaki tower. The least upper bound of this largest tower belongs to the tower itself and is a fixed point of the map. As a consequence, strictly progressive self-maps cannot exist in such posets. Combining this obstruction with a choice selector on strict upper cones yields a concise maximality principle: if every nonempty well-ordered subset has a least upper bound, then the poset has a maximal element. The contribution is methodological rather than axiomatic. The paper makes explicit a reusable proof architecture connecting well-ordered Bourbaki–Witt fixed points, strict progression obstructions, and least-upper-bound versions of Zorn-type maximality arguments.
Read the full abstract

General Mathematics

CrackMorph-XAI-Net: A Topology-Preserving and Explainable Framework for Automated Crack Morphology

Automated crack inspection is increasingly recognized as a critical component of infrastructure monitoring; however, cracks continue to be reported primarily as binary segmentation masks by many current vision-based systems. While localization is facilitated by such masks, limited structural information is provided for robust engineering interpretation. For practical crack assessment, measurable morphological features – including centerline geometry, branching behavior, junction locations, topology, and severity-related indicators – are required. In this work, CrackMorph-XAI-Net, an explainable morphology-aware framework for image-based crack analysis, is presented. Crack image and region-mask data are converted into a sequence of interpretable structural outputs through four distinct stages: topology-preserving skeleton extraction, junction detection via Gaussian heatmap regression, morphology descriptor computation, and severity-oriented screening. To support rigorous stage-wise evaluation, the standard CRACK500 benchmark is extended with aligned skeleton maps, junction heatmaps, and topology labels. Experimental validation demonstrates that a mean Dice coefficient of 0.991 is achieved by the learned skeleton extraction stage, with topology preserved in 98.5% of test images. Furthermore, a recall of 0.964 and an F1-score of 0.887 are obtained in the junction detection stage, highlighting the efficacy of heatmap regression for sparse structural targets. Strong agreement between predicted and reference morphology values is revealed by descriptor-level evaluation, with correlations exceeding 0.95 for length, width, orientation, junction count, and tortuosity.
Read the full abstract

Information Theory

Chase-like Decoding: Test-Pattern Design and Performance Analysis

Chase-like decoding algorithms are a popular choice for soft-input decoding of algebraic codes. We evaluate different test-pattern sets for Chase-like decoding. Structured sets, such as Chase-II patterns or patterns chosen by logistic weight, are analyzed using order statistics, while arbitrary sets are evaluated by calculating covered-space probabilities and by performing Monte Carlo simulation. We further propose an algorithm that designs test-pattern sets to cover likely error patterns, achieving comparable performance with half the number of test patterns compared with conventional sets for high-rate BCH codes.
Read the full abstract

All authors Tim Janz, Simon Obermüller, Andreas Zunker, Stephan ten Brink

Mathematical Physics · Cross-listed from cond-mat.stat-mech

Multiscale Structure of Eigenstate Thermalization

The eigenstate thermalization hypothesis provides a framework for understanding thermalization in isolated quantum many-body systems by characterizing statistical properties of local observables in energy eigenstates. Here we demonstrate that distributions of matrix elements in macroscopic systems may depend not only on the macrostate parameters, such as the densities of local conserved charges, but generally also on the properties of ensembles used in sampling eigenstates. To this end, we depart from the conventional analysis of microcanonical windows and consider statistical ensembles with an adjustable scale parameter prescribing the magnitude of charge fluctuations. We specifically consider an integrable field theory that permits efficient numerical sampling of matrix elements and reliable extrapolation to the thermodynamic limit. Moreover, in this system, we identify a class of states that enables explicit closed-form computation of the suppression rate of matrix elements. Our findings reveal an underlying multiscale structure of matrix elements captured by a non-analytic fluctuation-scale dependence of algebraic exponents governing their statistical properties.
Read the full abstract

Statistics Theory · Cross-listed from stat.ML

A Note on Non-Negative $L_1$-Approximating Polynomials

$L_1$-Approximating polynomials, i.e., polynomials that approximate indicator functions in $L_1$-norm under certain distributions, are widely used in computational learning theory. We study the existence of non-negative $L_1$-approximating polynomials with respect to Gaussian distributions. This is a stronger requirement than $L_1$-approximation but weaker than sandwiching polynomials (which themselves have many applications). These non-negative approximating polynomials have recently found uses in smoothed learning from positive-only examples. In this short note, we prove that every class of sets with Gaussian surface area (GSA) at most $\Gamma$ under the standard Gaussian admits degree-$k$ non-negative polynomials that $\eps$-approximate its indicator functions in $L_1$-norm, for $k=\tilde{O}(\Gamma^2/\varepsilon^2)$. Equivalently, finite GSA implies $L_1$-approximation with the stronger pointwise guarantee that the approximating polynomial has range contained in $[0,\infty)$. Up to a constant-factor, this matches the degree of the best currently known Gaussian $L_1$-approximation degree bound without the non-negativity constraint.
Read the full abstract

