Peer Revenue Past citations are no guarantee of future returns

Papers

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

Mathematics

math Watch Mathematics (sign in to use)

Friday 8 May

Analysis of PDEs

Global Dynamics and Synchronization of Hodgkin-Huxley-Wilson Neural Networks

Hodgkin-Huxley equations as a monumental breakthrough in neuroscience theory had been distilled into many simplified models to study, but this mathematical model itself was not fully investigated in terms of global and asymptotic dynamics due to its strong nonlinearity and higher dimensionality. In this paper a new model called Hodgkin-Huxley-Wilson neural networks is proposed and investigated. This model captures the essential features of the nonlinearity and the conductances of two dominant ionic current channels coupled with the neuron membrane voltage in gated firing and bursting functionality of biological neural networks in the original Hodgkin-Huxley model. Through the uniform and sharp a priori estimates purely by mathematical hard analysis on the solutions of the model equations and then the interneuron differencing equations of gap functions, it is rigorously proved that global solution dynamics are robustly dissipative with a sharp ultimate bound and that exponential synchronization of the Hodgkin-Huxley-Wilson neural networks at a uniform convergence rate occurs if the interneuron coupling strength satisfies an explicitly computable threshold condition. The synchronization result with fractional-power convergence rate is also proved for time-fractional memristive Hodgkin-Huxley-Wilson neural networks.
Read the full abstract

Optimization and Control

Scalable Distributed Stochastic Optimization via Bidirectional Compression: Beyond Pessimistic Limits

In centralized, distributed, and federated learning with stochastic gradients and $n$ workers, it was recently shown that it is infeasible to find an $\varepsilon$-stationary point faster than $\tilde{\Omega}\left(\min\left\{\frac{d \kappa L \Delta}{\varepsilon} + \frac{h L \Delta}{\varepsilon} + \frac{h \sigma^2 L \Delta}{n \varepsilon^2},\; \frac{h \sigma^2 L \Delta}{\varepsilon^2} + \frac{h L \Delta}{\varepsilon}\right\}\right)$ seconds in both homogeneous and heterogeneous settings under standard assumptions: $L$-smoothness, $\sigma^2$-bounded unbiased stochastic gradients, and lower boundedness of the function, i.e., $f(x) \geq f^*$ for all $x \in \mathbb{R}^d$, where $\Delta = f(x^0) - f^*$, $h$ is the computation time, $\kappa$ is the communication speed between the workers and the server, and $d$ is the dimension of the iterates and gradients. This result is pessimistic since it does not allow a complexity in which both $\frac{d \kappa L \Delta}{\varepsilon}$ and $\frac{h \sigma^2 L \Delta}{\varepsilon^2}$ improve with $n$, even when using random sparsification techniques; moreover, this lower bound can be matched by either non-distributed SGD or vanilla Synchronous SGD, which reduces the impact of recent progress in the design of compression-based methods. In this work, we challenge this limitation and propose new compressed methods, Inkheart SGD and M4, and show that under an additional structural assumption, which is necessary due to the lower bound and which does not restrict the class of considered problems, we achieve new state-of-the-art time complexities that break this pessimistic barrier and allow scaling with the number of workers $n$.
Read the full abstract

Numerical Analysis · Cross-listed from cs.LG

Neural Operators as Efficient Function Interpolators

Neural operators (NOs) are designed to learn maps between infinite-dimensional function spaces. We propose a novel reframing of their use. By introducing an auxiliary base-space, any finite-dimensional function can be viewed as an operator acting by composition on functions of the base-space. Through a range of benchmarks on analytic functions of increasing complexity and dimensionality, we demonstrate that NOs can match or outperform standard multilayer perceptrons and Kolmogorov–Arnold Networks in accuracy while requiring significantly fewer parameters and training time. As a real-world application, we apply a two-dimensional Tensorized Fourier Neural Operator (TFNO) to the nuclear chart, learning a correction to state-of-the-art nuclear mass models as a partially observed residual field. A TFNO ensemble reaches a held-out root-mean-square error of 198.2 keV, placing it among the best recent neural-network approaches while retaining high parameter efficiency and short training times. More broadly, these results introduce NOs as a scalable framework for finite-dimensional function interpolation, from analytic benchmarks to structured scientific data.
Read the full abstract

