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)

Thursday 7 May

Rings and Algebras

Automorphism groups of direct products of multiplicative monoids of certain rings

In this paper, we establish a rigidity result for automorphisms of multiplicative direct products of $D$-rings which are total ring of fraction that have pairwise distinct cardinalities. Under these assumptions, every automorphism acts independently on each factor, so that no interaction between distinct components occurs; in particular, the automorphism group decomposes canonically as the direct product of the automorphism groups of the factors. As a consequence, the automorphism group of the multiplicative monoid of integers modulo $n$ is entirely determined by its $p$-power components.
Read the full abstract

Information Theory

Identification for Inverse Gaussian Channels

We derive lower and upper bounds on the identification capacity of inverse Gaussian (IG) channels, a fundamental model for molecular communications in fluid environments. The analysis considers deterministic encoding schemes under a peak time constraint and characterizes the asymptotic growth of optimal codebook sizes. In the converse, by leveraging the Borel-Cantelli lemma and the super-exponential near-zero tail behavior of the IG distribution, we establish an almost-sure lower bound on the first-passage time of drifting diffusing molecules. This leads to the conclusion that the identification capacity exhibits super-exponential growth with the codeword length $n$, i.e., $\sim 2^{(n \log n)R_n}$, where $R_n$ is the coding rate.
Read the full abstract

Analysis of PDEs

$s$-harmonic functions in the small order limit

We study families $u_s$ of functions satisfying the equations $(-\Delta)^s u_s=0$, $s \in (0,1)$ in a smooth bounded open set $\Omega \subset \mathbb{R}^N$. The main purpose of this paper is twofold. First, we provide a detailed analysis of the asymptotics of these families in the zero order limit $s \to 0^+$. Second, we study the differentiability of $u_s$ as a function of $s$. Most of our results are devoted to the associated Poisson problem, where the family $u_s$ is determined by the exterior condition $u_s = g$ in $\mathbb{R}^N \setminus \Omega$ for some fixed function $g \in L^\infty(\mathbb{R}^N \setminus \Omega)$. Our results show that both the zero order asymptotics and the differentiability properties of $u_s$ can be expressed in terms of the logarithmic Laplacian of suitable extensions of $g$. This allows to deduce pointwise monotonicity properties of $u_s$ in the order parameter $s$ for a large class of functions $g$.
Read the full abstract

Mathematical Physics

Non-relativistic limit of generalized relativistic Pauli operators by Feynman-Kac formulae

The non-relativistic limit of a generalized relativistic Pauli operator\[H_c^{S,\alpha}=\left(2c^{\beta}\bigl(\sigma\cdot(-i\nabla-a)\bigr)^2+(mc^\gamma)^{2/\alpha}\right)^{\alpha/2}-mc^\gamma+V\]on $L^2(\mathbb{R}^3;\mathbb{C}^2)$ is investigated under the constraint$2\alpha=\gamma\beta+\gamma^2$.This operator generalizes the relativistic Pauli operator within the framework of Bernstein functions.The associated heat semigroup $e^{-tH_c^{S,\alpha}}$ admits a Feynman–Kac representation involving Brownian motion, a subordinator, and a Poisson process.Using this representation, we prove that the semigroup $e^{-tH_c^{S,\alpha}}$ converges strongly to $e^{-tH^{S,\alpha}}$ as $c\to\infty$, where the limiting generator is given by\[H^{S,\alpha}=\frac{\alpha}{2m^{\frac{2}{\alpha}-1}}\bigl(\sigma\cdot(-i\nabla-a)\bigr)^2+V.\]The non-relativistic limit of a generalized relativistic Schrödinger operator is also investigated.
Read the full abstract

Optimization and Control · Cross-listed from eess.SY

Absolute Stability of Nonlinear Negative Imaginary Systems with Application to Potential Energy Shaping

