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

Numerical Analysis

On a stochastic column-block bregman method for nonlinear systems

Sparse solution problems play an important role in both signal processing and image restoration. In this paper, we propose a stochastic column-block nonlinear Bregman method for efficiently computing sparse solutions to nonlinear systems. Under certain assumptions, we analyze the convergence of the proposed method and derive an upper bound for its convergence rate. Numerical experiments, including an image recovery problem, are presented to illustrate the efficiency of the proposed method.
Read the full abstract

All authors Wendi Bao, Naiyu Jiang, Lili Xing, Weiguo Li

Logic · Cross-listed from cs.SC

Relating the Computational and Logical Difficulty of Solving ODEs: From Polynomial to Discontinuous Right-Hand Sides

When a computer algebra system fails to solve an Ordinary Differential Equation, is this a limitation of its implementation, or a genuine computational barrier? Three traditions bear on the question. Modern computer algebra algorithms can be extremely efficient: Newton-type methods solve polynomial ODEs over $\mathbb{Q}[[X]]$ in quasi-linear time. Analog models of computation has shown that polynomial ODEs and Turing machines are two presentations of the same phenomenon, with solution length acting as time and precision as space. Computable analysis shows that ODEs can be intrinsically hard – undecidable, even $\mathsf{PSPACE}$-complete, over compact domains. Comparing these traditions is natural and necessary, yet such comparisons routinely reduce to comparisons of encodings rather than of underlying algorithmic content. We argue that reverse mathematics provides a representation-invariant lens in which algorithmic content is compared directly. We prove that every level of the Big Five hierarchy is inhabited by a natural statement from classical ODE theory, as an exact equivalence: the regularity of $f$ is an intrinsic algorithmic invariant placing the initial value problem $y'(t)=f(t,y(t))$, $y(t_0)=y_0$, into one of several computational strata, ranging from polynomial-time solvability to transfinite computation. The resulting stratification acts as a practical diagnostic common to the three traditions. By abstracting from representation, it separates fundamental barriers from the technical shortcomings of symbolic solvers, the artefacts of analog encodings, and the effectivity constraints of computable analysis, identifying the intrinsic parameters (length bounds, radii of convergence, moduli of continuity) under which feasibility is restored.
Read the full abstract

Numerical Analysis · Cross-listed from cs.LG

Monotone Neural Policy Iteration for High-Dimensional First-Order Hamilton–Jacobi–Bellman Equations

We analyze a neural semi-discrete method for high-dimensional first-order Hamilton-Jacobi-Bellman (HJB) equations with known or learned dynamics. Centered differences and an artificial viscosity $Nh=O(h)$ define a monotone operator evaluated through $2d+1$ shifted network queries; policy iteration solves the resulting Bellman equation without a tensor grid. At fixed $h$, the sharp componentwise condition $\max_i|f_i|\le2N$ turns every frozen-policy operator into a nearest-neighbor Markov-chain generator with a policy-independent total jump rate. Uniformization gives whole-space well-posedness for measurable feedbacks, an explicit Poisson-tail bound on the numerical domain of dependence, and boundary-free localization. The representation also yields a posteriori policy-evaluation bounds that account for residual and learned-model errors. A greedy-gap analysis controls inexact policy iteration at fixed $h$; a separate consistency estimate connects the semi-discrete equation to the continuous HJB equation. Experiments reproduce the extremal tail, show rates consistent with $O(\sqrt h)$ and nearly $h$-independent exact-policy-iteration decay, and assess empirical estimator effectivity. A nonsmooth example shows that the continuous residual can miss a non-viscosity solution, whereas the shifted residual detects the defect. Further tests provide a structured interval-verified certificate calibration, an early-budget benefit of policy freezing for bang-bang control, and learned-dynamics diagnostics. A structured nonlinear problem with active compact-control constraints is tested against a manufactured semi-discrete reference through $d=1024$.
Read the full abstract

All authors Minseok Kim, Yeongjong Kim, Namkyeong Cho, Yeoneung Kim

Optimization and Control · Cross-listed from cs.LG

Solving Max-Cut to Global Optimality via Feasibility-Preserving Graph Neural Networks