Classical Analysis and ODEs

Orthogonal Polynomials, a Szegő–Verblunsky Theorem and Baxter's Theorem on the Quaternionic Sphere

We introduce a theory of orthogonal polynomials on the unit sphere of the quaternions based on the notion of a $q$-positive measure (which originated in a work of Alpay, Colombo, the second author and Sabadini). The results we extend to this setting include the Szegő recurrences, the Zeros Theorem for orthogonal polynomials, the Szegő–Verblunsky theorem, and Baxter's theorem; to obtain these results, we utilise the Verblunsky coefficients (or Schur parameters) of Alpay, Colombo and Sabadini and a number of established results in the matricial setting. Our approach also requires matrix-valued analogues of Schur's recurrences for the coefficients of a Schur function and of Verblunsky's formula for the moments of a measure, which appear to be new.
Read the full abstract

Rings and Algebras · Cross-listed from cs.DS

Computing bases in Hermite normal form of lattices of integer relations

Given a full column rank $M \in \Z^{\ell \times m}$ and an $F \in \Z^{n \times m}$ we present an algorithm to compute the $n \times n$ basis in Hermite form of the integer lattice comprised of all rows $p \in \Z^{1 \times n}$ such that $pF \in \Z^{1 \times m}$ is in the integer lattice generated by the rows of $M$. The algorithm is randomized of the Las Vegas type, that is, it can fail with probability at most $1/2$, but if fail is not returned it guarantees to produce the correct result. When $M$ is square and $F=I_m$, then the computed basis is the Hermite normal form of $M$, and the algorithm uses about the same number of bit operations as required to multiply together two matrices of the same dimension and size of entries as $M$.
Read the full abstract

Analysis of PDEs · Cross-listed from cs.LG

Training-Induced Escape from Token Clustering in a Mean-Field Formulation of Transformers

Transformers perform inference by iteratively transforming token representations across layers. This layerwise computation has been studied empirically, and recent mean-field theories of Transformer dynamics explain how attention can drive token distributions toward clustering. However, existing analyses treat model parameters as prescribed, leaving open how training reshapes this clustering picture. We study this question in a noisy mean-field Transformer in which only a parameter-linear FFN is trained under $L^2$ regularization. We identify a training-induced phase of the inference dynamics across layers: after clustering in the early layers, the token distribution can leave the clustered regime near the final layers. A single gradient-descent update already produces this escape, and minimizers of the training objective satisfy an exponential turnpike bound under which clustering persists through the interior layers. Our analysis is based on an entropy-regularized interaction energy that captures the clustering bias of attention, for which we establish a double-well landscape and a local Polyak-Łojasiewicz inequality. More broadly, our results point toward a training-aware mean-field theory in which training and inference dynamics of Transformers are treated together.
Read the full abstract

Information Theory

SateBeam: Beam-Aware Radio Map Estimation With Physics-Consistent Parametric Modeling for Unknown Multiple Satellites

Dense low-Earth-orbit (LEO) networks increasingly reuse spectrum across moving satellites and overlapping beams, making a spatial radio map essential for interference diagnosis and coordination. Constructing such a map from sparse ground power measurements is difficult because the active satellite subset and its cardinality are unknown, candidate footprints can be highly coherent, and beam pointing, width, effective power, Doppler, and background must be inferred jointly. Known ephemerides do more than enumerate visible candidates. Propagated positions determine candidate-specific ranges, viewing geometry, and spatial footprints, while radial velocities induce distinct time-frequency signatures. Their joint spatial–Doppler structure rejects candidates whose geometry or time–frequency signature conflicts with the observations before continuous beam inference. SateBeam therefore screens all candidates through Doppler-resolved nonnegative fitting. GPU-accelerated grid correlations initialize continuous off-grid refinement; a Gamma-likelihood Bayesian information criterion (BIC) selects model cardinality, and backward deletion and exchange repair greedy support errors. Across paired real-ephemeris scenarios, SateBeam recovers the exact support in every case and reduces independent-query RMSE from $0.655$ to $0.050$ dB relative to orthogonal matching pursuit, with consistent advantages under model mismatch and larger candidate sets. The code is available at https://github.com/UNIC-Lab/SateBeam.
Read the full abstract

