Peer Revenue Play credits. Real papers.

Papers

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

Mathematics

math Watch Mathematics (sign in to use)

Sunday 26 April

Analysis of PDEs · Cross-listed from cs.CE

Partition-of-Unity Gaussian Kolmogorov-Arnold Networks

Gaussian basis functions provide an efficient and flexible alternative to spline activations in Kolmogorov-Arnold networks (KANs), but their performance is sensitive to the Gaussian scale parameter \epsilon. We introduce the partition-of-unity Gaussian KAN (PU-GKAN), a kernel-based normalization framework that transfers the Shepard partition-of-unity principle to trainable KAN edge functions. Each Gaussian edge basis is normalized by its local sum over fixed centers, preserving the standard edge-based architecture and essentially the same trainable coefficient structure as GKAN. The resulting feature map reproduces constants exactly at the edge level and admits an explicit normalized finite-feature kernel interpretation. Numerical experiments show that PU-GKAN reduces sensitivity to \epsilon, improves validation accuracy for most smooth and moderately non-smooth targets, and provides more stable training behavior. These benefits persist across variations in sample size and number of centers, higher-dimensional architectures, and Matern RBF bases. PU-GKAN also improves physics-informed neural networks for Helmholtz and wave equations, with validation-error reductions of up to 90%. These results establish partition-of-unity normalization as a mathematically grounded and broadly applicable framework for improving the reliability and accuracy of radial basis function KANs.
Read the full abstract

Functional Analysis

How to recognise extension domains

Let $\Omega \subset \mathbb{R}^n$ be a bounded domain and $1 < p < \infty$. We prove that there is a bounded extension operator $\dot{W}^{1,p}(\Omega)\to \dot{W}^{1,p}(\mathbb{R}^n)$ if and only if $\Omega$ satisfies the measure density condition and a Bourgain-Brezis-Mironescu type inequality (or limiting formula). As a key ingredient, we establish a fractional Poincaré-type inequality under the assumption of Ahlfors regularity alone, improving a result of Ponce (2004). We also prove that, under a mild Hausdorff measure condition on the boundary $\partial \Omega$, fractional extension (from $\dot{W}^{1,p}(\Omega)$ to $\dot{W}^{s,p}(\mathbb{R}^n)$) at a single exponent $s > 1/p$ self-improves to full first-order Sobolev extension (from $\dot{W}^{1,p}(\Omega)$ to $\dot{W}^{1,p}(\mathbb{R}^n)$). These results clarify the role of nonlocal estimates in the geometry of Sobolev extension domains.
Read the full abstract

Analysis of PDEs

Analysis and numerical simulations of a landfast ice model

In this manuscript, we consider a common modeling framework for Arctic landfast ice based on the work of Lemieux et al. [27], which is designed for use in large-scale climate models. This approach extends the classical viscous-plastic sea-ice model introduced by Hibler [18], which remains the most widely used model for simulating large-scale sea-ice dynamics. In particular, landfast ice refers to sea-ice that is attached to the coastline or grounded and therefore exhibits nearly vanishing motion. The model considered in this manuscript augments Hibler's viscous-plastic sea-ice model by tensile strength and a threshold-dependent basal stress. We present a rigorous analytical and numerical study of this landfast ice model. The analytical contributions are the local strong well-posedness, the global strong well-posedness in the absence of external forces and for initial data close to constant equilibrium solutions, and the existence of time-periodic solutions. The limiting case $k_{\mathrm{t}} \equiv 1$ is treated separately, since the pressure coupling then vanishes and a new smallness condition on the stress regularization parameter is required. For the system without scalar diffusion, we establish local strong well-posedness through a Lagrangian formulation and anisotropic maximal $\mathrm{L}^p$-regularity. Finally, we perform numerical simulations that illustrate key qualitative differences between landfast ice and classical viscous-plastic sea-ice models. The simulations reveal the formation of stationary equilibrium states characterized by vanishing ice velocity. These observations are consistent with the global-in-time existence result close to equilibria established in Theorem 4.1 and the time-periodic result in Theorem 5.2. The combined analytical and numerical results provide new insight into the structure, stability, and long-term behavior of landfast ice dynamics.
Read the full abstract

