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)

Friday 3 April

Number Theory

Mixed character sums modulo prime powers

We obtain explicit estimates for the mixed character sum $S= S(\chi,g,f,p^m) = \sum_{x=1}^{p^m} \chi (g(x)) e_{p^m}(f(x))$, where $p^m$ is a prime power, $\chi$ is a multiplicative character mod $p^m$ and $f,g$ are rational functions over $\mathbb Q$. Let $f=f_+/f_-$, $g=g_+/g_-$ in reduced form, and set $D=\text{deg}(f)+Z-1$ where $Z$ is the number of distinct complex zeros of $f_-g_+g_-$, and $\Delta= \text{deg}(f)+\text{deg}(g)$ for polynomial $f,g$, $\Delta=2(\text{deg}(f)+\text{deg}(g))$ otherwise. We show for example that for odd $p$, any non-degenerate sum has $|S|\le 3^{4/3}\, p^{m(1-\frac 1D)}$ if $\text{deg}_p(f) \ge 1$, and $|S| \le 3^{4/3}\, p^{m(1-\frac 1\Delta)}$ if $\text{deg}_p(g) \ge 1$. Analogous bounds are given for degenerate sums.
Read the full abstract

Dynamical Systems · Cross-listed from cs.LG

WGFINNs: Weak formulation-based GENERIC formalism informed neural networks

Data-driven discovery of governing equations from noisy observations remains a fundamental challenge in scientific machine learning. While GENERIC formalism informed neural networks (GFINNs) provide a principled framework that enforces the laws of thermodynamics by construction, their reliance on strong-form loss formulations makes them highly sensitive to measurement noise. To address this limitation, we propose weak formulation-based GENERIC formalism informed neural networks (WGFINNs), which integrate the weak formulation of dynamical systems with the structure-preserving architecture of GFINNs. WGFINNs significantly enhance robustness to noisy data while retaining exact satisfaction of GENERIC degeneracy and symmetry conditions. We further incorporate a state-wise weighted loss and a residual-based attention mechanism to mitigate scale imbalance across state variables. Theoretical analysis contrasts quantitative differences between the strong-form and the weak-form estimators. Mainly, the strong-form estimator diverges as the time step decreases in the presence of noise, while the weak-form estimator can be accurate even with noisy data if test functions satisfy certain conditions. Numerical experiments demonstrate that WGFINNs consistently outperform GFINNs at varying noise levels, achieving more accurate predictions and reliable recovery of physical quantities.
Read the full abstract

All authors Jun Sur Richard Park, Auroni Huque Hashim, Siu Wun Cheung, Youngsoo Choi, Yeonjong Shin

Analysis of PDEs

Chemotaxis models with signal-dependent sensitivity and a logistic-type source, II: Persistence and stabilization

This paper is Part II of a series on boundedness, global existence, and asymptotic behavior of positive classical solutions to the chemotaxis model $$\begin{cases} u_t=\Delta u-\chi_0\nabla\cdot\left(\frac{u^m}{(1+v)^\beta}\nabla v\right)+au-bu^{1+\alpha}, & x\in\Omega,\\ 0=\Delta v-\mu v+\nu u^\gamma, & x\in\Omega,\\ \frac{\partial u}{\partial n}=\frac{\partial v}{\partial n}=0, & x\in\partial\Omega, \end{cases} \tag{CM}$$ where $\Omega\subset\mathbb{R}^N$ is a bounded smooth domain, $\alpha,\gamma,m,\mu,\nu>0$, $\chi_0\in\mathbb{R}$, and $a,b,\beta\ge0$. Part I established biologically relevant parameter regimes ensuring boundedness and global existence of positive classical solutions. Here we study persistence and stabilization of globally defined bounded positive solutions, quantifying how $\beta$ in the sensitivity $\chi(v)=\chi_0(1+v)^{-\beta}$ influences long-time dynamics. The factor $(1+v)^{-\beta}$ models signal-dependent desensitization, so large $\beta$ should promote stabilization, though it considerably complicates the analysis. We take $\mu=\nu=1$ with either $a=b=1$ (logistic source) or $a=b=0$ (minimal model). We show that when $m\ge1$, every globally defined bounded positive solution stays uniformly away from zero in space. We determine exact $\beta$-dependent critical sensitivity thresholds for local stability of constant solutions and establish $\beta$-dependent thresholds for their global stability; both tend to $\infty$ as $\beta\to\infty$. Thus signal saturation (large $\beta$) or repulsion ($\chi_0<0$) can prevent aggregation and promote relaxation to spatially homogeneous states. For $\beta>0$ and $m,\alpha,\gamma$ not all 1, we develop new techniques, including a Lyapunov function, pointwise estimates for $\nabla v$, and a nontrivial extension of the rectangle/ODE method from $\beta=0$ to $\beta>0$.
Read the full abstract