Information Theory · Cross-listed from quant-ph

Covert Signaling for Communication and Sensing over the Bosonic Channels

Preventing signal detection in communication and active sensing requires careful control of transmission power. In fact, the square-root laws (SRL) for covert classical and quantum communication and sensing prescribe that the average output energy per channel use scales as $1/\sqrt{n}$ for $n$ channel uses. Diffuse and sparse signaling achieve this. The former transmits signals whose energy decays as $1/\sqrt{n}$ over all $n$ channel uses, which is convenient for mathematical analysis. The latter transmits constant-energy signals only approximately $\propto\sqrt{n}$ times out of $n$ channel uses, remaining silent on the others. This offers significant practical advantages in compatibility with modern digital transmitters. Here, we study sparse signaling over the lossy thermal-noise bosonic channel, which is a quantum model of many practical channels (including optical, microwave, and radio-frequency). We characterize the input signal state that minimizes detectability. We find an unintuitive optimal quantum state structure: a mixture of just two consecutive photon-number states. In particular, in the low-brightness regime, the optimal signal state is a mixture of vacuum and a single photon. Since these states are generally suboptimal for both communication and active sensing, we explore the resulting trade-off and identify input-power thresholds for transitions between optimizing covertness and performance in communication and sensing tasks.
Read the full abstract

All authors Tianrui Tan, Evan J.D. Anderson, Michael S. Bullock, Boulat A. Bash

Mathematical Physics

Hamiltonian formulation of the supersymmetric KdV equation

We studied the constrained Hamiltonian formulation of a supersymmetric Korteweg-de Vries (KdV) equation, which is observed to be a constrained system similar to its classical version. We found a nontrivial Lagrangian description, where we select $a=2$ for the free parameter $a$ in the supersymmetric extension. The corresponding degenerate Lagrangian requires an exclusive consideration and the utilization of the Dirac-Bergmann algorithm. We explicitly determined the full set of primary and secondary constraints and constructed the total Hamiltonian governing the dynamics of the system. In this analysis, in addition to a nontrivial constraint involving the fermionic fields, the consistency conditions give rise to a nonlocal contribution to the Hamiltonian density. This highlights a distinctive feature of this supersymmetric extension. We showed that the resulting Hamilton equations of motion reproduce the supersymmetric KdV system in the component form. Finally, we derived a compact superspace representation of the Hamiltonian and demonstrated its consistency with the component-level formulation.
Read the full abstract

Algebraic Geometry

Finite order symplectic birational self-maps on Kummer-type manifolds

A projective hyperkähler manifold of Kummer-type is said to be twisted modular if it is birational to the Albanese fiber of a moduli space of twisted sheaves on an abelian surface. We prove that, with the exception of certain cases of Picard rank 3, any projective Kummer-type manifold admitting a finite-order symplectic birational self-map that acts nontrivially on its second cohomology group is twisted modular. We provide a complete characterization of these exceptions in terms of their Néron-Severi lattices. We then investigate symplectic birational self-maps of modular Kummer-type manifolds, determining exactly which Mukai vectors allow the birational transformation induced by crossing the vertical wall, which acts on cohomology as a reflection, to correspond to a finite-order symplectic birational self-map. Additionally, we prove in an appendix several results concerning moduli spaces of twisted sheaves on abelian surfaces which were not readily available in the literature.
Read the full abstract