Information Theory

On the Minimum Distances of Some Families of BCH Codes

BCH codes form an important class of cyclic codes, which have applications in communication and data storage systems. Although the BCH bound provides a lower bound on the minimum distance of BCH codes, determining the true minimum distances of BCH codes is a very challenging problem. In this paper, we settle the minimum distances of a number of infinite families of narrow-sense BCH codes. By explicitly constructing the locator polynomials for minimum weight codewords, we obtain many families of primitive and non-primitive BCH codes with $d=\delta$, where $d$ is the minimum distance of a $q$-ary BCH code of length $n$, designed distance $\delta$, and offset $b$, denoted by $\mathbf{C}_{(q, n, \delta, b)}$. For primitive BCH codes, we obtain infinite families of BCH codes over $\mathbb{F}_3$ and $\mathbb{F}_4$ satisfying $d=\delta$, where $\delta \in \{5,6,7,8\}$. Moreover, we construct several infinite families of $q$-ary BCH codes with $d=\delta$, where $2 \le \delta \le q-1$. For $\delta=q^t+1$, we prove that the BCH code $\mathbf{C}_{(q, q^m-1, q^t+1, 1)}$ has $d=\delta$ for all $m$ satisfying $m \equiv 0 \pmod{pt}$, where $p$ denotes the characteristic of $\mathbb{F}_q$. In the paper by Ding et al., IEEE Trans. Inf. Theory 61(5): 2351-2356, it was conjectured that the minimum distance of $\mathbf{C}_{(q, q^m-1, q^t+1, 1)}$ is always equal to its Bose distance $d_B$. Our result confirms this conjecture for the case $m \equiv 0 \pmod{pt}$. For non-primitive BCH codes, we construct a family of BCH codes $\mathbf{C}_{(q,\frac{q^p-1}{\lambda},p+1,1)}$ with $d=\delta=p+1$, where $p$ is an odd prime, $q=p^e$ with $p \nmid e$ and $\lambda \mid q-1$.
Read the full abstract

All authors Yaqi Chen, Hao Chen, Cunsheng Ding, Huimin Lao

Algebraic Geometry

Exceptional loci of F-blowups and $G$-Hilbert schemes

We study exceptional loci of F-blowups of normal toric varieties. In the $\Q$-factorial case, this study amounts to studying the exceptional loci of $G$-Hilbert schemes. We give a formula for the dimension of the center of a prime divisor on the F-blowup in terms of combinatorial data, together with an algorithm for computing it. Moreover, we study the relation between F-blowups and essential divisors for three-dimensional terminal singularities and canonical singularities. Finally, we give a simple condition ensuring that a prime divisor over the given toric variety has a positive-dimensional center on the F-blowup.
Read the full abstract

Probability

Mean Waiting Times in Discrete-Time Priority Queues with Geometrically Distributed Idle Periods

This paper considers the mean waiting times in discrete-time preemptive-resume and nonpreemptive priority single-server queues fed by K independent batch Markovian arrival streams with geometrically distributed idle periods. While being active, the k-th (k = 1, 2, ..., K) arrival stream feeds at least one customer to the queue, where the number of arriving customers depends on the state of the underlying Markov chain. Service times of class $k$ customers are independent and identically distributed according to a general distribution. For these queues, we derive explicit formulae for the mean waiting times of customers in each class.
Read the full abstract

Optimization and Control · Cross-listed from econ.TH

Rigidity and default in production networks

This paper studies the transmission of productivity shocks in general equilibrium production networks, when firms in different sectors operate under informational rigidity and rely on external debt. Rigidity breaks the Modigliani-Miller irrelevance of leverage and may generate default following shocks, even in equilibrium. The economy consists of firms, banks, and consumers. Under proportional shock transmission, we prove that a unique Walrasian rigid equilibrium exists and provide explicit expressions for equilibrium quantities, prices, and interest rates. We show that, on the one hand, Hulten's theorem fails under rigidity, even without leverage. On the other hand, we prove that welfare is smaller than in the first best if and only if both leverage and rigidity exist. The latter increase the total cost of debt and have inflationary effects on the levered sectors, which propagate downstream, and shift consumption and labor upstream. The occurrence of default depends solely on real shocks and the network structure, while the magnitude of the losses depends also on the connectedness of the economy and the cost of debt of the connected sectors. We provide conditions for default cascades to occur and study two examples of default propagation.
Read the full abstract

All authors Giacomo Como, Fabio Fagnani, Elisa Luciano, Alessandro Milazzo, Marco Scarsini