This paper establishes absolute stability conditions for nonlinear negative imaginary (NI) systems interconnected with static nonlinear feedback. We first show that the NI property is preserved when the feedback nonlinearity can be expressed as the gradient of a continuously differentiable function, and the composite storage of the resulting system remains positive definite. This condition provides a direct connection between nonlinear static feedback and storage-function shaping along the measured output channels. Building on this result, conditions are derived for absolute stability of the closed-loop system under mild assumptions. The linear specialization of the results strictly generalizes prior absolute stability results for linear NI systems, allowing coupled nonlinearities not covered by existing slope-restricted or sector-bounded frameworks. Finally, the proposed theory is illustrated through a linear example highlighting this generalization and a nonlinear example that shows the utility of the proposed results in potential energy shaping.
Read the full abstract

Statistics Theory

Time-Inhomogeneous Preconditioned Langevin Dynamics

Langevin sampling from distributions of the form $p(x) \propto \exp(-\Psi(x))$ faces two major challenges: (global) mode coverage and (local) mode exploration. The first challenge is particularly relevant for multi-modal distributions with disjoint modes, whereas the second arises when the potential $\Psi$ exhibits diverse and ill-conditioned local mode geometry. To address these challenges, a common approach is to precondition Langevin dynamics with problem-specific information, such as the sample covariance or the local curvature of $\Psi$. However, existing preconditioner choices inherently involve a trade-off between global mode coverage and local mode exploration, and no prior method resolves both simultaneously. To overcome this limitation, we propose the TIPreL, which introduces a time- and position-dependent preconditioner. This design effectively addresses both challenges mentioned above within a single framework. We establish convergence of the resulting dynamics in the Wasserstein-2 distance both in continuous time and for a tamed Euler discretization. In particular, our analysis extends the existing state of the art by proving convergence under time- and space-dependent diffusion coefficients, and only locally Lipschitz drifts, which has not been covered by prior work. Finally, we experimentally compare TIPreL with competing preconditioning schemes on a two-dimensional, severely ill-posed example and on a Bayesian logistic regression task in higher dimensions, confirming the efficiency of the proposed method.
Read the full abstract

All authors Alexander Falk, Laurenz Nagler, Andreas Habring, Thomas Pock

Representation Theory

A Sugawara–Legendre mechanism for the four-point Heisenberg algebra

This paper provides a representation-theoretic explanation for the recent discovery that families of orthogonal polynomials arise in the centers of universal central extensions of superelliptic affine Lie algebras. Working in the Verma modules of the Heisenberg subalgebra, we compute the canonical contravariant form in closed form on a distinguished family of weight vectors, and show that this family is the Legendre family, forced by the curve itself. The classical squared norms follow by a rescaling available under a positivity condition on the central charge, and the module is irreducible exactly when that charge is non-zero. The main result identifies a Sugawara element carried, by an explicit intertwiner, to the classical Legendre differential operator. The results are proved for the genus-zero, four-point case.
Read the full abstract

Mathematical Physics

Accessory Parameter of Confluent Heun Equations, Voros Periods and classical irregular conformal blocks

For the Heun differential equation and all of its confluent equations, we derive formal series expansions of the accessory parameters using the Voros periods. We then compare these expansions with the classical conformal blocks recently obtained by Bonelli–Shchechkin–Tanzini, and examine the Zamolodchikov-type conjecture expected to hold between them, allowing for irregular singularities. In particular, as an extension of the previous works of Mironov–Morozov, Piatek–Pietrykowski and Lisovyy–Naidiuk, we provide a detailed prescription for choosing cycles on the spectral curve that yield the Voros period which corresponds to the classical (regular or irregular) conformal blocks through the accessory parameter.
Read the full abstract

Differential Geometry

A comparison theorem with applications to sharp geometric inequalities for submanifolds

In this paper, we derive an explicit expression for the Jacobian determinant of the normal exponential map on a submanifold, establishing a relationship with its ambient counterpart. This formula leads to a new comparison theorem which is closely related to the comparison theorem of Heintze-Karcher. As applications, we obtain a Fenchel-Borsuk-Chern-Lashof-type inequality and a Willmore-Chen-type inequality on closed submanifolds in complete noncompact manifolds with nonnegative curvature and Euclidean volume growth.
Read the full abstract