All authors Xiucheng Wang, Nan Cheng, Zhisheng Yin, Conghao Zhou, Ruijin Sun

Mathematical Physics · Cross-listed from hep-th

Characterizing gapped phases by smeared boundary conformal field theories: Duality in unusual ordering with spontaneously broken generalized symmetries

We study the classification of the gapped phases or massive renormalization group (RG) flows dual to the massless RG flows under changing the sign of the coupling constants. Whereas our formalism is based on combining Higgs- or Nambu-Goldstone-type arguments with Cardy's smeared boundary conformal field theories (SBCFTs), several puzzling structures arise. More specifically, the established Higgs or Nambu-Goldstone type arguments on the duality imply that the natural basis for the gapped states should be constructed from a set of smeared Ishibashi states, which are unphysical in boundary critical phenomena. Hence, the module of the gapped phases can be outside of that of boundary critical phenomena, whereas one can still calculate characterizing quantities by applying SBCFTs to the models. For example, we demonstrate that the massive RG flow dual to the massless RG flow from the tricritical Ising model to the Ising model, one of the simplest massless RG flows, has this unusual structure. This can be regarded as a quantum field-theoretic analogue of order-disorder coexistence in lattice models. More generally, the resultant gapped phases usually spontaneously break non-group-like symmetry (or noninvertible symmetry). Our work provides systematic quantum field theoretic descriptions of such unusual phases with spontaneous symmetry breaking of non-group-like (or noninvertible) symmetries.
Read the full abstract

Differential Geometry

A Diffeological Construction of Singer's Universal Connection

We provide a rigorous construction of I.M. Singer's universal connection, a natural connection on a bundle of paths associated to any manifold, using the theory of diffeology. Furthermore, we generalize the universal connection to the diffeological setting, which enables the reconstruction of diffeological principal bundles with connections from their holonomy representations. We show that any two diffeological bundle-connection pairs with conjugate holonomy representations must be equivalent in a certain sense. These constructions are functorial in that, ultimately, our results can be summarized as an equivalence of categories between the so-called holonomy category and the category of diffeological bundle-connection pairs.
Read the full abstract

Algebraic Geometry · Cross-listed from hep-th

Genus drop involving non-hyperelliptic curves in Feynman integrals

For both theoretical and phenomenological studies, it is important to analyze the function types of Feynman integrals. The phenomenon of genus drop between different representations of hyperelliptic Feynman integrals was discussed in \cite{Marzucca2024Genusdrop}. In this paper, we reformulate the extra-involution mechanism of \cite{Marzucca2024Genusdrop} as a special case of an unramified double covering between algebraic curves, and show that this covering mechanism also explains genus drops accompanied by a curve-type change from non-hyperelliptic to hyperelliptic for a class of three-loop Feynman diagrams. We also demonstrate that within a specific framework, the origin of the discrete spacetime symmetry that leads to the genus drop in hyperelliptic cases is manifest. This work also points out that there exist non-hyperelliptic Feynman integrals that exhibit no apparent genus drop.
Read the full abstract

Algebraic Topology · Cross-listed from stat.ML

TopoFisher: Learning Topological Summary Statistics by Maximizing Fisher Information