Thursday 2 April

Rings and Algebras

Unitary-Invariant Decomposition of Reducible Total Least Squares Core Problems

The analysis of a total least squares problem can be reduced to that of an associated core problem, which typically has lower dimension and improved solubility properties. Nevertheless, even a core problem may remain reducible, admitting further decomposition into irreducible component sub- problems with simpler structure and better analytical properties. We seek the finest decomposition under unitary transformations over C; for real data this decomposition can be strictly finer than the finest one under real orthogonal transformations. We split the data matrix into row blocks associated with the singular subspaces of the system matrix and consider the graph whose ver- tices are the blocks and whose edges correspond to nonzero cross-products. After preliminary splittings that make every nonzero cross-product a scalar multiple of a unitary matrix, we show that the components can be obtained directly from these blocks if and only if the covariance matrices along all cycles of the graph are scalar matrices. Otherwise, there is a chordless cycle whose covariance matrix is not a scalar matrix, and splitting the blocks on this cy- cle along the eigenspaces of the covariance matrix removes the obstruction without changing the decomposition. Based on these results, we propose a recursive strategy that finds all irreducible components over the complex field. The strategy is exact and theoretical rather than an efficient numerical algo- rithm; a small floating-point computation illustrates it. The results depend only on the core problem and therefore apply to general linear approximation problems AX B.
Read the full abstract

Combinatorics

$RD_\alpha$-Spectra of Joined Union Graphs with Applications to Power Graphs of Finite Groups

The generalized reciprocal distance matrix of a graph $\mathscr{G}$, denoted by $RD_\alpha(\mathscr{G})$, is defined as $RD_\alpha(\mathscr{G})=\alpha\,RT_r(\mathscr{G})+(1-\alpha)\,RD(\mathscr{G}), \, \alpha\in[0,1],$ where $RT_r(\mathscr{G})$ represents the diagonal matrix of reciprocal vertex transmissions, and $RD(\mathscr{G})$ is the Harary (reciprocal distance) matrix of $\mathscr{G}$. In this paper, we investigate the $RD_\alpha$-spectrum of graphs obtained through the joined union operation. We derive explicit formulas for the characteristic polynomial of $RD_\alpha(\mathscr{G})$ when $\mathscr{G}$ is formed as a joined union of regular graphs. These results provide closed-form expressions for the corresponding spectra of several important graph classes. Moreover, we show that the power graphs of the dihedral group $D_{2n}$ and the generalized quaternion group $Q_{4n}$ admit representations as joined union graphs. Using this structural characterization, we determine the $RD_\alpha$-spectra of power graphs arising from various classes of finite groups, including cyclic groups $\mathbb{Z}_n$, dihedral groups $D_{2n}$, generalized quaternion groups $Q_{4n}$, elementary abelian $p$-groups, and certain non-abelian groups of order $pq$.
Read the full abstract