Combinatorics

Solvable and unsolvable instances of the equal sum partition problem

We consider the equal sum partition problem, motivated by distance magic graph labeling: Given $n,k \in \N$ such that $k\, | \sum_{i=1}^ni$ and a partition $p_1+\cdots+p_k=n$, when is it possible to find a partition of the set $\{1,2,\ldots,n\}$ into $k$ subsets of sizes $p_1,\dots,p_k$, such that the element sum in each subset is the same? A known necessary condition is the slack condition, requiring that for all $j$, placing the largest possible elements in the $j$ smallest sets yields a total sum that is at least what is needed. However, this condition is not sufficient, and known counterexamples exist. This work clarifies the boundary between solvable and unsolvable instances of the problem. We extend the list of unsolvable problem instances satisfying the slack condition by exhibiting infinite families where the $n/k$ ratio is any rational number in the interval $(2,\frac{24}{7})$, and a new criterion for unsolvability. Furthermore, we show that the slack condition is natural, as it is both necessary and sufficient for the fractional relaxation of the problem. Based on this result, we prove that the problem is solvable for the class of linear partitions, where $k$ is fixed, $p_1,\ldots,p_k$ grow linearly with $n$, and where the slack condition holds in a strong sense. We do this by applying a randomized rounding algorithm to a solution of the fractional relaxation of the problem and proving that the algorithm has an exponentially small failure probability.
Read the full abstract

Algebraic Topology · Cross-listed from cs.LG

Geometry-Aware Simplicial Message Passing

The Weisfeiler–Lehman (WL) test and its simplicial extension (SWL) characterize the combinatorial expressivity of message passing networks, but they are blind to geometry, i.e., meshes with identical connectivity but different embeddings are indistinguishable. We introduce the Geometric Simplicial Weisfeiler–Lehman (GSWL) test, which incorporates vertex coordinates into color refinement for geometric simplicial complexes. In addition, we show that (i) the expressivity of geometry-aware simplicial message passing schemes is bounded above by GSWL, and (ii) that there exist parameters such that the discriminating power of GSWL is matched by these schemes on any fixed finite family of geometric simplicial complexes. Combined with the Euler Characteristic Transform (ECT), a complete invariant for geometric simplicial complexes, this yields a geometric expressivity characterization together with an approximation framework. Experiments on synthetic and mesh datasets serve to validate our theory, showing a clear hierarchy from combinatorial to geometry-aware models.
Read the full abstract

Geometric Topology

Poincaré-Hopf Theorem for Isolated Determinantal Singularities

Let $X$ be a projective algebraic $d$-variety endowed with isolated determinantal singularities, and let $\omega$ be a $1$-form on $X$ exhibiting a finite number of singularities (in the stratified sense). Under some technical conditions, we use two generalizations of Poincaré-Hopf index with the goal of proving a Poincaré-Hopf Type Theorem for $X$.
Read the full abstract

Analysis of PDEs

Standing waves for defocusing nonlinear Schrödinger equations with point interaction

We consider standing waves of the nonlinear Schrödinger equation $i\partial_t u = -\Delta_\alpha u + |u|^{p-1}u$ in the defocusing case in dimensions $N=2$ and $N=3$. Here, $-\Delta_\alpha$ denotes the Laplacian with a point interaction. This operator is bounded from below by a negative constant; consequently, unlike in the free case, the associated energy functional admits non-trivial minimizers. We establish existence and uniqueness of standing waves, and prove further qualitative properties, including radial symmetry, positivity, and stability. Moreover, we build an appropriate functional space for the zero-mass case and establish sharp decay estimates in this case.
Read the full abstract

Number Theory

Evaluation of eight different families of cubic Euler sums

We present a study on cubic Euler sums of degree four, five and six, where three different types of denominators $1/k^n$, $1/((2k-1)^n)$ and $1/(k(2k-1))$ will be considered We demonstrate that for all three orders the complete variety of corresponding nonlinear Euler sums belonging to the eight different families can be explicitly calculated in terms of zeta values and polylogarithmic values $Li_4(1/2)$, $Li_5(1/2)$, $Li_6(1/2)$, $Li_6(-1/2)$ and $Li_6(-1/8)$.
Read the full abstract