Exact solution of hard combinatorial optimization problems often relies on strong convex relaxations, but solving these relaxations repeatedly inside a branch-and-bound algorithm can be prohibitively expensive. Hence, we consider this challenge for Max-Cut, where branch and bound commonly uses semidefinite programming (SDP) relaxations to bound subproblems. We propose a Max-Cut-specific graph neural network that serves as a principled, lightweight neural proxy for these SDP solvers and can be plugged directly into an exact branch-and-bound framework. The proposed architecture has update steps of complexity $\mathcal{O}(n^2 + ne)$, and predicts both primal- and dual-feasible SDP solutions. The primal SDP solutions yield feasible Max-Cut solutions via the Goemans–Williamson algorithm. In addition, it is trained in a self-supervised fashion without requiring solved SDP relaxations as labels. Empirically, we show that our architecture can substantially reduce the cost of bounding in exact Max-Cut solving by up to $10.6 \times$ compared with using the state-of-the-art SDP solver Mosek. Our work highlights the potential of learned, validity-preserving surrogates for accelerating exact optimization over structured convex relaxations.
Read the full abstract

All authors Hao Chen, Chendi Qian, Christopher Morris, Andrea Lodi, Can Li

Information Theory

Sub-Gaussian Concentration and Entropic Normality of the Maximum Likelihood Estimator

It is well known that, under standard regularity conditions, the maximum likelihood estimator (MLE) satisfies a central limit theorem and converges in distribution to a Gaussian random variable as the sample size grows. This paper strengthens this classical result by developing several stronger forms of asymptotic normality for the normalized MLE. With additional assumptions on the score, we first establish sub-Gaussian tail bounds and convergence of all moments for the normalized estimation error. We then prove an entropic central limit theorem for a smoothed version of the estimator, showing convergence in relative entropy to the limiting Gaussian law. When the Fisher information of the normalized estimate is bounded, or its density has bounded first derivative, we further show that the smoothing can be removed, yielding entropic normality of the MLE itself. The proofs develop auxiliary tools that may be of independent interest, including exponential consistency bounds, high-moment estimates, and entropy-control arguments for the estimator.
Read the full abstract

Information Theory · Cross-listed from cs.LG

Theoretical Limits of Language Model Alignment

Language model (LM) alignment improves model outputs to reflect human preferences while preserving the capabilities of the base model. The most common alignment approaches are (i) reinforcement learning, which maximizes the expected reward under a KL-divergence constraint, and (ii) best-of-$N$ alignment, which selects the highest-reward output among $N$ independent samples. Despite their widespread use, the fundamental limits of reward improvement under a KL budget remain poorly understood. We characterize the information-theoretic limits of KL-regularized alignment by deriving the maximum achievable expected reward gain for a fixed KL-divergence budget. Our first result provides a closed-form expression for the optimal reward improvement, governed by a Jeffreys divergence term rather than the $\sqrt{\texttt{KL}}$ used in prior analyses. We further reformulate this expression as a covariance under the base model, yielding a practical estimator that predicts achievable alignment gains from base model samples alone. We extend our analysis to the proxy reward setting, showing that the gap between ideal and proxy alignment (reward hacking) grows with the magnitude of reward error and when the KL penalty factor decreases. We then prove that reward ensembling mitigates reward hacking, providing a theoretical justification for this technique used in practice. Empirically, we compute the KL-reward Pareto frontier for two tasks for LMs, safety and summarization, and show that best-of-$N$ closely approaches the theoretical limit, while PPO and GRPO remain substantially suboptimal. Our theoretical results shed light on several empirically observed phenomena in the alignment literature and suggest that algorithmic improvements are needed to achieve optimal alignment without high inference costs.
Read the full abstract

All authors Lucas Monteiro Paes, Natalie Mackraz, Barry-John Theobald, Federico Danieli

Optimization and Control · Cross-listed from cs.LG

Almost Sure Convergence Rates of Stochastic Approximation and Reinforcement Learning via a Poisson-Moreau Drift