General Mathematics

Finite Field Tarski-Maligranda Inequalities

Let $\mathbb{F}$ be a sub-modulus field such that $2 \neq 0$. Let $\mathcal{X}$ be a sub-normed linear space over $\mathbb{F}$. Then we show that \begin{align*} \bigg|\|x\|-\|y\|\bigg|\leq \frac{2}{|2|}\|x+y\|+\frac{2}{|2|}\max\{\|x-y\|, \|y-x\|\}-(\|x\|+\|y\|) \end{align*} and \begin{align*} \bigg|\|x\|-\|y\|\bigg|\leq \|x\|+\|y\|-\frac{2}{|2|}\|x+y\|+\frac{2}{|2|}\max\{\|y-x\|, \|x-y\|\}. \end{align*} Above inequalities are finite field versions of important Tarski-Maligranda inequalities obained by Maligranda [Banach J. Math. Anal., 2008].
Read the full abstract

Analysis of PDEs

Sharp mean Hadamard inequalities and polyconvex integrands that give rise to convex functionals

We investigate several instances of the Hadamard inequality in the mean in two dimensions. As a consequence, we prove the uniqueness of minimizers of an integral functional with a polyconvex integrand, subject to mixed Dirichlet and Neumann boundary conditions. The theoretical findings are complemented by computational experiments that illustrate the behavior of the minimizers.
Read the full abstract

Number Theory

Series for $1/\pi$ arising from Cauchy product

In this note, we evaluate a series for $1/\pi$ conjectured by Sun. Our proof uses the Cauchy product and hypergeometric transformations. From this result, we derive two additional analogous series for $1/\pi$ involving polynomials of degree $3$. Further identities can be proved using our method; these are presented in a table at the end of the note.
Read the full abstract

Logic

The first fatal axiom for weakened sequential products on finite MV-effect algebras: Local obstruction, exact low-rank classification, and the rank-one boundary case

We study total binary operations on effect algebras obtained by truncating the Gudder-Greechie axiom package for a sequential product. The point is not to reprove the known nonexistence of non-Boolean full sequential products on finite chains, but to determine, axiom by axiom, where finite MV-effect algebras first fail. We prove two structural facts valid on every effect algebra. First, the operation \sigma_E(a,b) = 0 if a=0, and b if a \neq 0, satisfies (S1)-(S3), so (S3) is never fatal by itself. Second, any operation satisfying (S1)-(S4) already has the right-unit property a \circ 1 = a, even without (S5). From this we derive a local obstruction theorem: if an effect algebra contains an atom of finite isotropic index at least 2, then it admits no (S1)-(S4) operation. Consequently, a finite MV-effect algebra admits such an operation if and only if it is Boolean. In this precise sense, (S4) is the first fatal axiom on finite MV-effect algebras. On the constructive side, let E_u = [0,u] \subseteq Z^r be the simplicial interval representation of a finite MV-effect algebra. We show that additive maps E_u \to E_v are exactly the restrictions of positive group homomorphisms Z^r \to Z^s, equivalently maps x \mapsto Mx given by nonnegative integer matrices with Mu \le v. This yields a complete classification of (S1)+(S2) operations by row-wise subunital matrices. We then solve the first genuinely higher-rank (S1)-(S3) problem: on the rank-two Boolean algebra B_2 = E_{(1,1)} \cong C_1^2, all such operations are classified and there are exactly 34. Thus the finite-chain collapse at (S3) is a rank-one boundary phenomenon, whereas on finite MV-effect algebras the sharp threshold for nonexistence occurs exactly at (S4).
Read the full abstract

Analysis of PDEs · Cross-listed from cs.LG

General Explicit Network (GEN): A novel deep learning architecture for solving partial differential equations

Machine learning, especially physics-informed neural networks (PINNs) and their neural network variants, has been widely used to solve problems involving partial differential equations (PDEs). The successful deployment of such methods beyond academic research remains limited. For example, PINN methods primarily consider discrete point-to-point fitting and fail to account for the potential properties of real solutions. The adoption of continuous activation functions in these approaches leads to local characteristics that align with the equation solutions while resulting in poor extensibility and robustness. A general explicit network (GEN) that implements point-to-function PDE solving is proposed in this paper. The "function" component can be constructed based on our prior knowledge of the original PDEs through corresponding basis functions for fitting. The experimental results demonstrate that this approach enables solutions with high robustness and strong extensibility to be obtained.
Read the full abstract