Persistence diagrams provide stable, interpretable summaries of geometric and topological structure and are useful for simulation-based inference when low-order statistics miss key information. Yet persistence-based pipelines require hand-chosen filtrations, vectorizations, and compressors, typically without an objective tied to parameter uncertainty. We introduce TopoFisher, a differentiable persistent-homology pipeline that learns topological summaries by maximizing local Gaussian Fisher information. Using simulations near a fiducial parameter, TopoFisher optimizes trainable filtrations, diagram vectorizations, and compressors without posterior samples or supervised regression targets, while retaining stable topological inductive bias. We also give sufficient regularity conditions for the log-determinant Fisher loss to be locally Lipschitz in trainable parameters. Controlled experiments on noisy spirals and Gaussian random fields, where total Fisher information is known, show that TopoFisher recovers much of the available information and outperforms fixed topological vectorizations. Our main results are on weak gravitational lensing, a high-dimensional non-Gaussian cosmological field-inference problem. Learned topological summaries reach higher Fisher information than state-of-the-art cosmological summaries and approach an unconstrained Information Maximising Neural Network baseline with up to $\sim80\times$ fewer parameters. The learned filtrations also generalize better: under simulator shift from lognormal to LPT-based maps it retains most Fisher information, while the neural baseline drops, and in neural posterior estimation they give tighter constraints than the neural baseline, and of state-of-the-art cosmological summaries. These results support Fisher-based topological optimization as a robust, parameter-efficient front end for simulation-based inference.
Read the full abstract

All authors Matteo Biagetti, Mathieu Carrière, Francesco Conti, Enrico Maria Ferrari, Sven Heydenreich, Karthik Viswanathan

Information Theory · Cross-listed from cs.LG

Future Validity is the Missing Statistic: From Impossibility to $\Phi$-Estimation for Grammar-Faithful Speculative Decoding

Grammar-constrained generation is often combined with local vocabulary masking and speculative decoding, but the resulting sampling law is not the grammar-conditional distribution users usually intend. We show that any speculative decoder with local mask access, Leviathan rejection, and rollback soundness samples from the locally projected distribution $\mu^{\mathrm{proj}}$ rather than the grammar-conditional distribution $\mu^\star$. This extends the GAD impossibility result to speculative decoding; on Dyck grammars with Qwen3-8B, the total-variation gap can reach 0.996. We identify the future-validity function $\Phi_t(y)=\Pr_p[\mathrm{valid\ completion}\mid y]$ as the missing correction statistic. The target distribution is a Doob transform of the base model with $h=\Phi$, while local masking corresponds to setting $h$ to one. With exact $\Phi$, our oracle decoder FVO-Spec samples exactly from $\mu^\star$; with approximate $\Phi$, we bound the resulting total-variation error. Because exact future validity is hard for general context-free grammars, we evaluate estimator hierarchies on tractable Dyck and finite JSON languages. OneStep reduces Dyck TV by 14% with under 1% throughput overhead, exact dynamic programming reduces it by 97%, and finite-language correction closes JSON gaps to numerical precision. All fidelity claims are scoped to enumerable grammars and token tries.
Read the full abstract

All authors Wenhua Nie, Zijie Meng, Kun Zou, Zheng Lin, Ziwei Li, Haoran Zheng, Jyh-Shing Roger Jang, Hao Zhang

Spectral Theory

Quantum Ergodicity on large hyperbolic surfaces for local and pseudolocal operators

We prove a quantum ergodicity theorem for sequences of closed hyperbolic surfaces converging to the Poincaré disc in the Benjamini-Schramm sense. Assuming a uniform lower bound on the injectivity radius and a spectral gap, we establish vanishing of quantum variance on fixed spectral windows for a class of observables that contains differential operators and finite-propagation smooth operators. This generalises a result of Le Masson and Sahlsten from scalar observables to both local and 'pseudolocal' operator settings.
Read the full abstract

Analysis of PDEs

Dissipative structures in compressible inviscid fluids

This note aims at the following problem. In an ideal density dependent fluid system, is the total energy dissipated on shock type discontinuities? To this end, we study the local energy balance for weak solutions to the isentropic compressible Euler system and derive fine properties of the defect measure. This is done by a careful analysis of the small scale properties of the solutions at the shock discontinuity. By means of the same technique, we also consider the inhomogeneous incompressible case, and, comparing the result, we confirm the general principle that the accumulation of the total energy on codimension one singular structures is a feature that distinguishes compressible and incompressible models.
Read the full abstract