Establishing almost sure convergence rates for stochastic approximation and reinforcement learning under Markovian noise is a fundamental theoretical challenge. We make progress towards this challenge for a class of stochastic approximation algorithms whose expected updates are contractive, a setting that arises in many reinforcement learning algorithms such as $Q$-learning and linear temporal difference learning. Specifically, for a power-law learning rate $O(n^{-\eta})$ with $\eta \in (1/2, 1)$, we obtain an almost sure convergence rate arbitrarily close to $o(n^{1 - 2\eta})$. For the harmonic learning rate $\alpha_n=\alpha/(n+1)$, the almost sure squared-error rate is $o(n^{-\zeta})$ for every $\zeta<\min\{1,2(1-\kappa)\alpha\}$, where $\kappa$ is the contraction factor. Thus exponents arbitrarily close to one are available when $2(1-\kappa)\alpha\geq1$. This approaches the $n^{-1}$ polynomial exponent of classical law-of-the-iterated-logarithm benchmarks. Key to our analysis is a novel Lyapunov drift construction that applies a Poisson-equation based correction for Markovian noise to the well-established Moreau-envelope smoothing for the contractive mapping.
Read the full abstract

Logic · Cross-listed from stat.ML

Every Feedforward Neural Network Definable in an o-Minimal Structure Has Finite Sample Complexity

We show that, in a precise sense, a broad class of feedforward neural networks learn (have finite sample complexity) in the PAC model: every fixed finite feedforward architecture whose layers are definable in an o-minimal structure has finite sample complexity in the agnostic PAC setting, even with unbounded parameters. This covers standard fixed-size MLPs, CNNs, GNNs, and transformers with fixed sequence length, together with the operations and layers typically used in such architectures, including linear projections, residual connections, attention mechanisms, pooling layers, normalization layers, and admissible positional encodings. Hence, distribution-free learnability for modern non-recurrent architectures is not an exceptional property of particular activations or architecture-specific VC arguments, but a consequence of tame feedforward computation. Our results reposition finite-sample PAC learnability as a baseline rather than a differentiator: they shift the focus of architectural comparison toward inductive biases, symmetries and geometric priors, scalability, and optimization behaviour.
Read the full abstract

All authors Anastasis Kratsios, Gregory Cousins, Haitz Sáez de Ocáriz Borde, Bum Jun Kim, Simone Brugiapaglia

Optimization and Control

Cross-validated feature selection using least-squares support vector machines for binary classification

Support vector machines (SVMs) are widely used for binary classification. However, irrelevant or redundant features can degrade classification accuracy and model interpretability, making feature selection important for SVMs. Most SVM feature selection methods rely on sequential search or regularization and do not directly identify the best feature subset. Mixed-integer optimization (MIO) is a direct approach to best subset selection. Cross-validation (CV) is broadly applicable to feature selection because it evaluates predictive performance on validation data without model assumptions. Although Purcaru et al. (2026) proposed an MIO formulation for CV-based linear-SVM feature selection (SVM-CV), this method suffers from high computational cost. We develop computationally efficient CV-based MIO formulations for feature selection in linear SVM classification by using least squares support vector machines (LS-SVMs). We formulate the problem as a bilevel MIO problem, where the lower-level problem trains LS-SVM classifiers and the upper-level problem selects features based on validation sets. We then reformulate it as a single-level MIO problem by exploiting the optimality conditions of LS-SVM. We propose two formulations: a hinge-loss-based validation model and a least-squares-based validation model. We compare our methods with SVM-CV and other SVM-based and logistic-regression-based methods through simulation experiments on synthetic data. The hinge-loss-based validation model attained high feature selection accuracy under the high-noise and moderate-noise conditions, particularly with more candidate features, and the least-squares-based validation model performed well in feature selection under the low-noise condition at the lower correlation level. Furthermore, our two models achieved substantially faster computation than the existing SVM-CV model.
Read the full abstract

All authors Masaharu Mori, Shunnosuke Ikeda, Ryuta Tamura, Yuichi Takano, Ryuhei Miyashiro

Combinatorics

Topological Zeta Functions of Matroids: Operations and Computations

The topological zeta function of a matroid is a rational function as well as a valuative invariant of the matroid, encoding rich combinatorial information. We analyze topological zeta functions of matroids from the vantage point of several matroid operations and operations on lattices of flats. We prove a clean recurrence relation for the Möbius inversion and use it to describe the topological zeta function of the truncation and free extension of a matroid in relation with that of the original matroid. We also characterize the Taylor coefficients of the topological zeta functions for matroids in terms of a matroid invariant, which we call the girth, and generalize an earlier result by Jensen-Kutler-Usatine.
Read the full abstract

Probability

Modified ruin probability for a Cramér-Lundberg model driven by a compound mixed Poisson process