All authors Genwei Ma, Ting Luo, Ping Yang, Xing Zhao

Logic

A Concise Proof of the $L_0$ Dichotomy

Carroy, Miller, Schrittesser, and Vidnyánszky established the $L_0$ dichotomy: there is a Borel graph of Borel chromatic number three that admits a continuous homomorphism to every analytic graph of Borel chromatic number at least three. Their proof relies on a transfinite analysis of terminal approximations over a decreasing $\omega_1$-sequence of analytic sets. I give a new, substantially shorter proof of this result by adapting the graph-theoretic framework recently introduced by Bernshteyn for the $G_0$ dichotomy. The central device is a $\sigma$-ideal of small sets of homomorphisms from finite path approximations into the target graph, where smallness is witnessed by a bounded odd-walk condition on vertex projections. The key lemma that largeness is preserved under the doubling operation is established via the First Reflection Theorem, replacing the original transfinite construction with a single Borel reflection argument. The continuous homomorphism from the canonical graph $\mathbb{L}_c$ into the target is then obtained as a limit of shrinking families of copies, in direct analogy with Bernshteyn's proof for $G_0$.
Read the full abstract

Functional Analysis

A Classification of Order Convergence via a Transfinite Fatou Hierarchy

We investigate the descriptive complexity of order convergence in separable Banach lattices. While uniform convergence is Borel and $\sigma$-order convergence is known to be ${\bf \Delta}^1_2$, it is unclear in general when $\sigma$-order convergence is analytic. We introduce a transfinite hierarchy of weakenings of the classical Fatou property, indexed by countable ordinals, and show that it provides a complete structural classification of this definability problem. For a separable Banach lattice $X$, we prove that the following are equivalent: (i) the set of decreasing positive sequences with infimum zero is Borel; (ii) $\sigma$-order convergence is analytic; and (iii) $X$ satisfies the $\alpha$-Fatou property for some countable ordinal $\alpha$. We further establish that the hierarchy is proper: for every countable ordinal $\alpha$ there exists a separable Banach lattice with a countable $\pi$-basis that fails to be $\alpha$-Fatou, but is $\beta$-Fatou for some $\beta>\alpha$. Thus the Borel definability of order convergence is governed by a canonical ordinal invariant intrinsic to the lattice, and the descriptive complexity can be arbitrarily high below $\omega_1$. These results identify projective complexity as a genuine structural invariant in Banach lattice theory.
Read the full abstract

All authors Antonio Avilés, Christian Rosendal, Mitchell A. Taylor, Pedro Tradacete

Combinatorics

The Invariance Reduction Process – a New Tool to Solve Circular Nim and Related Games

We introduce the notion of invariant vectors of a game and develop the Invariance Reduction Process, which first uses reduction of positions via invariance and then zero and merge reductions of games to arrive at smaller, solved sub-games for closed subspaces of the positions. This process makes it much easier to prove that there are moves from N-positions to P-positions, and can also be used in some cases to show that there are no moves between P-positions. This process is suitable for all variations of the game Nim whose rule sets form a simplicial complex. We rephrase Simplicial Nim as Set Nim SN($n,A$) and derive results on the structure of the P-positions in terms of invariant vectors, without needing the background and notation of simplicial complexes. We also show that invariant vectors differ from the circuits used to describe the P-positions in Simplicial Nim and that invariant vectors have wider applicability compared to circuits. We apply the Invariance Reduction Process to derive results on the P-positions of the family of Path Nim games where play is allowed on at least half the stacks, as well as for the Circular Nim games CN($n,k$) with $n=7, k=3$ and $n=8,k=3$.
Read the full abstract