Differential Geometry

Virasoro extensions for diffeomorphisms with breaks

We study homeomorphisms of the circle that are smooth diffeomorphisms away from a finite set of $n$ points. These "broken diffeomorphisms" do not form a Lie group, but instead naturally assemble into a Lie groupoid. We construct an explicit nontrivial $n$-dimensional central extension of this groupoid, which restricts to the classical Virasoro group when confined to smooth diffeomorphisms. We further describe the associated "broken Virasoro" algebroid, defined as a nontrivial $n$-dimensional central extension of the Lie algebroid of vector fields on the circle that are smooth except at $n$ points. This construction generalizes the Virasoro algebra. As a byproduct, we analyze a related setting on an interval: we construct a nontrivial central extension of the Lie algebra of vector fields vanishing at the endpoints, together with the corresponding central extension of the group of diffeomorphisms fixing the endpoints. We also describe the associated Lie algebroid and groupoid obtained by allowing the endpoints to vary.
Read the full abstract

Combinatorics

On uniform Higmanian association schemes

An imprimitive symmetric indecomposable association scheme of rank $5$ is said to be Higmanian. In the present paper, we prove a necessary and sufficient condition for a Higmanian association scheme with two nontrivial parabolics to be uniform. We also provide examples of uniform Higmanian Cayley schemes.
Read the full abstract

Classical Analysis and ODEs

Hermite expansions of functions from the weighted Hardy class

In this paper, we analyze a function space consisting of functions for which both the function and its Fourier transform exhibit Gaussian decay together with exponential growth governed by suitable weight functions. First, we examine logarithmic-type weights, in which case these function spaces are equivalent to Pilipović spaces. In this setting, we establish a decay estimate for the Hermite coefficients of functions. Furthermore, by combining these estimates with the asymptotic behavior of Hermite functions, we prove a decay rate for solutions to the harmonic oscillator Schrödinger equation. Second, we consider a class of weights and prove the exponential decay of the Hermite projection operators on these spaces by analyzing Laguerre expansions and the short-time Fourier transform. Additionally, we revisit the subcritical Hardy uncertainty principle and obtain a partial improvement toward a conjecture posed by Vemuri.
Read the full abstract

Combinatorics

On separability of Tatra association schemes

A Tatra association scheme is an association scheme arising from a symmetric bilinear form defined on the equivalence classes of nonzero $2$-dimensional vectors modulo some subgroup of the multiplicative group of a finite field. In the present paper, we prove that every such association scheme is $4$-separable, i.e., it is determined up to isomorphism by the tensor of its $4$-dimensional intersection numbers.
Read the full abstract

Optimization and Control · Cross-listed from cs.GT

The Endogeneity of Miscalibration: Impossibility and Escape in Scored Reporting

An agent's probability report is paid for twice: by a strictly proper scoring rule, and by an approval rule for the decision it triggers. In this classical decision-coupled setting, non-affine approval is known to defeat truthful reporting. We show the conflict is endogenous: when feasible, the welfare-maximizing approval rule is never affine. The distortion, however, is predictable and can be designed around. There is a reserve report at which pretending to be the marginal type costs exactly the approval prize. Approving at or above the reserve screens types perfectly under every strictly proper score, and the reserve does not depend on the type distribution. A Lipschitz rule with a single kink attains first-best exactly; under strict feasibility no continuously differentiable rule does. The binding constraint is steepness, not smoothness. First-best is attainable within a slope budget if and only if the budget is at least the critical slope: the steepest chord of the pretending cost up to the reserve. Below it the welfare loss is cubic in the shortfall. Where the pretending cost is convex up to the reserve, as for Brier, log and power scores, the critical slope is closed-form. The instances are AI-agent oversight and marketplace operation.
Read the full abstract

Analysis of PDEs

The Cauchy problem for the improved Boussinesq equation with spatially quasi-periodic initial data