We study modified ruin probabilities in a Cramér-Lundberg model driven by a compound mixed Poisson process. In the heavy-tailed regime, if the integrated claim-size distribution is subexponential and the upper endpoint of the mixing distribution stays below the net-profit boundary, the modified and classical ruin probabilities are asymptotically equivalent. In the light-tailed regime, we prove a fixed-intensity ratio theorem and obtain both an endpoint-atom result and a sharp endpoint-density asymptotic with an explicit constant.
Read the full abstract

Numerical Analysis

Proximal Galerkin for the isometry constraint

We resolve a longstanding open problem in the computational modeling of nonlinear plates by introducing a numerical method that exactly enforces the isometry constraint, namely, that the first fundamental form of the mid-surface coincides with the identity tensor. Several numerical methods have been proposed to approximate solutions of such manifold-constrained variational problems using gradient flows with tangent space updates. However, this class of methods presents two main challenges. First, a preprocessing step is required to enforce the boundary conditions and generate an initial guess sufficiently close to an isometry. Second, each step of the gradient flow typically increases the isometry defect. We adopt an alternative approach based on the proximal Galerkin framework, originally introduced for variational problems with convex inequality constraints. The resulting method preserves the geometric structure of the feasible set and yields an efficient algorithm in which each iterate is an exact isometry at the barycenter of every mesh cell. In contrast to existing methods, no preprocessing step is required, enabling broader applicability of this important category of mathematical models. Numerical experiments on standard benchmarks demonstrate that the method converges to a prescribed error tolerance in an asymptotically mesh-independent number of iterations and requires substantially fewer iterations than previous methods, even on coarse meshes.
Read the full abstract

Thursday 7 May

General Mathematics

Bayesian fuzzy optimization using fuzzy Gaussian process

In many real-life problems, decision-making gets complicated due to dual sources of uncertainty, known as randomness and fuzziness or imprecision, which can be challenging for traditional optimization methods. Most of the existing fuzzy optimization techniques that optimize fuzzy-valued objective functions ignore fuzziness, while the probabilistic optimization techniques ignore randomness. To handle this dual source of uncertainty, Kwakernaak introduced the concept of a fuzzy random variable as “random variables whose values are not real, but fuzzy numbers". This work aims to derive a theoretical background for the Gaussian fuzzy process and fuzzy acquisition functions, which will be used to develop a novel Bayesian fuzzy optimization (BFO) technique that optimizes a fuzzy-values objective function. Based on fuzzy random variables, the Gaussian fuzzy process is developed, which is used as a prior belief about the fuzzy-valued objective function in the BFO. Fuzzy acquisition functions are defined to act as a guide for the search process of BFO with the help of posterior fuzzy mean and fuzzy variance. The proposed method demonstrated effective performance in both fuzzy mean-variance portfolio allocation and Indian temperature data analysis, showing robust predictive accuracy and adaptability. The proposed method can have broader applications in various fields like healthcare, material science, agriculture, etc.
Read the full abstract

All authors Sourav Das, Debdas Ghosh, Debjani Chakraborty, Pabitra Mitra

Differential Geometry

Hineva Inequality for Submanifolds of Real Space Forms with Semi-Symmetric Non-Metric Connection

In this paper, we establish the Hineva inequality for submanifolds of a real space form endowed with a semi-symmetric non-metric connection. We derive a sharp lower bound for the Ricci curvature of the submanifold in terms of the mean curvature vector and the squared norm of the second fundamental form. We apply this inequality to derive the Hineva inequality for several classes of submanifolds.
Read the full abstract

Representation Theory

A problem on Hecke algebras for $\mathrm{GL}_n(F)$ for $n>2$ over $p$-adic field $F$

We study the Hecke algebra $\mathcal{H}_G(F)$ for $G = \mathrm{GL}_n$ and $n>2$ where $F$ is a non-Archimedean local field of characteristic zero. We show that for $G = \mathrm{GL}_n$ and $n>2$ and any two such fields $E$ and $F$, there is a Morita equivalence $\mathcal{H}_G(E) \sim \mathcal{H}_G(F)$, by using the Bernstein decomposition of the Hecke algebra and by determining the intertwining algebras that yield the Bernstein blocks up to Morita equivalence.
Read the full abstract

Analysis of PDEs

Corrigendum to "Ramification of multiple eigenvalues for the Dirichlet-Laplacian in perforated domains" published in Journal of Functional Analysis 283 (2022) 109718

We fix the proof of Theorem 1.17 in the quoted reference, which does not in general gives the correct coefficients in the asymptotic behavior of eigenvalues for the Dirichlet Laplacian when a small hole is removed from the domain.