Numerical Analysis · Cross-listed from stat.ML

Learning interacting particle systems from unlabeled data

Learning the potentials of interacting particle systems is a fundamental task across various scientific disciplines. A major challenge is that unlabeled data collected at discrete time points lack trajectory information due to limitations in data collection methods or privacy constraints. We address this challenge by introducing a trajectory-free self-test loss function that leverages the weak-form stochastic evolution equation of the empirical distribution. The loss function is quadratic in potentials, supporting parametric and nonparametric regression algorithms for robust estimation that scale to large, high-dimensional systems with big data. Systematic numerical tests show that our method outperforms baseline methods that regress on trajectories recovered via label matching, tolerating large observation time steps. We establish the convergence of parametric estimators as the sample size increases, providing a theoretical foundation for the proposed approach.
Read the full abstract

Probability

Finite reservoirs lead to Wentzell boundary conditions for independent random walks and exclusion process

We analyze the scaling limits (hydrodynamics/propagation of local equilibrium) of two particle systems, namely independent random walks (IRW) and symmetric exclusion processes (SEP), both in the discrete one-dimensional segment where the left boundary is in contact with a reservoir, which may hold any number of particles. At rate one a particle jumps from site $1$ to the reservoir, and at rate $\alpha \eta(0) N^{-\theta}$ a particle jumps from the reservoir to site $1$ (if site $1$ is empty in SEP), where $\eta(0)$ is the total number of particles in the reservoir and $\theta\geq 0$ is a parameter whose tuning leads to a dynamical phase transition. For every $\theta\geq 0$, the hydrodynamic equation is the heat equation with Neumann boundary conditions at the right boundary for both processes, while the boundary conditions at the left boundary depend on the value of $\theta$. For $0\leq \theta<1$, it is given by Neumann boundary conditions, meaning the deposit is asymptotically empty, acting as a barrier. For $\theta>1$, and for IRW (resp. SEP) it is given by a non-homogeneous (resp. homogeneous) Dirichlet boundary condition, meaning the reservoir becomes asymptotically infinite, acting as a heat bath (resp. the reservoir behaves as a sink). For $\theta=1$, we obtain a non-local Dirichlet boundary condition, which is additionally non-linear for SEP. As a by-product, we obtain an equivalence between solutions to the heat equation with Wentzell boundary conditions and with non-local Dirichlet boundary conditions related to the total mass of the system.
Read the full abstract

Quantum Algebra

Representation Category of Free Wreath Product of Classical Groups

In this paper, we construct a rigid concrete $C^*$-tensor category whose associated compact quantum group, reconstructed via Woronowicz–Tannaka–Krein duality, is the free wreath product of classical groups.

Optimization and Control · Cross-listed from quant-ph

RFOX (Rotated-Field Oscillatory eXchange) quantum algorithm: Towards Parameter-Free Quantum Optimizers

We introduce RFOX (Rotated-Field Oscillatory eXchange), a parameter-free quantum algorithm for combinatorial optimization that combines an almost constant non-stoquastic $XX$ catalyst with a weak harmonic $ZX$ counter-diabatic term. Using the Floquet-Magnus expansion, we derive an effective Hamiltonian whose leading-order $\mathcal{O}(\delta/\omega)$ corrections yield local $Y$ fields, field-modulated 2-body terms, and poly-local 3-body topological interactions driven by graph connectivity. This structure ensures a nearly flat instantaneous spectral gap, preventing the unpredictable gap collapses typical of conventional $X$ (stoquastic), $XX$, and $X+sXX$ (non-stoquastic) driver schedules. Extensive noiseless simulations and physical hardware experiments on IBM Quantum processors (up to 20 qubits) validate our spectral predictions. RFOX consistently attains near-optimal or exact ground states in the random-field Ising model using up to an order of magnitude fewer Trotter slices, with an advantage that grows alongside problem disorder. These results suggest that fixed-gap, non-stoquastic drivers augmented with analytically derived counter-diabatic terms offer a scalable, tuning-free route for quantum optimization.
Read the full abstract