All authors Yajnaseni Dutta, Dominique Mattei, Stevell Muller, Howard Nuer

Probability

Rapid phase ordering of Ising dynamics on $\mathbb Z^2$

We consider the phase ordering problem for the low-temperature Ising dynamics initialized from a biased and disordered initialization. Work of Fontes, Schonmann, Sidoravicius (2002) showed that at zero-temperature, Ising Glauber dynamics on $\mathbb Z^d$ for $d\ge 2$ initialized from i.i.d. spins on each vertex that are $+1$ with sufficiently large probability, absorbs into the all-plus configuration quickly. We prove that analogous behavior holds throughout the low-temperature regime of the Ising model in two dimensions. Namely, there exists $p_0 <1$ such that Ising Glauber dynamics initialized from i.i.d. spins that are $+1$ with probability $p>p_0$, run at any low temperature $\beta>\beta_c$ converges rapidly to the plus phase measure $\pi^+$. The result is proved using a spacetime multiscale coupling valid in any $d\ge 2$, that boosts a uniform-in-$\beta$ quasi-polynomial bound on the mixing time of Ising dynamics with plus boundary conditions, into rapid phase ordering from biased initializations with no boundary conditions.
Read the full abstract

Numerical Analysis

Invariant domain preserving limiting of time explicit and time implicit discretizations for systems of conservation laws

This work concerns the design and analysis of a limiting technique that allows the preservation of invariant domains for high-order numerical approximations of nonlinear hyperbolic systems of conservation laws. The method can be applied to any conservative discretization method in space as well as to a wide range of explicit and implicit time integration schemes. The method limits the high-order solution around a low-order accurate solution that is known to preserve all the invariant domains. It generalizes the flux-corrected transport limiter [J. P. Boris and D. L. Book, J. Comput. Phys., 11, 1973; S. T. Zalesak, J. Comput. Phys., 31, 1979] to systems of conservation laws and relies on the limitation of antidiffusive fluxes, but defines the limiting coefficients so as to express the limited solution as a convex combination of invariant domain preserving quantities similarly to the convex limiting framework [Guermond et al., Comput. Methods Appl. Mech. Engrg., 347, 2019]. We give details on the derivation of this limiting technique and provide some illustration with finite volume or discontinuous Galerkin (DG) space discretizations associated to explicit or implicit Runge-Kutta methods as well as to time DG integrations. The limiter is applied iteratively to refine the limited solution around the high-order one, while preserving the invariant domains, and a heuristic is proposed to accelerate its convergence. Numerical experiments solving one- and two-dimensional problems involving scalar hyperbolic equations and the compressible Euler equations are presented to illustrate the properties of these schemes.
Read the full abstract

Combinatorics

Weak Order on the MacNeille Completion of Bruhat Order

Let $\mathrm{Mac}(W)$ be the MacNeille completion of the Bruhat order of a Coxeter group $W$. We introduce an action of the $0$-Hecke monoid of type $W$ on $\mathrm{Mac}(W)$, which allows us to define a weak order and a descent set statistic on $\mathrm{Mac}(W)$. When $W$ is of type $A$, we recover constructions of Hamaker and Reiner, which were originally formulated in terms of monotone triangles and alternating sign matrices. Using this action, we prove that certain unions of Knutson–Miller subword complexes are vertex-decomposable. By specializing to type $A$, we prove a conjecture of Escobar, Klein, and Weigandt regarding Cohen–Macaulay ASM varieties. Along the way, we also exhibit a counterexample to a conjecture of Hamaker and Reiner regarding the poset topology of intervals in the ASM weak order. When $W$ is finite, our action also gives rise to a Boolean cell complex with one facet for each element of $\mathrm{Mac}(W)$. We prove that this is an abstract simplicial complex and that every linear extension of the weak order on $\mathrm{Mac}(W)$ is a shelling order for its facets. Finally, when $W$ is finite and irreducible, we use our $0$-Hecke action to introduce a noninvertible dynamical system on $\mathrm{Mac}(W)$ that we call the MacNeille pop-stack operator, and we prove that the maximum number of iterations of this operator needed to reach the bottom state is $h-1$, where $h$ is the Coxeter number of $W$. This article is meant to serve as a case study in using large language models to automate the workflow of mathematical research. The proof of the conjecture of Escobar–Klein–Weigandt and the disproof of the conjecture of Hamaker–Reiner were obtained autonomously by ChatGPT 5.4 Pro. Other aspects of the paper were obtained mostly by the author, but ChatGPT expedited the process. We provide a detailed account of this interaction.
Read the full abstract