History and Overview of Mathematics

Besant quadrilaterals

We solve the following problem of W.H. Besant using a formula for the coefficients of an ellipse inscribed in a quadrilateral, $Q$: \enquote{If an ellipse be inscribed in a quadrilateral so that one focus is equidistant from the four vertices(call that point $EP$), the other focus must be at the intersection of the diagonals(call that point $IP$).} We also prove somewhat more than just solving Besant's problem itself, though it would be nice to see the details of the geometric approach proposed by Besant. More precisely, we also prove the converse result and additional results when $Q$ is a trapezoid. Finally, we show that such an inscribed ellipse exists if and only if $Q$ is orthodiagonal.
Read the full abstract

Optimization and Control

Unbiased Gradients for a Class of Conditional Stochastic Optimization Problems

In this paper we consider the conditional stochastic optimization (CSO) problem. This consists of optimizing a function which can be written as the expectation of a function which is itself a function of a conditional expectation, i.e.~of the type $F(\xi) := \mathbb{E}\left[f\left(Z,\mathbb{E}[g(Z,X,\xi)|Z]\right)\right]$, where precise definitions are given in the main text. We address a particular class of CSO problems where the joint law of the random variables $X,Z$ cannot be exactly sampled; this case has been addressed in Goda & Kitade (2023). We introduce a method that combines Markovian stochastic approximation with unbiased approximation methods which allows one to find the optimizer of $F(\xi)$ in the context of interest. We illustrate our methodology on two examples associated to parameter estimation with model averaging and portfolio selection associated to high-dimensional full factor multivariate stochastic volatility models.
Read the full abstract

Combinatorics

The partition dimension and $k$-domination number of a family of non-distance regular graph

A partition $\Sigma = \{S_1, S_2, \dots, S_k\}$ of the vertex set $V(G)$ is a resolving partition if every pair of distinct vertices in $G$ has a unique representation relative to $\Sigma$. The partition dimension, $pd(G)$, is the minimum cardinality of such a partition. Additionally, a subset $D \subseteq V(G)$ is a $k$-dominating set if every vertex in $V(G) \setminus D$ has at least $k$ neighbors in $D$; the $k$-domination number, $\gamma_k(G)$, denotes the minimum size of such a set. Determining these parameters is NP-complete and particularly challenging for non-distance-regular graphs. This paper consider the Toeplitz graph $T_{2n}(W)$, a family of non-distance-regular graphs. While some resolving parameters for this family have been established, its partition dimension and $k$-domination number remain unknown. We close this gap by computing both parameters for $T_{2n}(W)$.
Read the full abstract

Information Theory · Cross-listed from stat.ML

Decentralized Conformal Novelty Detection via Quantized Model Exchange

This work studies decentralized novelty detection with global false discovery rate (FDR) control across heterogeneous composite null distributions, without sharing the raw data due to privacy and bandwidth considerations. We propose a framework based on the exchange of quantized surrogate models, allowing independent agents to share low-precision representations of locally learned non-conformity score functions. We prove that evaluating data against these quantized composite scores preserves conditional exchangeability, providing rigorous finite-sample guarantees for global FDR control. Empirical studies on synthetic datasets confirm our theoretical results, demonstrating that the proposed approach maintains competitive statistical power while drastically reducing the communication cost.
Read the full abstract

Probability · Cross-listed from quant-ph

Multiplayer parallel repetition without dependency-breaking and anchoring variables: monotonic, concave amplification

We obtain quantitative estimates on the decay of the multiplayer optimal value under parallel repetition. In comparison to a previous work of the author in 2025 (arXiv: 2508.09380) which sought to generalize dependency-breaking and anchoring variables from two-player Quantum games, being able to establish quantitative estimates on the decay of the optimal value of a multiplayer game under parallel repetition is of interest to establish under different assumptions. Specifically, independently of the dependency-breaking and anchoring variables that have previously been employed to remove correlations from entangled information shared between Alice and Bob (hence removing dependencies), monotonic concave functions can be used in place of such variables to obtain rates of decay on the optimal value. The game-theoretic setting with two players was first analyzed with monotonic concave functions by Lanzenberger and Maurer. For $q_i , x_i > 0$ $\forall 1 \leq i \leq N$ where $N > 0 $ is the total number of players we adddress an open question raised in their work regarding potential generalizations of two-player monotonic concave functions, through amplification functions of the form $\Psi_{\textit{Mult}} \equiv \Psi = N - \underset{1 \leq i \leq N}{\prod} \mathrm{exp} \big[ - q_i x_i \big]$, which in the multiplayer game-theoretic setting have more intricate combinatorial structures.
Read the full abstract