Algebraic Topology

Integral bases for the second degree cohomology of 4-dimensional toric orbifolds

We study toric orbifolds of real dimension four with vanishing odd-degree cohomology and obtain a basis for its degree-two equivariant cohomology with integral coefficients by identifying it with the intersection of certain lattices. As applications, we provide an alternative construction of the algebraic cellular basis for integral ordinary cohomology \cite{FSS2}. In addition, when the toric orbifold is an algebraic variety, we determine its Cartier divisor group and Picard group.
Read the full abstract

Commutative Algebra

On topologies on the space of valuations and the valuative tree

In this paper, we discuss topological aspects of the space of valuations $\mathbb{V}$ and the valuative tree $\mathcal{T}(v,\Lambda)$. We present a relation between the weak tree topology and the Scott topology in $\mathcal{T}(v,\Lambda)$ and describe the supremum of an increasing family of valuations in a special subtree. We also view the valuative tree as a subset of the product $(\Lambda_\infty)^{K[x]}$ and prove that it is closed if we consider the natural product topology.
Read the full abstract

Optimization and Control · Cross-listed from cs.LG

Communication-Efficient Distributed Learning with Differential Privacy

We address nonconvex learning problems over undirected networks. In particular, we focus on the challenge of designing an algorithm that is both communication-efficient and that guarantees the privacy of the agents' data. The first goal is achieved through a local training approach, which reduces communication frequency. The second goal is achieved by perturbing gradients during local training, specifically through gradient clipping and additive noise. We prove that the resulting algorithm converges to a stationary point of the problem within a bounded distance. Additionally, we provide theoretical privacy guarantees within a differential privacy framework that ensure agents' training data cannot be inferred from the trained model shared over the network. We show the algorithm's superior performance on a classification task under the same privacy budget, compared with state-of-the-art methods.
Read the full abstract

All authors Xiaoxing Ren, Yuwen Ma, Nicola Bastianello, Karl H. Johansson, Thomas Parisini, Andreas A. Malikopoulos

Combinatorics

Cascade-free sequences, dispersion index, and state avoidance for stateful digit-wise operations

We show that cascade-free counting from carry theory is a special case of a general transfer matrix construction. For any binary stateful digit-wise operation with GEN/PROP/KILL decomposition, the number of cascade-free sequences of length $L$ depends on only two parameters: the alphabet size $N$ and the product $d = |\text{GEN}| \cdot |\text{PROP}|$. The resulting sequence satisfies $a(L) = N a(L-1) - d a(L-2)$ and equals a scaled Chebyshev polynomial of the second kind with coupling parameter $x = N/(2\sqrt{d}) \geq 1$. We instantiate this for digit-wise addition and doubling in base $p$. For odd primes the exact relation $a_{\text{carry}}(L) = p^L a_{\text{dbl}}(L)$ holds. For $p = 3$ the cascade-free doubling count equals the Fibonacci bisection $F(2L+2)$ via $U_L(3/2) = F(2L+2)$ (OEIS A001906); we are not aware of this interpretation in the existing literature. We analyse the dispersion index $D = \text{Var}(\nu)/E[\nu]$ of the state count for uniformly distributed inputs. For symmetric chains ($g = k$) the Poisson transition $D_\infty = 1$ occurs at $\mu = 1/3$, corresponding to base 3 where the Fibonacci bisection appears. The finite Poisson transition point $\mu^*(L)$ decreases strictly to $1/3$ with rate $1/(6L) + O(1/L^2)$. We generalise to state spaces $|S| > 2$ via state avoidance. The restricted transfer matrix has dimension $s-1$; the Chebyshev representation persists for $|S| = 3$.
Read the full abstract