Optimization and Control

Approximate directional stationarity and associated qualification conditions

Approximate stationarity conditions provide necessary optimality conditions without requiring additional assumptions by demanding that a perturbed stationarity system possesses solutions as the involved perturbations tend to zero. Together with associated approximate constraint qualifications, which are typically rather mild, they raised much interest in the optimization community during the last decade. In parallel, directional stationarity conditions became quite popular as they sharpen standard stationarity conditions by incorporating data associated with underlying critical directions. The purpose of this paper is twofold. First, we melt the aforementioned concepts of approximate and directional stationarity to formulate and study so-called approximate directional stationarity. For the underlying model problem, an optimization problem with nonsmooth geometric constraints is chosen, which covers diverse practically relevant applications. The role of approximate directional stationarity as a necessary optimality condition is investigated in much detail, complementing results from the literature. Second, we formulate a qualification condition which, based on an approximately directionally stationary point, can be exploited to infer its directional stationarity. The latter condition depends on one particular sequence verifying approximate directional stationarity and merely requires to check a simple condition of Mangasarian–Fromovitz type stated in terms of the directional tools of limiting variational analysis. This contrasts standard approximate constraint qualifications that typically demand a certain stable behavior of all sequences validating approximate stationarity. Throughout, various approaches to verify directional stationarity of local minimizers are established, and illustrative examples are presented to make the theoretical results more accessible.
Read the full abstract

Differential Geometry

Asymptotics of small eigenvalues on degenerations of Kähler manifolds

We derive the exact asymptotic rates of the small eigenvalues of the Laplacian on one-parameter degenerations of compact Kähler manifolds equipped with induced background metrics. This generalizes a recent result of Dai and Yoshikawa to higher dimensions. To achieve this, we combine Li's uniform Skoda inequality with the method of auxiliary Monge-Ampère equations, introduced by Guo–Phong–Song–Sturm–Tong and adapted by Guedj–Tô. As an application, we establish estimates for degenerations of compact Kähler manifolds with reducible singular fibers.
Read the full abstract

Dynamical Systems · Cross-listed from q-bio.NC

Dynamical mechanisms of flexible phase-locking in cortical theta oscillators

Oscillatory activity in auditory cortex is thought to play a central role in auditory and speech processing by synchronizing neural rhythms to external acoustic features of the speech stream. To support this function, cortical oscillators must flexibly phase-lock to inputs spanning a wide range of timescales, including rhythms substantially slower than their intrinsic frequency. Here we identify a general dynamical mechanism by which intrinsic inhibitory currents operating on multiple timescales enable such flexible phase-locking. Using tools from dynamical systems theory, we show that interactions between slow and superslow inhibitory processes generate prolonged post-input recovery delays through delayed Hopf phenomena, thereby substantially expanding the frequency range over which entrainment can occur. We demonstrate this mechanisms in a biophysically grounded cortical theta oscillator model for speech segmentation. Specifically, we show that both a theta-timescale (4-8 Hz) inhibitory current $I_m$ and a slower delta-timescale (1-4 Hz) inhibitory potassium current $I_{\rm K_{SS}}$ are crucial for entrainment flexibility. Their interaction creates a three-timescale structure that gives rise to pronounced delay phenomena associated with a delayed Hopf bifurcation (DHB). Interestingly, the superslow $I_{\rm K_{SS}}$ and the associated DHB play little role in the unforced oscillatory dynamics, but are recruited to support phase locking under external forcing. Moreover, the intermediate-timescale current $I_m$, rather than being redundant, further expands the phase-locking range by prolonging delayed recovery along the superslow manifold. Together, these results suggest that coordination among intrinsic inhibitory currents operating on multiple timescales may represent a key mechanism supporting flexible phase locking to rhythmic inputs in the brain.
Read the full abstract

Optimization and Control

Penalty-Based First-Order Methods for Bilevel Optimization with Minimax and Constrained Lower-Level Problems

We study a class of bilevel optimization problems in which both the upper- and lower-level problems have minimax structures. This setting captures a broad range of emerging applications. Despite the extensive literature on bilevel optimization and minimax optimization separately, existing methods mainly focus on bilevel optimization with lower-level minimization problems, often under strong convexity assumptions, and are not directly applicable to the minimax lower-level setting considered here. To address this gap, we develop penalty-based first-order methods for bilevel minimax optimization without requiring strong convexity of the lower-level problem. In the deterministic setting, we establish that the proposed method finds a lifted $\epsilon$-KKT point with $\tilde{O}(\epsilon^{-4})$ oracle complexity. We further show that bilevel problems with convex constrained lower-level minimization can be reformulated as special cases of our framework via Lagrangian duality, leading to an $\tilde{O}(\epsilon^{-4})$ complexity bound that improves upon the existing $\tilde{O}(\epsilon^{-7})$ result when the lower-level Slater's condition holds. Finally, we extend our approach to the stochastic setting, where only stochastic gradient oracles are available, and prove that the proposed stochastic method finds a nearly lifted $\epsilon$-KKT point with $\tilde{O}(\epsilon^{-9})$ oracle complexity.
Read the full abstract