Optimization and Control

Wireless Mobile Charging for Emergency Electric Vehicle Routing: A Mixed-Integer and Metaheuristic Framework for In-Motion Energy Transfer

As electric vehicles (EVs) become central to decarbonization efforts, the need for uninterrupted power supply in time-critical logistics, particularly in medical transportation, poses unique challenges for power systems integration. Conventional fixed or mobile charging infrastructure requires vehicle downtime, which makes them unsuitable for nonstop operations such as organ delivery. This work introduces the Wireless Mobile Charging Electric Vehicle Routing Problem, a novel framework in which mobile charging trucks wirelessly transfer energy to moving EVs via inductive coupling, eliminating the need for stationary charging stops. We formulate a mixed-integer programming model that co-optimizes routing and in-motion energy transfer between heterogeneous vehicle fleets under temporal and spatial alignment constraints. To address computational complexity, we develop a hybrid Bitmask Dynamic Programming and Large Neighborhood Search algorithm, tailored to arc-level charging decisions and dual-fleet synchronization. Experiments using real-world hospital logistics data from Singapore demonstrate significant runtime improvements and higher-quality solutions compared with commercial solvers. This study advances computational power system modeling by incorporating dynamic, motion-based energy delivery that is both time- and space-dependent. The results provide actionable insights into planning mobile energy infrastructure, enhancing demand-side flexibility, strengthening resilience in emergency logistics, and integrating wireless charging into urban power systems to support flexible and zero-carbon electrified mobility.
Read the full abstract

All authors Jingyi Zhao, Haoxiang Yang, Youxuan Pan, Yang Liu

Combinatorics

New infinite families of $q$-analogs of group divisible designs with arbitrary block dimension

This paper is mainly devoted to constructions of \(q\)-analogs of group divisible designs and their applications. We give a complete description of the action of \(G=\GL(m,q^l)\) on \(\Omega_k^{k-1}\), where $3\leq k\leq \min\left\lbrace m+1,l\right\rbrace $ and \(\Omega_k^{k-1}\) is the set of \(k\)-subspaces of $\GF(q)^{ml}$ whose \(\GF(q^l)\)-span has dimension \(k-1\). We do this by relating the \(G\)-orbits on \(\Omega_k^{k-1}\) to the corresponding Singer cycle orbits on subspaces of $\GF(q)^l$. From the properties of the $G$-incidence matrix between $2$-subspaces and $k$-subspaces, we obtain plenty of new infinite families of simple \(q\)-analogs of group divisible designs with arbitrary block dimension. We further establish a recursive construction for simple \(q\)-analogs of pairwise balanced designs and then produce new infinite families of such designs. We also obtain plenty of infinite families of non-simple subspace \(2\)-designs through the above two types of designs.
Read the full abstract

Representation Theory

Super-Chevalley Restriction and Relative Lie Algebra Cohomology over the 2|3 Algebra

Let $A:=\mathbb{C}[z_+,z_-]\otimes \Lambda(\theta_1,\theta_2,\theta_3)$, with $z_\pm$ even and $\theta_1,\theta_2,\theta_3$ odd. For a reductive Lie algebra $\mathfrak g$, let $\mathfrak g[A]:=\mathfrak g\otimes A$ be the corresponding current Lie superalgebra. Motivated by the Chang–Yin description of weak-coupling $1/16$-BPS cohomology in $\mathcal N=4$ super-Yang–Mills, we study the relative Lie algebra cohomology $H^\bullet(\mathfrak g[A],\mathfrak g;\mathbb{C})$. We isolate three finite-rank phenomena. First, the natural $3|2$ super-commuting restriction map, viewed as a super analogue of Chevalley restriction and its commuting-scheme variants, already fails to be an isomorphism for $\mathfrak g=\mathfrak{so}_7$; the obstruction is a non-Cartan class. Second, the same algebra produces explicit fortuitous classes for $\mathfrak{sl}_2$ and $\mathfrak{so}_7$, giving concrete counterexamples to naive stable-image expectations suggested by the type-A Loday–Quillen–Tsygan theorem and its current-algebra refinements. Third, the classical relative cohomologies for the Langlands-dual pair $(\mathfrak{so}_7,\mathfrak{sp}_6)$ are not isomorphic. We then record the conjectural quantum deformation of the differential expected to restore duality, together with first-order evidence pairing the fortuitous and non-Cartan $\mathfrak{so}_7$ classes.
Read the full abstract