We study the Cauchy problem for the improved Boussinesq equation \[ u_{tt}-u_{xx}-u_{xxtt}-(u^2)_{xx}=0 \] on the real line with spatially quasi-periodic initial data. For a non-resonant frequency vector $\omega\in\mathbb R^\nu$, we prove local existence and uniqueness of classical spatially quasi-periodic solutions with the same frequency vector $\omega$ in two Fourier-side classes. First, for exponentially decaying initial Fourier coefficients, we obtain a spatially quasi-periodic solution whose Fourier coefficients remain exponentially decaying on an explicit time interval. Second, for initial Fourier coefficients $c(n)$ and $d(n)$ satisfying the polynomial decay $ |c(n)|+|d(n)|\lesssim (1+|n|)^{-r}, \; r>\nu+2, $ we prove that the corresponding spatially quasi-periodic solution preserves the same polynomial decay rate as the initial data. We also extend these results to the nonlinearity $u^p$ with integer $p \geq 3$.
Read the full abstract

Classical Analysis and ODEs

Dyadic Martingale Transforms and Weighted Walsh-Carleson Operators

We study weighted Walsh–Carleson maximal operators arising from dyadic martingale transforms associated with Walsh–Fourier partial sums. For weights satisfying a uniform dyadic variation condition and a uniform bound at the top dyadic scale, we prove weak type~$(1,1)$ estimates for the corresponding maximal operators along subsequences. We also give divergence criteria in terms of the behavior of the weights near the top dyadic scale and, under suitable admissibility assumptions, relate these criteria to explicit ratio conditions. As applications, we obtain results on matrix transforms of Walsh–Fourier partial sums, including de la Vallée Poussin means, Cesàro means with varying parameters, Nörlund logarithmic means, and general Nörlund means. In particular, we prove a Walsh–Paley analogue of the Leindler–Tandori theorem and establish everywhere divergence results for several summability methods.
Read the full abstract

Rings and Algebras

Commutativity preserving mappings in Banach algebras

Let $A$ and $B$ be unital complex Banach algebras having no quotients isomorphic to $\mathbb{C}$ or $M_2(\mathbb{C})$. Assume additionally that $B$ is semisimple. If a surjective additive mapping $\Phi\colon A\to B$ satisfies $[\Phi(x^2),\Phi(x)] = 0$ for all $x\in A$, then there exist a surjective direct sum of an additive homomorphism and an additive anti-homomorphism $\Psi\colon A\to B$, an invertible element $\lambda\in\mathcal{Z}(B)$, and an additive mapping $\zeta\colon A\to\mathcal{Z}(B)$ such that $\Phi(x)=\lambda\Psi(x)+\zeta(x)$ for all $x\in A$.
Read the full abstract

All authors M. Brešar, G. M. Escolano, A. Peralta, A. R. Villena

Combinatorics

Benjamini-Schramm convergence and subtrees of trees

In this paper, we study the asymptotic behaviour of the number of subtrees and the subtree density for a sequence of trees that converges in the Benjamini-Schramm sense. Benjamini-Schramm convergence, also called local weak convergence, describes the local behaviour of a sequence of graphs. Here we show that for a Benjamini-Schramm-convergent sequence of trees, the subtree entropy, i.e., the logarithm of the number of subtrees divided by the order, converges to a constant depending only on the limit. The same holds true for the subtree density, i.e., the probability of a uniformly random vertex being contained in a uniformly random subtree, provided that long paths are ruled out in the limit. Related to this, we show that the subtree density and the average subtree entropy are dense in different parts of the unit interval $[0,1]$ for both general trees and series-reduced trees.
Read the full abstract

Numerical Analysis · Cross-listed from cs.LG

Direction-Preserving Number Representations

Low-precision number formats are widely used in modern machine learning systems due to their efficiency. Accurate direction representation is key to the accuracy of vector operations. This work precisely explores the extent to which the direction of a vector can be represented by selecting its scalar elements from a common finite alphabet of a given size. This is standard practice in machine learning, where low-precision significands may be narrow-width floating-point or integer values. A geometric framework is introduced for analyzing the directional coverage of such product-structured codes. This work analytically quantifies the suboptimality gap between such product-structured codes and spherical codes for the vector as a whole, in both low and asymptotically high dimensions. Furthermore, within the product code class, it is proven that the standard formats of two's complement, fixed-point, and floating-point are suboptimal, again with quantified gap, pointing to the potential to develop new scalar number formats. Such scalar alphabets are numerically optimized across multiple block dimensions for directional coverage, including the dimension used in NVIDIA's NVFP4 format. Experimental results are presented comparing the performance of standard formats and the optimized alphabet. We find that for four bits, NVIDIA's choice of E2M1 closely approximates the optimized alphabet, providing a geometric explanation for its strong performance in low-precision machine learning workloads and an analytical understanding of the link between that superiority and block size. We provide open-source formal proofs in Lean for the theorems in this work, along with the experimental code and the optimized alphabets obtained.
Read the full abstract