All authors Yiyang Shen, Yutian He, Weiran Wang, Qihang Lin

Operator Algebras

Functoriality of the KSGNS Construction for Intertwiners of Strict Positive $C^*$-Correspondences

We prove that the KSGNS construction can be viewed as an endofunctor on a category whose objects are strict positive $C^*$-correspondences from a fixed $C^*$-algebra and morphisms are given by intertwiners which account for automorphisms of the fixed $C^*$-algebra. Using this perspective, we provide a functorial perspective for strict positive equivariant $C^*$-correspondences of $C^*$-dynamical systems and show every strict positive equivariant $C^*$-correspondence of $C^*$-dynamical systems unitarily uniquely dilates under the KSGNS construction to an equivariant $C^*$-correspondence of the dynamical systems.
Read the full abstract

Statistics Theory · Cross-listed from stat.ME

Cellwise and Casewise Robust Multivariate Regression with Inference

Multivariate linear regression is a fundamental statistical task, but classical estimators such as ordinary least squares are highly sensitive to outliers. These may occur as casewise outliers that affect entire observations, or as outlying cells, that are individual contaminated entries in the predictor and/or response matrix. Moreover, modern datasets frequently contain missing values and are high-dimensional. To address these challenges we propose the cellwise multivariate regression (cellMR) estimator, a robust regression method that simultaneously accommodates casewise and cellwise outliers, missing data, and high dimensionality. The approach builds on a cellwise robust covariance estimator and uses ridge regularization for numerical stability. We further introduce cellBoot, a novel bootstrap-based inference procedure tailored to the cellMR framework. Relying on indirect inference, cellBoot provides asymptotically valid confidence intervals that are robust to casewise and cellwise contamination. We derive influence functions and breakdown bounds of the regression estimator and prove the asymptotic validity of the cellBoot confidence intervals. Simulations and a real genomics application illustrate the strong finite-sample performance of the proposed methods.
Read the full abstract

Statistics Theory

Scale selection for geometric medians on product manifolds

Geometric medians on product manifolds are sensitive to the relative scaling of factor metrics because the median objective couples the factors rather than separating them. We study this scale-selection problem and first prove that naive joint minimization over location and scale is degenerate: the scale is driven to the boundary and the problem collapses to a marginal median, effectively discarding one factor. Thus relative scale is not identifiable from the raw median loss alone. We develop three alternatives to mitigate this issue. The first treats scale as indexing a sensitivity path and establishes uniform consistency, a functional central limit theorem, and a derivative-based sensitivity measure. The second constructs a robust scale-calibrated median using marginal radial median scales, yielding unit invariance, consistency, a two-step central limit theorem, and bounded influence. The third introduces a bounded balance equation for direct scale estimation, with monotonicity, uniqueness, joint asymptotic normality, and bounded influence. Simulations illustrate boundary collapse, sensitivity, unit invariance, and balanced estimation in Euclidean and Bures-Wasserstein settings.
Read the full abstract

Information Theory

Semantic Smoothing for Language Models via Distribution Estimation and Embeddings

We propose semantic smoothing, a smoothing method for language models that uses embeddings to share statistical observations across semantically similar contexts. The starting point is a decomposition of log-perplexity that motivates smoothing as a collection of distribution-estimation problems under Kullback-Leibler (KL) loss. We then show that, under a Lipschitz-logit model for embedding-based language generation, proximity of context embeddings implies proximity of the corresponding next-word distributions in KL divergence. Combining these observations, we formulate semantic smoothing as distribution estimation in KL loss with KL-proximity side information. For $n$ samples on a $d$-symbol alphabet with a side-information distribution at KL distance $\Delta$, we give an interpolation estimator with worst-case KL risk $O(\min\{\Delta,d/n\})$, and prove a matching-order lower bound for uniform side information. We extend the estimator to multiple and empirically estimated synonymous distributions. Experiments on synthetic Markov data and WikiText-103 bigram models using Word2Vec, GloVe, and GPT-2 embeddings show that semantic smoothing consistently reduces test perplexity when applied to add-constant and Kneser-Ney estimates.
Read the full abstract