Optimization and Control · Cross-listed from cs.CE

Unsupervised Learning for AC Optimal Power Flow with Fast Physics-Aware Layer

Learning to solve the Alternating Current Optimal Power Flow (AC-OPF) problem by neural networks (NNs) is a promising approach in real-time applications. Existing methods to ensure the physical feasibility of NN outputs embed a power flow (PF) solver within networks. However, the gradient through the PF solver, namely, implicit differentiation, needs manual Jacobian derivation and the solution of linear systems, which is computationally prohibitive and hinders integration with modern automatic differentiation (AD) frameworks. To address these challenges, we propose FPL-OPF, a novel unsupervised learning framework that incorporates a Fast Physics-aware Layer for AC-OPF problems. FPL-OPF embeds a fast PF iterative solver within the NN and takes solely the last few or even the final iterations into the AD graph. This design ensures high computational efficiency for both the forward and backward passes, circumventing complex custom backward implementations. Theoretically, we rigorously prove that the gradient from this design serves as a high-fidelity surrogate of the true implicit gradient under mild conditions. Extensive experiments demonstrate that FPL-OPF achieves significant speedups over state-of-the-art unsupervised learning approaches, while maintaining near-zero constraint violations and competitive optimality. Our code is available at https://github.com/wowotou1998/fpl-opf
Read the full abstract

All authors Jiebao Zhang, Haoyu Yan, Zhichao Sheng, Hongwen Yu, Shuang Ye, Haoyu Wang, Ye Shi

Mathematical Physics

Regularization of Divergent Power Sums via Fractional Extension of Differential Generators

We reconsider the problem of regularizing the divergent series $\sum_{n=1}^{\infty}n^{\alpha}$ for $\operatorname{Re}\alpha>-1$, and offer a regularization prescription that yields the Riemann zeta regularization as a special case. The development of the regularization is framed as a two-step problem. The first step is prescribing a regularization of the divergent sum $\sum_{n=1}^{\infty}n^m$ for every non-negative integer $m$; and the second step is the extension of the sum for non-integer $\alpha$. The extension is obtained under the consistency condition that the regularized sum for integer $m$ emerges continuously from the sum for non-integer $\alpha$. The scheme is specified by a differential generator $L=L(\mathrm{d}/\mathrm{d}t)$ through which a generalized spectral function (GSF), $K_L(t)$, is constructed. Under the condition that the GSF has a holomorphic complex extension $K_L(z)$ with $z=0$ as a pole, the case for integer $m$ takes the regularized value $\sum_{n=1}^{\infty} n^m = (2\pi i)^{-1}\oint_C L^m K_L(z) z^{-1}\mathrm{d}z$, where $C$ is a closed contour enclosing only the pole of $K_L(z)$ at the origin. On the other hand, under the consistency condition, the case for non-integer $\alpha$ takes the value $\sum_{n=1}^{\infty}n^{\alpha}=(2\pi i)^{-1}\int_{\tilde{C}} L^{\alpha} K_L(z) z^{-1}\mathrm{d}z$, where $L^{\alpha}$ is the fractional extension of $L^m$ and $\tilde{C}$ is an appropriate deformation of the contour $C$. Here, we obtain the regularization corresponding to the generator $L=h(t) \mathrm{d}/\mathrm{d}t$, with $h(t)$ positive for all $t>0$, monotonically non-increasing, and admitting complex extension $h(z)$ such that $1/h(z)$ is entire. We find that the regularized sum is equal to the Riemann zeta regularized value plus terms determined by the generator $L$.
Read the full abstract

Number Theory

Picking up the partial sums of the Möbius function problem with probabilistic number theory