Numerical Analysis

Two-sided eigenvalue bounds for the Euler-Bernoulli beam

We derive novel guaranteed lower bounds for eigenvalues of the Euler-Bernoulli beam with variable bending stiffness. While the standard finite element Rayleigh-Ritz method automatically yields upper bounds, we obtain lower bounds by employing interpolation error estimates with the explicitly known value of the associated constant. This approach is especially efficient and easy to apply for piecewise constant bending stiffness. For general variable material parameters, we obtain guaranteed lower bounds through an auxiliary beam-bending problem. The first eigenvalue is of primary interest in applications because it represents the critical load that causes buckling of the beam. Our method is, however, suitable also for the higher buckling modes. In addition, it can be applied to the physically more relevant nonlinear Gao beam model with piecewise constant bending stiffness, which has the same first eigenvalue as the classical Euler–Bernoulli beam. The presented numerical experiments illustrate the performance of the proposed eigenvalue bounds, demonstrating their convergence rates.
Read the full abstract

Optimization and Control · Cross-listed from physics.flu-dyn

Topology optimization of two-fluid turbulent heat exchangers: A Darcy flow-based multifidelity approach

This paper presents a topology optimization method for designing two-fluid heat exchangers under turbulent conditions using a Darcy flow-based low-fidelity (LF) model. The LF model is calibrated against a high-fidelity (HF) model based on the Reynolds-averaged Navier-Stokes (RANS) equations to increase the accuracy of predictions for fluid flow and heat transfer characteristics. Since the discrepancies between the LF and HF models can be significant, particularly for pressure drops, a multifidelity topology optimization framework is adopted to leverage the strengths of both models. Using the calibrated LF model, we perform topology optimization for various inlet velocities in the boundary conditions and trade-off parameters in the objective function to obtain diverse optimized designs. The optimized designs are then evaluated using the HF model to assess their performance with higher accuracy. The results demonstrate that the optimized designs significantly improve overall heat transfer coefficients while maintaining manageable pressure drops, achieving up to a 22% higher performance evaluation criterion (PEC) compared to a reference design enhanced by conventional twisted tape insertion. The improvements are attributed to the optimized configurations that promote enhanced fluid mixing and increased surface area for heat exchange, yet maintain streamlined flow paths to minimize pressure losses. Overall, the proposed topology optimization method using the Darcy flow-based LF model proves effective in designing high-performance double pipe heat exchangers, showcasing the potential of the multifidelity approach in overcoming the challenges of optimizing heat exchangers under turbulent flow conditions.
Read the full abstract

All authors Hiroki Kawabe, Kaito Ohtani, Kentaro Yaji, Ryota Fukunishi, Akira Ogawara

Classical Analysis and ODEs

Fourier coefficients of continuous functions with sparse spectrum

Let $(r_k)$ be an increasing sequence and $(w_k)$ a positive sequence. We study the following question: is it true that for every sequence $(a_k)$ satisfying $\sum_{k=0}^\infty |a_k|^2 w_k^2 < \infty$ there exists a function $f\in C(\mathbb{T})$ such that $\hat{f}(2^k) = a_k$ and $\hat{f}(n) = 0$ for $n\notin \cup_k [2^k-r_k,2^k+r_k]$? We show that this is possible if and only if $\sup_{k\in\mathbb{N}}\sum_{n=[\log_2 r_k]}^k w_k^{-2} < \infty$.
Read the full abstract

Group Theory

Concise formulae in groups of non-positive curvature

We show that first-order formulae are concise in acylindrically hyperbolic groups and certain extensions thereof. We study further classes of groups, including Burnside groups, icc groups, groups with the `Big Powers' condition, torus knot groups and more, and prove conciseness for wide classes of formulae. We also explore properties of definable sets in these groups, such as their finiteness, depending on the type of formula considered.
Read the full abstract