All authors Haricharan Balasundaram, Swathi Shree Narashiman, Pranay Mathur, Andrew Thangaraj

Algebraic Geometry

On Bands and Limit Theorems in Tropical Geometry

We review the basic theory of bands and band schemes introduced by Baker-Jin-Lorscheid, which is an algebraic framework for tropicalization, analytification, and $\mathbb{F}_1$-geometry. For an affine scheme $X$ over a non-Archimedean valued field $k$, one can associate to every affine embedding $\iota$ of $X$ a naturally defined affine band scheme $Y_\iota$ whose rational points over the tropical band $\mathbb{T}$ recover the tropicalization $Trop(X,\iota)$. We prove that $X$ is the limit of the $Y_\iota$ in the category of band schemes, thereby obtaining a scheme-theoretic enhancement of Payne's limit theorem. By taking $\mathbb{T}$-rational points, this recovers Payne's theorem for affine tropicalizations from the perspective of band scheme theory and the same method provides an analogous result in the real tropical setting.
Read the full abstract

Statistics Theory · Cross-listed from cs.LG

Susceptibilities and Patterning: A Primer on Linear Response in Bayesian Learning

These notes introduce the theory of susceptibilities as developed in [arXiv:2504.18274, arXiv:2601.12703] for interpreting neural networks. The susceptibility of an observable $\phi$ to a data perturbation is defined as a derivative of a posterior expectation, which by the fluctuation–dissipation theorem equals a posterior covariance. Different choices of $\phi$ yield different objects: per-sample losses give the influence matrix (the Bayesian influence function of [arXiv:2509.26544]), while component-localized observables give the structural susceptibility matrix that pairs model components with data patterns. The susceptibility matrix is (up to a factor of $n\beta$) the Jacobian of the map from data distributions to structural coordinates; its pseudo-inverse provides a linearized solution to the patterning problem of [arXiv:2601.13548]: finding data perturbations that produce a desired structural change. We motivate the theory from its statistical-mechanical foundations, then give a detailed exposition of susceptibilities, their empirical estimators, and their connection to the geometry of the loss landscape.
Read the full abstract

Algebraic Geometry

Tensor product surfaces and graded syzygies

Let $U\subseteq H^0(\mathcal{O}_{\mathbb{P}^1\times \mathbb{P}^1}(a,b))$ be a four-dimensional vector space and consider the rational map $\phi_U:\,\mathbb{P}^1\times \mathbb{P}^1 \dashrightarrow \mathbb{P}^3$ defined by its basis of bihomogeneous polynomials. The tensor product surface $X_U\subseteq \mathbb{P}^3$ is the closed image of $\phi_U$, and a fundamental problem in this setting is to determine its implicit equation. As these surfaces are ubiquitous within the field of geometric modeling and design, knowledge of their implicit equations is particularly advantageous, allowing for more effective and efficient computations. In this article, we expand upon work of Duarte-Schenck and work of the present author to solve this implicitization problem when the bigraded ideal $I_U$ admits a singly graded syzygy.
Read the full abstract

Statistics Theory

Linear Response Estimators for Singular Statistical Models

We define susceptibilities as a measure of the response of an observable quantity of a parameterized statistical model to a perturbation of the data for a general class of observables. We define estimators for these susceptibilities as statistics in a sequence of n data-points and prove that these estimators are consistent and asymptotically unbiased in the large n regime.
Read the full abstract

Information Theory · Cross-listed from cs.LG

When Diffusion Model Can Ignore Dimension: An Entropy-Based Theory

Diffusion models perform remarkably well on high-dimensional data such as images, often using only a modest number of reverse-time steps. Despite this practical success, existing convergence theory does not fully explain why such samplers remain efficient in high dimensions. Many prior KL guarantees bound the discretization error in terms of the ambient dimension, while other improved results replace this dependence using intrinsic-dimensional or geometric structure assumptions. In this work, we develop an alternative information-theoretic perspective on diffusion sampler convergence. We prove that, for Gaussian mixture targets, the discretization error is controlled by the Shannon entropy of the latent mixture component rather than by the ambient dimension. Consequently, the leading step complexity scales linearly with latent entropy and depends only logarithmically on the second moment of the data. Our analysis also extends to discrete target distributions, where the relevant complexity is the entropy of the target rather than the dimension of the embedding space. These results suggest that diffusion sampling can remain efficient in high-dimensional spaces when the data distribution admits a compact latent representation, as is widely believed to be the case for natural images.
Read the full abstract