We revisit several hybrid multiplicative-to-additive type functions from a recent preprint article. These functions, $g(n)$ with Dirichlet generating function (DGF) $\zeta(s)^{-1} (1+P(s))^{-1}$ for $\Re(s) > 1$ where $P(s) = \sum_p p^{-s}$ is the prime zeta function, $|g(n)| = \lambda(n) g(n)$ with DGF $\zeta(2s)^{-1}(1-P(s))^{-1}$, and $C_{\Omega}(n)$ with DGF $(1-P(s))^{-1}$. Each of these function variants are defined in terms of the additive (respectively, strongly additive) functions $\omega(n)$ and $\Omega(n)$. These two auxiliary functions are used in the prior manuscript to relate partial sums of the classical Möbius function, $\mu(n)$, to signed partial sums involving the prime counting function, $\pi(x)$, and the Liouville lambda function, $\lambda(n) := (-1)^{\Omega(n)}$. In this article, we explore summing the identities from the first manuscript using several probabilistic assumptions about the independence of the values of $\Omega(n)$ and $\mu^2(n)$ for $n \leq x$ at large $x$. We recover proofs of the limiting asymptotic growth of $|M(x)| / \sqrt{x}$ whose hypotheses promise to be substantially more attainable to make rigorous than past results from other authors relying on the Riemann Hypothesis or assumption of the linear independence of the simple, non-trivial zeros of $\zeta(s)$.
Read the full abstract

Statistics Theory

When Does Dynamic Preconditioning Preserve the Polyak-Ruppert CLT? A Stabilization Threshold

Polyak-Ruppert averaging yields an asymptotically normal estimator with sandwich covariance $H^{-1}SH^{-1}$, the foundation of online inference. When the gradient step is preconditioned by a data-driven matrix $P_t$, we ask how fast $P_t$ must stabilize for the central limit theorem (CLT) to remain valid. We resolve this via an exact preconditioner-isolating decomposition of the averaged error that confines $P_t$ to a dynamic remainder $R_n$, leaving the martingale and Taylor terms preconditioner-free. Let $M_t = (P_t H)^{-1}$ denote the effective inverse drift matrix, with $\|M_t - M_{t-1}\|_{\mathrm{op}} \lesssim t^{-\beta}$ and step-size exponent $\alpha \in (1/2, 1)$. We identify a stabilization-rate threshold $\beta > (\alpha+1)/2$ and prove that, within the class of polynomial rate hypotheses used in our upper bound, it cannot be weakened: the dynamic remainder $\sqrt{n}\,R_n$ vanishes in $L^2$ whenever $\beta > (\alpha+1)/2$, and we exhibit sequences satisfying those hypotheses for which it does not vanish when $\beta \le (\alpha+1)/2$. A single stabilization argument certifies three SA variants - SA-AdaGrad, SA-RMSProp, and SA-ONS - with gain $\rho_t = c/t$, each delivering one-step $L^2(\mathrm{op})$ stabilization of order $t^{-1}$, yielding the CLT $\sqrt{n}(\bar{x}_n - x^*) \to N(0, H^{-1}SH^{-1})$; under bounded inputs the pathwise rate $\beta = 1$ further preserves the $n^{-1/6}$ Wasserstein rate at $\alpha^* = 2/3$. Under standard regularity conditions, Wald-type online inference remains valid for dynamically preconditioned averaged SGD whose stabilization rate exceeds the threshold.
Read the full abstract

Mathematical Physics

A Note on the Resolvent Algebra and Functional Integral Approach to the van Hove Model

This paper is a collection of the author's computational notes on the van Hove model and contains no essentially new results. We discuss, from both the operator-algebraic perspective via the Weyl algebra and the resolvent algebra and the functional integral approach, the removal of infrared and ultraviolet cutoffs and the existence of the ground state and $\beta$-KMS states in the case of a point source. In the infinite-volume system at finite temperature, Bose–Einstein condensation can arise.
Read the full abstract

Mathematical Physics

Q-Manifolds and Sigma Models

Recent developments of Batalin-Vilkovisky (BV) formalism and related geometry are reviewed. Mathematical structures of BV formalism are summarized as a Q-manifold and a QP-manifold. Lie algebras, Lie algebroids and other higher algebroids are explained as typical examples of Q- and QP-manifolds. Finally, the BV action functionals are constructed by geometric structures of Lie algebroids and Courant algebroids.
Read the full abstract

Optimization and Control

Optimality Conditions and Numerical Algorithms for a Class of Minimax Bilevel Optimization Problems