Mathematical Physics

Superintegrability in the interaction of two particles with spin: First-order pseudo-scalar integrals of motion

In a recent work, we initiated a research program aimed at the systematic investigation of quantum superintegrable systems describing the interaction of two non-relativistic spin-$1/2$ particles in three-dimensional Euclidean space. In that study, we classified all such superintegrable systems admitting additional first-order scalar integrals of motion. In the present paper, we continue this program by focusing on systems that admit additional pseudo-scalar integrals of motion. Starting from the most general rotationally invariant Hamiltonian for two interacting spin-$1/2$ particles, we construct the most general pseudo-scalar operator of differential order at most one in the form of a matrix polynomial in the momenta. Imposing the commutativity of this operator with the Hamiltonian leads to a system of determining equations. By solving these equations, we obtain a complete classification of Hamiltonians in the considered class admitting additional pseudo-scalar integrals of differential order at most one, and determine these integrals explicitly. The resulting classification provides new families of spin-dependent Hamiltonians with additional pseudo-scalar symmetries. These systems enrich the class of integrable models relevant to nucleon–nucleon interactions and contribute to the broader program of classifying superintegrable quantum systems with spin. For selected cases, we further construct the associated polynomial symmetry algebras generated by the integrals of motion, providing additional insight into the algebraic structure of the systems.
Read the full abstract

Algebraic Geometry

Root bounds of vertical systems using tropical geometry

Sparse polynomial systems with vertical coefficient dependencies arise naturally when describing the critical points of optimization problems and, when augmented with linear forms, the steady states of chemical reaction networks. Moreover, any polynomial system is the specialization of such a parametrized system. We prove that the generic number of complex zeros of an augmented vertically parametrized system is the tropical intersection number of a tropical linear space and a classical linear space. In the special case when the matroid of the tropical linear space is cotransversal, we express this number as a mixed volume. We also obtain bounds on the maximal number of positive zeros, which is often the significant number in applications. We derive lower bounds from the number of intersections between positive tropicalizations, and when the positive zeros have toric structure, we provide upper bounds that are simpler and in some cases smaller than the generic root count. The resulting algorithms are implemented in Julia.
Read the full abstract

All authors Elisenda Feliu, Paul Alexander Helminck, Oskar Henriksson, Yue Ren, Benjamin Schröter, Máté L. Telek

Algebraic Geometry

Relative $\mathbb{A}^1$-Contractibility of Smooth Schemes

We study smooth morphisms $f \colon X \to S$ that are $\mathbb{A}^1$-contractible in the unstable $\mathbb{A}^1$-homotopy category $\mathcal{H}(S)$. For base schemes $S$ of finite Krull dimension, we show that $\mathbb{A}^1$-contractibility is a fiberwise property: such a morphism is $\mathbb{A}^1$-contractible if and only if all its geometric fibers are $\mathbb{A}^1$-contractible. We apply this criterion to $\mathbb{A}^n$-fiber spaces, obtaining a geometric description of their $\mathbb{A}^1$-contractibility in terms of local factorizations as towers of torsors under vector bundles, building on results of Asanuma. In low relative dimensions, we establish rigidity results. In relative dimension $1$, $\mathbb{A}^1$-contractible morphisms over normal bases are precisely Zariski locally trivial $\mathbb{A}^1$-bundles. In relative dimension $2$, we show that over bases with characteristic zero residue fields, $\mathbb{A}^1$-contractible morphisms are $\mathbb{A}^2$-fiber spaces, and we obtain Zariski local triviality under additional hypotheses on the base. We also exhibit counterexamples in positive and mixed characteristic and formulate open problems concerning the existence of exotic $\mathbb{A}^1$-contractible surfaces.
Read the full abstract