Optimization and Control

Solving Constrained Affine Heaviside Composite Optimization Problems by a Progressive IP Approach

This paper discusses the computational resolution and presents numerical results for solving affine combinations of Heaviside composite optimization problems (abbreviated as A-HSCOPs) by a progressive integer programming (abbreviated as PIP) method. The characteristics of these problems are that the Heaviside functions, which appear in the objective and define the constraints, are discontinuous, and their mixed-signed combinations result in the overall objective lacking the matching semicontinuity needed for the optimization and in the feasible set being not necessarily closed. Added to these challenging properties is the nondifferentiability of the inner functions in the composition. In this paper, we propose resolutions to all these challenges by first an approximation to remedy the lack of semicontinuity in the objective and closedness in the constraints, followed by a progressive integer programming approach with successive decomposition to handle the intrinsically discrete nature of the Heaviside function. Convergence to the local optimizers of the given Heaviside optimization problem is established. The effectiveness of the overall solution strategy is supported by extensive computational experiments on the score-based and tree-based multiclass classification problems with precision constraints.
Read the full abstract

All authors Ke Zheng, Junyi Liu, Yurui Wang, Jong-Shi Pang

Operator Algebras

Geometric mean and Lebesgue-type decomposition of completely positive maps

We introduce the geometric mean and the parallel sum of completely positive (CP) maps between von Neumann algebras, based on the Pusz–Woronowicz theory of positive sesquilinear forms. We provide a concrete characterization via a block matrix positivity condition and establish their fundamental properties, including the AM–GM–HM inequality with respect to the CP order. In finite-dimensional settings, our construction is compatible with the Choi–Jamiolkowski correspondence, under which the geometric mean of CP maps corresponds to the Kubo–Ando geometric mean of their Choi matrices. This yields a natural operator-theoretic framework for interpolating quantum channels. As an application, we obtain index-type inequalities for conditional expectations in subfactor theory. Finally, we establish a Lebesgue-type decomposition of CP maps via a parallel sum construction, thereby providing a unified framework that simultaneously generalizes Ando's decomposition of bounded positive operators and Kosaki's decomposition of normal positive functionals on von Neumann algebras.
Read the full abstract

Probability · Cross-listed from cs.LG

Matrix-Decoupled Concentration for Autoregressive Sequences: Dimension-Free Guarantees for Sparse Long-Context Rewards

Sequence-level evaluations in autoregressive Large Language Models (LLMs) rely on highly dependent token generation. Establishing tight concentration bounds for these processes remains a challenge due to two fundamental bottlenecks in existing frameworks: (i) classical inequalities typically separate dependency structures from target sensitivities, leading to a scalar collapse that inflates the variance proxy to a suboptimal $\mathcal{O}(N)$ for sparse terminal rewards; (ii) conversely, while certain spatial methods achieve tighter bounds, they lack the strictly causal filtration required by sequential generation, rendering them inapplicable to the autoregressive setting. To resolve both bottlenecks, we establish a sharp McDiarmid-type inequality for dependent sequences, governed strictly by the exact matrix-vector multiplication of the causal dependency resolvent and the target sensitivity vector. This Matrix-Decoupled Concentration (MDC) framework natively recovers optimal constants for Markov chains and exploits directed $d$-separation to yield order-optimal bounds for causal trees. Crucially, by exactly preserving the coordinate-wise sparsity of rewards within a strictly causal framework, MDC mathematically prevents scalar collapse, guaranteeing a dimension-free $\mathcal{O}(1)$ variance proxy and providing a rigorous mathematical justification for the stability of long-context reasoning.
Read the full abstract

Optimization and Control

A Unified Zeroth-Order Proximal Newton-Type Framework for Composite Optimization

We propose a unified derivative-free proximal Newton-type algorithm framework for solving composite optimization problems formulated as the sum of a black-box function and a known regularization term. We establish the iteration and oracle complexity bounds for the algorithm to attain an $\epsilon$-optimal solution under both nonconvex and strongly convex settings. We also establish its local R-superlinear convergence based on the Dennis–Moré condition, and theoretically address an open problem by showing that the BFGS scheme is more compatible with finite-difference gradient estimators than with smoothing-based ones. Numerical experiments are further presented to demonstrate the efficiency of the proposed method.
Read the full abstract