In many applications, including Stackelberg games, machine learning, and power systems \cite{Mackay2018Selftuning,Heinrich1952The,Wang2021Bi-Level}, the decisions in a minimax optimization problem can be constrained by a solution to an optimization problem. In this paper, we introduce optimality conditions of this novel minimax bilevel optimization problem and develops efficient first-order algorithms for this class of problems. Firstly, we establish the optimality conditions for minimax bilevel problems by reconstructing the lower-level problem through its Karush-Kuhn-Tucker (KKT) conditions and value function. Secondly, we develop a penalty method framework to approximately solve the minimax bilevel problem by transforming it into a single-level minimax problem. Thirdly, we design a projected gradient multi-step ascent descent method to solve the resulting minimax problem, which can find an $\epsilon$-KKT solution for the original minimax bilevel problem within $\mathcal{O}(\epsilon^{-3} \log(\epsilon^{-1}))$ iterations. To improve {the convergence rate} of the algorithm, we provide its Nesterov accelerated extension with $\mathcal{O}(\epsilon^{-3} \log(\epsilon^{-1}))$ iteration complexity. Finally, we demonstrate the effectiveness of our model and algorithms through numerical experiments on various minimax bilevel optimization problems and a bilevel economic dispatch in the power system.
Read the full abstract

All authors Yaling Hu, Jiani Wang, Yu-hong Dai, Xiaojiao Tong

Number Theory

The Absolute Anabelian Geometry of Virtual Curves of Arbitrary Genus

The objective of this paper is to further study the anabelian object referred to as pointed virtual curves. Building upon previous work that investigated these fundamental-group-theoretic pullbacks of Galois sections in the genus-zero situation, we extend the central anabelian results to curves of arbitrary genus. To facilitate this generalization, we introduce the group-theoretic notion of an inclusion of CAVC-type and the categorical-theoretic notion of a virtual decuspidaloid. Furthermore, we establish a criterion regarding the "geometricity" of certain virtual curves, providing group-theoretic conditions under which a section of an arithmetic fundamental group arises from a rational point.
Read the full abstract

Operator Algebras

Projectional Skeletons of Fourier Algebras

The preduals of $W^*$-algebras are 1-Plichko spaces. A natural question arises: does every predual possess a projectional skeleton (PS) $\{P_s:s\in J\}$ such that each $P_s^*$ is a conditional expectation? In this note, we answer this question affirmatively for the preduals of the group von Neumann algebras of locally compact groups.
Read the full abstract

Number Theory

$2$-Selmer groups, $2$-class groups, and congruent numbers

In this article, we study necessary conditions for certain square-free integers to be congruent numbers. Our method uses divisibility properties of class numbers of related imaginary quadratic fields. We first consider positive square-free integers of the form $n = p_1 p_2 \cdots p_t q,$ where each prime $p_i \equiv 5 \pmod{8}$ and $q \equiv 7 \pmod{8}$. We show that if such an integer $n$ is a congruent number, then the class number $h(-n)$ of the quadratic field $\mathbb{Q}(\sqrt{-n})$ satisfies a specific divisibility condition. Furthermore, we provide quantitative lower bounds on the number of non-congruent numbers of this form. Next, we study integers of the form $n = p_1 p_2 \cdots p_t q,$ with $p_i \equiv 5 \pmod{8}$ and $q \equiv 3 \pmod{8}$. Assuming that $n$ is a congruent number, we obtain a congruence modulo powers of $2$ between the class numbers of the fields $\mathbb{Q}(\sqrt{-n})$ and $\mathbb{Q}\!\left(\sqrt{-p_1 p_2 \cdots p_t}\right)$.
Read the full abstract

Saturday 25 April

Optimization and Control

Multi-Path Routing in Decentralized Exchange Networks: Convex Allocation and an Improving-Path Certificate

We present a graph-theoretic and convex optimization framework for multi-path routing in decentralized exchange networks, together with its implementation and empirical evaluation on Ethereum mainnet. The framework models the market as a directed token multigraph whose arcs carry AMM exchange functions. Routing is decomposed into two implemented layers: candidate path generation via gas-aware marginal k-shortest-path enumeration, where edge scores embed expected execution cost directly into graph traversal with an explicit pool-simple constraint tracked during path construction, and continuous flow allocation over the selected candidates solved as a concave maximization over a simplex with a per-pool price-impact cap. Under standard concavity and monotonicity assumptions, the KKT conditions imply marginal-output equalization across active paths. The central technical contribution is an improving-path certificate: after solving the allocation on k=20 candidate paths, the KKT multiplier is used as a threshold to determine via a single shortest-path query whether any omitted pool-simple path could improve the current solution; in our implementation the certificate confirms sufficiency in the majority of epochs. Execution is protected by an on-chain slippage tolerance enforced at the smart-contract level. We evaluate the implemented engine against four production DEX aggregators on repeated WETH-USDT quote observations across six trade sizes on Ethereum mainnet: median shortfall is below 5 bps across all sizes and top-3 quote rank exceeds 57% of epochs.
Read the full abstract