Complex Variables

Meromorphic functions and linearization phenomena in partial differential equations

In this paper, we investigate meromorphic solutions of certain nonlinear partial differential equations in several complex variables involving differential and functional operators. Let $f$ be a non-constant meromorphic function in $\mathbb{C}$, $g$ an entire function in $\mathbb{C}^n$, and $h(z)=f(z_1+z_2+\ldots+z_n)$. We study the equations \begin{align*} \frac{\partial h(z)}{\partial z_i}=a G^g_{h}(z)+bh(z)+c\;\;\text{and}\;\;\frac{\partial h(z)}{\partial z_i}=a(z)G^g_{h}(z)+b(z)h(z)+c(z), \end{align*} where $z\in\mathbb{C}^n$, $i\in\{1,2,\ldots,n\}$, $a(\neq 0), b, c\in\mathbb{C}$ or $a(z)(\not\equiv 0), b(z),c(z)$ are polynomials in $\mathbb{C}^n$, and $G^g_h(z)=h(g(z),g(z),\ldots,g(z))$. The results obtained in the paper, extend previous studies on meromorphic solutions of functional-differential equations to the setting of several complex variables, and further illustrate the rigidity imposed by value distribution properties on nonlinear functional equations.
Read the full abstract

Optimization and Control

Robust stochastic first order methods in heavy-tailed noise via medoid mini-batch gradient sampling

We consider a first order stochastic optimization framework where, at each iteration, $K$ independent identically distributed (i.i.d.) data point samples are drawn, based on which stochastic gradients can be queried. We allow gradient noise to be heavy-tailed, with possibly infinite variances. For the considered heavy-tailed setting, many algorithmic variants have recently been proposed based on gradient clipping or other nonlinear operators (e.g., normalization) applied over noisy gradients. In this paper, we take an alternative approach and propose a novel stochastic first order method dubbed Robust Stochastic Gradient Descent with medoid mini-batch gradient sampling, R-SGD-Mini for short. The core idea of R-SGD-Mini is to split the $K$-sized data batch into $M$ distinct data chunks, form for each chunk the stochastic gradient, and update the solution estimate with respect to the stochastic gradient direction of the chunk that is medoid of gradients of all data-chunks. Under a general class of symmetric heavy-tailed gradient noises and a standard non-convex setting, we establish explicit bounds on the expected time-averaged squared gradient norm. More precisely, we show that the latter quantity converges at rate $\mathcal{O}(T^{-1})$ to a small neighborhood of zero; we explicitly characterize this neighborhood in terms of noise and algorithm's parameters. Moreover, if the time horizon is known in advance, we establish the rate of $\mathcal{O}(T^{-\frac{1}{2}}).$ Furthermore, when clipping is incorporated, we obtain convergence guaranties in the high-probability sense and recover the same rate. Experimental results indicate that R-SGD-Mini and its clipped variant consistently perform favorably compared to SGD, clipped SGD and Median-of-Means based methods.
Read the full abstract

Optimization and Control

Distributed Seeking for Fixed Points of Biased Stochastic Operators: A Communication-Efficient Approach

This paper investigates the distributed fixed point seeking problem of sum-separable stochastic operators over the multi-agent network. Based on inexact Krasnosel'skiĭ–Mann iterations, the communication-efficient distributed algorithm is proposed under the relaxed growth bias and variance conditions, generalizing traditional unbiased and bounded additive variance assumptions. To enhance communication efficiency, we integrate communication compression and dynamic period skipping techniques, particularly adopting a unified compressor that allows both relative and absolute compression errors. By introducing a surrogate function for general non-contractive and contractive operators, we establish convergence guarantees of the distributed fixed point iteration, achieving among the first theoretical unifications with distributed non-convex optimization algorithms. Finally, numerical simulations validate the effectiveness of the theoretical results.
Read the full abstract

All authors Fan Li, Lei Xu, Xinlei Yi, Guanghui Wen, Yang Shi, Tao Yang