Optimization and Control

Distortion-minimized de-homogenization for optimization of cell-size distribution in TPMS structures

This paper presents a homogenized topology optimization (TO) method for spatially optimizing cell-size distribution of triply-periodic minimal surface (TPMS) structures, with high accuracy in the optimized structural response after de-homogenization. To achieve this, we introduce a novel de-homogenization technique that directly minimizes the difference between the wavenumbers obtained from the target and actual size distributions. This minimization problem is efficiently solved as a typical Poisson's equation utilizing the discrete cosine transform. We first verify the proposed de-homogenization method through numerical examples, showcasing its capability in significantly reducing the known distortion of the de-homogenized TPMS structures from the conventional periodic modulation (PM) method. Then, we apply the proposed method to a stiffness maximization problem, to demonstrate its effectiveness in improving the structural response compared to the PM method. The proposed method successfully reduced the distortion of the de-homogenized structures compared to the PM method, leading to 0.8% difference in the strain energy compared to the homogenized model, as opposed to 63.6% difference in the PM method. The optimized structure from the proposed method shows a significant improvement in the strain energy by 50.1% compared to the uniform case in the FE analysis on the de-homogenized models, while the PM method results in a significant decrease of 45.8%. The experimental validation shows that the effective stiffness of the optimized structure from the proposed method is 54.2% higher than that of the uniform case, while the PM method results in a significant decrease by 77.3%. These results exhibit the proposed method effectively increases the accuracy of the de-homogenization, thereby maximizing the potential of the homogenized TO for the spatial cell-size optimization of TPMS structures.
Read the full abstract

All authors Hiroki Kawabe, Kaito Ohtani, Yusibo Yang, Musaddiq Al Ali, Kentaro Yaji

Probability

Exponential mixing for the stochastic Allen–Cahn equation with localized white noise

This paper studies the 1D stochastic Allen–Cahn equation on a bounded domain driven by localized white noise. We prove that the associated Markov process admits a unique invariant measure and is exponential mixing. The main challenge lies in the interaction between localized nature of the noise and non-trivial global dynamics of the system. To overcome this, our approach relies on two ingredients from PDE control theory: stabilization for the linearized system and global steady-state controllability for the nonlinear equation. The stabilization result is derived using the weak observability and Fenchel–Rockafellar duality, while the global controllability relies on quasi-static deformations combined with global dynamics.
Read the full abstract

Statistics Theory · Cross-listed from cs.LG

A Fine-Grained Understanding of Uniform Convergence for Halfspaces

We study the fine-grained uniform convergence behavior of halfspaces beyond worst-case VC bounds. For inhomogeneous halfspaces in $\mathbb{R}^d$ with $d\ge 2$, we show that standard first-order VC bounds are essentially tight: even consistent hypotheses can incur population error $\Theta(d\ln(n/d)/n)$, and in the agnostic setting the deviation scales as $\sqrt{\tau\ln(1/\tau)}$ at true error $\tau$. In contrast, homogeneous halfspaces in $\mathbb{R}^2$ exhibit a markedly different behavior. In the realizable case, every hypothesis consistent with the sample has error $O(1/n)$. In the agnostic case, we prove a bandwise, log-free deviation bound on each dyadic risk band via a critical-wedge localization argument. Unioning over bands incurs only a $\ln\ln n$ overhead, and we establish a matching lower bound showing this overhead is unavoidable. Together, these results give a fine-grained and nearly complete picture of uniform convergence for halfspaces, revealing sharp dimensional and structural thresholds.
Read the full abstract