Optimization and Control

Moreau Envelope-Based Clustering for Generalized Multi-Source Weber Problem

In this paper, we propose an efficient algorithm for data clustering based on the Moreau envelope, which approximates nonsmooth and nonconvex components of the generalized multi-source Weber problem. The number of clusters is not fixed in advance and is determined automatically by progressively removing empty or redundant clusters. The smoothing induced by the Moreau envelope transforms the original problem into a structured optimization task that can be efficiently solved using first-order methods and simple matrix vector operations. Numerical experiments on synthetic and real datasets show that the proposed approach is fast, scalable, and competitive with existing methods in both clustering quality and computational efficiency.
Read the full abstract

Dynamical Systems

Hidden Harmonic Structure, Universal Damping, and Stability Bounds in Nonlinear Contact Dynamics

Nonlinear contact dynamics are widely regarded as intrinsically nonlinear systems whose behaviour depends strongly on geometry and impact conditions. Here we show that any one-dimensional conservative contact system satisfying monotone energy-consistent conditions admits two complementary structures: (i) a canonical action-angle representation in physical time, and (ii) an exact harmonic oscillator representation under an energy-based coordinate transformation combined with time reparametrisation. This reveals a hidden linear structure underlying nonlinear contact interactions. Building on this result, we derive a unique universal damping law that preserves linear dissipative dynamics in the transformed harmonic space, and establish a rigorous, closed-form lower bound for the critical timestep in numerical simulations. The framework generalises classical power-law contact models and provides a unified basis for restitution control across arbitrary geometries, recovering known exact solutions as explicit monomial special cases.
Read the full abstract

Probability · Cross-listed from physics.ao-ph

On the White-Noise Limit of the Colored Linear Inverse Model

A recent paper by Lien et al. (2025) introduces the "colored linear inverse model" (colored LIM), in which stochastic forcing is modeled using Ornstein-Uhlenbeck colored noise rather than idealized white noise. In that work, it is shown that the derivative-based identification formulas used to estimate model parameters do not admit a regular white-noise limit due to the loss of differentiability of the lag-correlation function at zero lag. Here we revisit the white-noise limit from the perspective of the underlying stochastic differential equations. Treating the colored LIM as an augmented Ornstein-Uhlenbeck system, we show that as the correlation time tau -> 0 the colored-noise-driven system reduces to the classical LIM, and the corresponding stationary covariance satisfies the standard fluctuation-dissipation relation. Re-examining the same linear system used by Lien et al. (2025), we illustrate this convergence numerically. These results highlight a distinction between the singular behavior of derivative-based identification formulas and the regular limiting behavior of the underlying stochastic model. Taken together with recent results showing convergence of estimated parameters in the white-noise limit, they provide a consistent interpretation in which the colored LIM recovers the classical LIM at the level of stochastic dynamics even though certain estimation procedures become ill-defined in that limit.
Read the full abstract

Analysis of PDEs

Viscosity solutions of the integro-differential equation for the Cramér–Lundberg model with annuity payments and investments

This note is an addendum to the work initiated by Promyslov on the integro-differential equation arising in the ruin problem for annuity payment models. First, the existence of viscosity solutions is proved. Then the regularity of these solutions is established, showing that they are indeed classical solutions.
Read the full abstract

Algebraic Topology

A rational model for the fiberwise THH transfer I: Sullivan algebras

Given a map $f$ of fibrations over a space $B$ such that the fiber of $f$ is simply connected and finitely dominated, we prove that its fiberwise THH transfer, considered as a map of parametrized spectra over $B$, is rationally modeled by the Hochschild homology transfer of a Sullivan model of $f$. The proof goes in two steps. Firstly, we use the machinery of higher categorical traces to show that the fiberwise THH transfer can be computed internally to parametrized spectra. Secondly, we model the resulting description rationally using work of Braunack-Mayer, who proved that parametrized spectra can be modeled by modules over Sullivan algebras. In Part II, we will use our result to obtain a rational model of the Becker-Gottlieb transfer, and for applications to manifold topology.
Read the full abstract