2605.08259 · 7 May · Template (51 pages). Video summary available at: https://youtu.be/9BYgP_0N0Yc?si=XYwrVFEeaA6mDmEM. Related topics at: https://youtube.com/playlist?list=PL3rTBtU0TK_C05u3pbB8fOLI2XFwWPorX&si=nYEtKTgjAt7NiKS0; https://youtube.com/playlist?list=PL3rTBtU0TK_DEFrUY_h-ZX4KLBKcuaF92&si=RWqdJWFmi6qGNnL_; https://youtube.com/playlist?list=PL3rTBtU0TK_CvLJPBvmvNOQHdoDxey0oW&si=Jgs4H6RnArmotDes

Number Theory

Frequency Ordered Ratio Families Arising from the Factorization of $p_{m-1}+1$

We investigate a ratio sequence derived from the factorization of $p_{m-1} + 1$, where $p_n$ denotes the $n$th prime. For each $m \geq 3$, write $p_{m-1} + 1 = L_m R_m$ with $L_m$ the largest prime factor. Restricting to those $m$ for which $L_m > m$ (equivalently, $m \in \text{A223881}$), we obtain a multiset of values $R_m$. Since $p_{m-1}+1$ is even and $L_m > 3$ is odd, all values of $R_m$ are strictly even. Sorting the distinct $R_m$ by decreasing frequency yields a new sequence beginning $2, 6, 4, 8, 10, 12, 14, 16 \dots$. This article explains how this construction arises naturally from the structure of A223881, why the “family” phenomenon appears in plots of $p_{m-1} + 1$, and how the frequency ordering of $R_m$ captures the dominant families. Additionally, we propose a heuristic asymptotic model explaining the observed frequency ordering via classical results on primes in arithmetic progressions and support the model with numerical log-log analysis.
Read the full abstract

Differential Geometry

Geometry of tangent bundles of statistical manifolds equiped with Cheeger-Gromoll type metrics

In this paper, we investigate the geometry of the tangent bundle $TM$ of a statistical manifold $(M,g,\nabla)$ endowed with a two-parameter family of generalized Cheeger–Gromoll metrics $g_{p,q}$. We compute the associated the Levi–Civita connection $\nabla^{p,q}$ and express its curvature in terms of the Riemannian curvature and the skewness tensor $K$ of the base statistical manifold. We further analyze the behavior of geodesics, identify conditions under which the fibers of $TM$ are totally geodesic, and determine when the geodesic flow associated with $g_{p,q}$ is incompressible. Moreover, we establish necessary and sufficient conditions for the tangent bundle to admit constant sectional curvature. Several examples are provided to illustrate the theory, including statistically deformed Euclidean spaces and information geometric models such as the manifold of normal distributions. The sectional curvature of $(TM, g_{p,q})$ is computed for horizontal, vertical, and mixed directions, leading to a concise expression for the corresponding scalar curvature.
Read the full abstract

Probability

Universality of the fluctuations of the free energy in generalized Sherrington-Kirkpatrick models and the log likelihood ratio in spiked Wigner models

We consider the fluctuations of the free energy in generalized Sherrington-Kirkpatrick models and the log likelihood ratio of spiked Wigner models in the high temperature/subcritical regime. We prove that the limiting laws of the fluctuations are Gaussian under suitable assumptions, and the result is universal in the sense that it does not depend on the distribution of the disorder or the prior except that the means and the variances of the limiting laws depend on a few parameters of the model. The proof is based on the multigraph expansion that provides a unified approach to analyze both models.
Read the full abstract

Probability

Brownian-time Change of measure

We prove a fundamental change of measure theorem for the Brownian-time Brownian motion and its associated Brownian-time processes class introduced by Allouba and Zheng in 2001. This result, together with Allouba's prior work on (1) Brownian-time processes and their PDEs/SPDEs links and on (2) change of measure for SPDEs, is a critical building block in analyzing the behaviors of SDEs and SPDEs – of different types and orders – driven by Brownian-time noises and their relatives.
Read the full abstract