Metric Geometry

Progress in Formalizing Sphere Packing in Dimension 8

In 2016, Viazovska famously solved the sphere packing problem in dimension $8$, using modular forms to construct a 'magic' function satisfying optimality conditions determined by Cohn and Elkies in 2003. In March 2024, Hariharan and Viazovska launched a project to formalize this solution and related mathematical facts in the Lean Theorem Prover. A significant milestone was achieved in February 2026: the result was formally verified, with the final stages of the verification done by Math, Inc.'s autoformalization model 'Gauss'. We discuss the techniques used to achieve this milestone, reflect on the unique collaboration between humans and Gauss, and discuss project objectives that remain.
Read the full abstract

All authors Sidharth Hariharan, Christopher Birkbeck, Seewoo Lee, Ho Kiu Gareth Ma, Bhavik Mehta, Auguste Poiroux, Maryna Viazovska

Statistics Theory · Cross-listed from stat.ME

A theory of ROC analysis of rule-out and rule-in diagnostics with applications to mammography data

Multiple diagnostic tests are frequently used to determine the presence of a disease condition in patients. In this paper, we use bivariate copulas to examine the properties of receiver operating characteristic (ROC) curves formed when two correlated diagnostic tests are used together to rule-out ("believe the negative") and rule-in ("believe the positive") patients for disease. We use this theory to analyze three mammography data sets where AI devices are applied to reduce radiologists' workload or improve diagnostic performance. Our analysis shows with generality that increasing the radiologist-AI correlation for diseased cases enhances the area under the ROC curve (AUC) of a radiologist-AI rule-out curve, whereas decreasing correlation for non-diseased cases has a similar effect. The opposite trends hold for rule-in scenarios. Applications to clinical mammography data show that projected empirical radiologist performance under a rule-out or rule-in scenario is consistent with the theory.
Read the full abstract

All authors Michelle Mastrianni, Kwok Lung Fan, Yee Lam Elim Thompson, Jessie J.J. Gommers, Ioannis Sechopoulos, Fredrik Strand, Weijie Chen, Gary Levine, Mukul Sherekar, Frank W. Samuelson

Probability

Invariant measures for the open KPZ equation: the Gaussian case

In [arXiv:2409.08465], Quastel and Gu use Stein's equation and integration by parts to give a direct proof that drifted Brownian motions are stationary (modulo height shifts) for the full-line KPZ equation. In this article, we consider the open KPZ equation with boundary conditions $\partial_x h(t,0) = \partial_x h(t,1) = \alpha$ for a general real parameter $\alpha$, and emulate the approach of Quastel and Gu to provide a similar proof that Brownian motion with constant drift $\alpha$ is invariant (modulo height shifts) in this case.
Read the full abstract

Probability

Scaling limit of Sinkhorn-rescaled Random Matrices via Stability of Static Schrödinger Bridges

We analyze the asymptotic behavior and scaling limits of large random matrices rescaled via the Sinkhorn algorithm to match prescribed row and column margins. For a random matrix with independent sub-exponential entries, we show that its Sinkhorn rescaling concentrates around the rescaling of its mean matrix, both at the level of the Schrödinger potentials and as random measures on the unit square, with explicit non-asymptotic rates. As the dimensions grow, the rescaled random matrix converges to the continuous static Schrödinger bridge (SSB) determined by the limiting margins and reference density. Around this scaling limit we develop a fluctuation theory: bulk rigidity for the empirical spectral distribution of the associated sample covariance matrix, and a central limit theorem for the empirical Schrödinger potentials of the rescaled empirical mean. Our analysis is driven by a new quantitative stability theory for the SSB, developed in three forms: Lipschitz continuity in the Hellinger distance under perturbations of the reference measure (kernel stability); Hölder-$1/2$ continuity in the Hellinger distance under $L^1$ perturbations of the margins (margin stability); and $L^\infty$ stability of the discrete Schrödinger potentials under margin perturbation (potential stability). Translated to the discrete random-matrix setting, these bounds yield the concentration and scaling-limit results, while a local law for random Gram matrices with a non-uniform variance profile drives the bulk rigidity. Our SSB stability theory may be of independent interest.
Read the full abstract