Functional Analysis

Implications of an affirmative solution to the Lindenstrauss Problem

The question regarding the location of Banach spaces inside their biduals has been investigated and answered reasonably satisfactorily in the linear theory of Banach spaces. Thus, for instance, whereas it is known that a dual Banach space is complemented inside its bidual, the space of all null sequences is not! However, the latter space is a Lipschitz retract of its bidual. In his famous paper of 1964, Lindenstrauss asked if every Banach space is a Lipschitz retract of its bidual. In this short note, we show how to relate the Lindenstrauss problem (LP) to certain other important and well-known questions that remain open in the Lipschitz theory of Banach spaces and how these latter questions may be settled in the affirmative under the assumption of (LP) having a positive solution.
Read the full abstract

Dynamical Systems

Deep-Koopman-KANDy: Dictionary Discovery for Deep-Koopman Operators with Kolmogorov-Arnold Networks for Dynamics

Symbolic library – or Koopman dictionary – selection is a fundamental challenge in data-driven dynamical systems. Extended Dynamic Mode Decomposition (EDMD), Sparse Identification of Nonlinear Dynamics (SINDy), and Kolmogorov–Arnold Networks for Dynamics (KANDy) all require the practitioner to commit to a function library at training time; Deep-Koopman Operators avoid this commitment but produce uninterpretable latent observables. We propose Deep-Koopman-KANDy, a structured approach to post-hoc symbolic dictionary readout that combines Deep-Koopman modeling with Kolmogorov-Arnold Networks for Dynamics (KANDy). The encoder and decoder of a Deep-Koopman Operator are replaced with two-layer Kolmogorov–Arnold Networks (KANs), and a level-set construction together with a chain-rule gradient identity exposes the compositional structure of the learned observables in a basis chosen after training. We evaluate the method on the Lorenz system, the Chirikov standard map, the Ikeda map, and the Arnold cat map. On Lorenz it recovers the target dictionary $\{x,y,z,xy,xz\}$ with perfect recall and Jaccard score $0.79\pm0.06$; on the standard map it recovers a low-order Fourier basis matching the analytical structure; on Ikeda – which has no sparse polynomial representation – a misspecified polynomial readout still recovers the correct foliation coordinate $g\approx x^2+y^2$ together with a nontrivial outer function; and on the Arnold cat map – used as a negative control because finite-dimensional Koopman closure is provably impossible – the method fails to find a sparse closure, as expected.
Read the full abstract

Algebraic Geometry

On a generalized Poincaré series of plane valuations

Earlier, there were defined two generalized (“motivic”) versions of the Poincaré series of a collection of plane valuations on the algebra ${\mathcal O}_{{\mathbb C}^2,0}$ of germs of holomorphic functions in two variables. One of them was defined as an integral with respect to the generalized Euler characteristic over the projectivization of the extended semigroup of the collection. One has a natural version of it for valuations on the algebra ${\mathcal E}_{{\mathbb K}^2,0}$ of germs of holomorphic functions in two variables whose Taylor coefficients are from a fixed subfield ${\mathbb K}$ of the field ${\mathbb C}$ of complex numbers. In this setting the usual Poincaré series were computed for one plane curve or divisorial valuation on ${\mathcal E}_{{\mathbb K}^2,0}$. We give equations for the corresponding generalized Poincaré series.
Read the full abstract

Optimization and Control

Verifiable Model-Free Safety Filters via Reinforcement Learning

This paper presents a reinforcement learning approach of a model-free safety filter, drawing inspiration from the framework of model-based Predictive Safety Filters (PSFs). Similar to conventional PSFs, our method adopts a Quadratic Programming (QP) formulation by representing the filter as an unrolled QP solver network. However, unlike existing PSFs that derive QP parameters explicitly from system models, we learn these parameters directly through Deep Reinforcement Learning (DRL), thereby eliminating the dependency on accurate system identification. Furthermore, compared to traditional neural network-based methods, this QP structure allows us to furnish a formal certificate for the persistent safety of the learned filter. Numerical results demonstrate that our method outperforms both conventional model-based PSFs and RL-trained Multi-Layer Perceptron (MLP) baselines in terms of safety guarantees, minimal intervention, and per-step computational load.
Read the full abstract

All authors Bihui Yin, Yiwen Lu, Yuchen Jiang, Yilin Mo