Dynamical Systems

A Structurally Flat Triangular Form for Three-Input Systems

We present a broadly applicable structurally flat triangular form for x-flat control-affine systems with three inputs. Building on recent results for the derivative structure of flat outputs, we define the triangular form together with regularity conditions that guarantee structural flatness, and derive necessary and sufficient conditions for a system with a given x-flat output to be static feedback equivalent to this form. Further, we present sufficient conditions under which general x-flat three-input systems can be rendered static feedback equivalent to the proposed triangular form after a finite number of input prolongations.
Read the full abstract

Optimization and Control

Observer-Based Performance-Barrier Event-Triggered Control of $2\times2$ Linear Hyperbolic PDEs

Performance-barrier event-triggered control (P-ETC) is a methodology implemented to increase the dwell-times between events while still preserving a prescribed performance of the system under event-triggered control (ETC). This is achieved by considering the performance residual of the system, which is a measure of the system performance with respect to the prescribed performance. This allows the Lyapunov function candidate to deviate from decreasing monotonically. In order to determine the performance residual, it is required to know the full-state information, leading to all work related to P-ETC to be under full-state feedback. In this article, we propose a novel dynamic performance-barrier under output feedback with an exponentially convergent observer. We consider event-triggered boundary control of a class of $2\times2$ linear hyperbolic PDEs with anti-collocated measurements with the control input. Under the proposed P-ETC mechanism, we prove the existence of a minimum dwell-time, and show the global exponential stability of the spatial $L^2$ norm of the solution of the system. Simulation results are presented to validate the theoretical claims.
Read the full abstract

Optimization and Control

Optimal Projection-Free Adaptive SGD for Matrix Optimization

Recently, Jiang et al. [2026] developed Leon, a practical variant of One-sided Shampoo [Xie et al., 2025a, An et al., 2025] algorithm for online convex optimization, which does not require computing a costly quadratic projection at each iteration. Unfortunately, according to the existing analysis, Leon requires tuning an additional hyperparameter in its preconditioner and cannot achieve dimension-independent convergence guarantees for convex optimization problems beyond the bounded gradients assumption. In this paper, we resolve this issue by proving certain stability properties of Leon's preconditioner. Using our improved analysis, we show that tuning the extra hyperparameter can be avoided and, more importantly, develop the first practical variant of One-sided Shampoo with Nesterov acceleration, which does not require computing projections at each iteration. As a side contribution, we obtain improved dimension-independent rates in the non-smooth non-convex setting and develop a unified analysis of the proposed algorithm, which yields accelerated projection-free adaptive SGD with (block-)diagonal preconditioners.
Read the full abstract

Optimization and Control

On vehicle routing problems with stochastic demands – Scenario-optimal recourse policies

Two-Stage Vehicle Routing Problems with Stochastic Demands (VRPSDs) form a class of stochastic combinatorial optimization problems where routes are planned in advance, demands are revealed upon vehicle arrival, and recourse actions are triggered whenever capacity is exceeded. Following recent works, we consider VRPSDs where demands are given by an empirical probability distribution of scenarios. Existing approaches rely on integer L-shaped (ILS) cuts, whose coefficients are tailored for specific recourse policies. In contrast, we propose a framework that casts recourse policies as solutions of a higher-dimensional mixed-integer program, and we characterize its convex hull in the original lower-dimensional space via a new class of inequalities called scenario recourse inequalities (SRIs). We show that SRIs are valid for any recourse policy satisfying mild assumptions and are sufficient for formulating the VRPSD under a scenario-optimal recourse policy, where the recourse actions are chosen optimally for each scenario. Under this latter policy, we also demonstrate that SRIs dominate several known classes of ILS cuts. We conduct computational experiments on the VRPSD with scenarios under both the classical and the scenario-optimal recourse policies. By using the SRIs, our algorithm solves 329 more instances to optimality than the previous state-of-the-art ILS algorithm.
Read the full abstract