Differential Geometry

On Einstein-type manifold with cyclic parallel Ricci tensor

In this article, we derive an integral formula involving the tensor $D_{ijk}$ for compact Einstein-type manifolds with constant scalar curvature. As an application, we classify three-dimensional compact Einstein-type manifolds satisfying the cyclic parallel Ricci tensor condition, obtaining rigidity results that extend and unify previous work in the literature.
Read the full abstract

All authors M. Andrade, H. Baltazar, A. da Silva, D. Tavares

Numerical Analysis

A robust a posteriori error estimator for the Oseen problem

A residual-based a posteriori error estimator is proposed for the incompressible Oseen problem in the convection-dominated regime. The SUPG/PSPG/grad-div stabilized finite element method is used as discretization. The error estimator estimates the global error in a norm that is used in the a priori error analysis of the method. Based on several hypotheses concerning the error and interpolation errors, the robustness of the estimator in the convection-dominated regime is proved. Numerical studies support the analytic results. Finally, the extension of the a posteriori error estimator to the steady-state Navier–Stokes equations is discussed.
Read the full abstract

Number Theory

A necessary condition for a congruent number of the form $8k+3$

A positive square-free integer is called a congruent number if it arises as the area of a right triangle with rational side lengths. Let $ n = p_1p_2 \cdots p_t q $ be a square-free integer, where each $ p_i \equiv 1 \pmod{8} $ and $ q \equiv 3 \pmod{8} $, with the $ p_i $ and $ q $ being distinct primes. In this article, we present a congruence relation modulo powers of 2 between the 2-part of the class numbers of $ \mathbb{Q}(\sqrt{-n}) $ and $ \mathbb{Q}(\sqrt{-p_1p_2 \cdots p_t}) $, under the assumption that $ n $ is a congruent number, using a modified Rédei matrix.
Read the full abstract

Functional Analysis

Invariant Subspaces and the $C_{00}$-Property of $3$-Brownian Shifts

In this paper, we introduce a $3$-Brownian shift $T_{\sigma, \theta}$ on the Hilbert space $H^2(\mathbb D^2)\oplus H^2(\mathbb D)\oplus \mathbb C,$ which is a natural extension of the classical Brownian shift $B_{\sigma, \theta}$ on $H^2(\mathbb D)\oplus \mathbb C$. This is motivated by Brownian extensions in the context of 3-isometries recently developed by A. Crăciunescu and L. Suciu. We investigate the problem of unitary equivalence for $3$-Brownian shifts on invariant subspaces of the type $M_0 \oplus M_1,$ where $ M_0 \subseteq H^2(\mathbb D^2)$ and $ M_1 \subseteq H^2(\mathbb D)\oplus \mathbb C.$ Here, $M_1$ turns out to be an invariant subspace of the respective Brownian shift $B_{\sigma, \theta}$. We also study the asymptotic behaviour of the normalized $3$-Brownian shifts. This work is motivated by Richter \cite{R88} and very recently by work on Brownian shift on $H^2(\mathbb D)\oplus \mathbb C$ in \cite{DDS2025}.
Read the full abstract

Optimization and Control · Cross-listed from stat.ML

Inference of Online Newton Methods with Nesterov's Accelerated Sketching

Reliable decision-making with streaming data requires principled uncertainty quantification of online methods. While first-order methods enable efficient iterate updates, their inference procedures still require updating proper (covariance) matrices, incurring $O(d^2)$ time and memory complexity, and are sensitive to ill-conditioning and noise heterogeneity of the problem. This costly inference task offers an opportunity for more robust second-order methods, which are, however, bottlenecked by solving Newton systems with $O(d^3)$ complexity. In this paper, we address this gap by studying an online Newton method with Hessian averaging, where the Newton direction at each step is approximately computed using a sketch-and-project solver with Nesterov's acceleration, matching $O(d^2)$ complexity of first-order methods. For the proposed method, we quantify its uncertainty arising from both random data and randomized computation. Under standard smoothness and moment conditions, we establish global almost-sure convergence, prove asymptotic normality of the last iterate with a limiting covariance characterized by a Lyapunov equation, and develop a fully online covariance estimator with non-asymptotic convergence guarantees. We also connect the resulting uncertainty quantification to that of exact and sketched Newton methods without Nesterov's acceleration. Extensive experiments on regression models demonstrate the superiority of the proposed method for online inference.
Read the full abstract