Analysis of PDEs

The Free Boundary in a Higher-Dimensional Long-Range Segregation Model

We consider a system of elliptic equations, depending on a small parameter $\eps$, that models long-range segregation of populations. The diffusion is governed by the Laplacian. This system was previously investigated by Caffarelli, Patrizi, and Quitalo in \cite{CL2} as a model in population dynamics, and they established the regularity of the free boundary in two dimensions. In this paper we study the free boundary in the higher dimensional case. We extend the concept of angles and asymptotic cones to higher dimensions, and give a characterization of regular and singular points in terms of their densities and angles. We obtain a structure result of the free boundary and show that, if the angles at the singular points are away from $\frac{n\omega_n}{2}$, the regular set is open in the free boundary and locally a $C^1$ manifold of dimension $n-1$. We also show that, if the supports of the populations are convex, they are convex polytopes. A weak form of the equality of angles for the convex configuration is also derived.
Read the full abstract

General Topology

Nielsen coincidence theory of $(n,m)$-valued pairs of maps

We consider pairs of maps $(f,g)$, where $f$ is an $n$-valued map and $g$ is an $m$-valued map, defined on connected finite polyhedra. A point $x$ such that $f(x)\cap g(x)\neq \emptyset$ is called a coincidence point of $f$ and $g$. A useful device for studying coincidence points would be a Nielsen-type invariant which provides a lower bound for the number of coincidence points of all $(n, m)$-valued pairs of maps homotopic to $(f,g)$. The construction of such an invariant $N(f:g)$ was proposed in [J. Fixed Point Theory Appl. 14, 309–324 (2013)]. Unfortunately, this approach has some flaws. In this paper, we present a modified construction that yields a corrected form of the invariant, defined in terms of the intersection points of the graphs of $f$ and $g$. In the case of $(n, m)$-valued pairs of maps of the circle our invariant provides a sharp lower bound, which we precisely determine.
Read the full abstract

Combinatorics

Improved Upper Bounds on the Pebbling Numbers of the Blanuša Snarks

The two Blanuša snarks $B_1$ and $B_2$ are 3-regular graphs on 18 vertices. Dantas, Lordelo, Niedermaier and Nogueira (Discrete Appl. Math. 361, 2025, pp. 336-346) established the first systematic bounds $23 \le \pi(B_i) \le 34$ for $i=1,2$. Bridi, Marquezino and Figueiredo (arXiv:2505.16050, 2025) then sharpened the upper side to $\pi(B_1) \le 31$ and $\pi(B_2) \le 30$ via a Weight Function Lemma heuristic. We push the upper bounds further to $\pi(B_1) \le 28$ and $\pi(B_2) \le 29$. The route is again Hurlbert's Weight Function Lemma, but applied one automorphism orbit at a time, with optimal weight functions coming from a linear program over a corpus of roughly $30{,}000$ rooted-subtree strategies per target. For the lower bound $\pi(B_i) \ge 23$ we re-derive the witnesses of Dantas et al. and re-verify them with two independent oracles: an exhaustive forward state-space search, and a sound-and-complete MILP encoding whose acyclicity constraint is motivated by the Milans-Clark No-Cycle Lemma. The interval for $B_1$ shrinks from $[23, 31]$ to $[23, 28]$, and for $B_2$ from $[23, 30]$ to $[23, 29]$.
Read the full abstract

2605.07014 · 7 May · v3: add missing citation to Bridi-Marquezino-Figueiredo (arXiv:2505.16050, 2025) and correct the framing: their Weight Function Lemma heuristic had already pushed the Blanusa upper bounds from 34 to 31 (B1) and 30 (B2); this paper's improvements 28 (B1) and 29 (B2) are over their bounds, not Dantas et al.'s

Algebraic Geometry

Weight of the De Rham-Betti Structures of Abelian Varieties

In this note, we prove that for any abelian variety defined over $\overline{\mathbb{Q}}$, its de Rham-Betti (dRB) group necessarily contains $\mathbb{G}_{m}$ as the group of homotheties. Consequently, this rules out the existence of non-zero dRB classes in odd-degree cohomology groups of abelian varieties over $\overline{\mathbb{Q}}$. This generalises results of the first part of arXiv:2511.01072.
Read the full abstract