Peer Revenue Returns subject to peer review

Papers

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

Mathematics

math Watch Mathematics (sign in to use)

Wednesday 18 March

Representation Theory

Representations of categories of finite relational structures and associated endomorphism monoids

We develop a unified representation theory for the categories of finite subsets and relation-preserving maps of highly homogeneous relational structures classified by Cameron. For any commutative coefficient ring $k$, we extend the classical Dold-Kan correspondence to this setting, with the sole exception of the category $\mathrm{FA}$, and prove that finitely generated representations are noetherian (resp., artinian) when $k$ is noetherian (resp., artinian). When $k$ is a field, we obtain a precise structural description of these representation categories. We classify irreducible representations, showing that canonical quotients of indecomposable standard modules are either irreducible (the regular case) or has length 2 (the singular case). In the case that $k$ has characteristic 0, we establish a (direct sum or triangular) decomposition into a singular component governed by classical Dold-Kan theory and a regular component exhibiting semisimplicity or representation stability. Finally, we establish a monoidal generalization of Artin's reconstruction theorem for topological groups, proving an equivalence between uniformly continuous representations of infinite topological transformation monoids and sheaves on the associated categories of finite subsets. In the special case where the transformation monoid is a permutation group, our result recovers Artin's theorem.
Read the full abstract

2603.17371 · 18 Mar · Two major revisions. 1) In the current version we assume that the set $\Omega$ is equipped with an underlying relation (not restricted to the orbital relation given by a permutation group), so that more examples can be included. 2) Classifications of irreducible representations are extended from characteristic 0 to all characteristics

Probability · Cross-listed from cs.LG

Variational Kernel Design for Internal Noise: Gaussian Chaos Noise, Representation Compatibility, and Reliable Deep Learning

Internal noise in deep networks is usually inherited from heuristics such as dropout, hard masking, or additive perturbation. We ask two questions: what correlation geometry should internal noise have, and is the implemented perturbation compatible with the representations it acts on? We answer these questions through Variational Kernel Design (VKD), a framework in which a noise mechanism is specified by a law family, a correlation kernel, and an injection operator, and is derived from learning desiderata. In a solved spatial subfamily, a quadratic maximum-entropy principle over latent log-fields yields a Gaussian optimizer with precision given by the Dirichlet Laplacian, so the induced geometry is the Dirichlet Green kernel. Wick normalization then gives a canonical positive mean-one gate, Gaussian Chaos Noise (GCh). For the sample-wise gate used in practice, we prove exact Gaussian control of pairwise log-ratio deformation, margin-sensitive ranking stability, and an exact expected intrinsic roughness budget; hard binary masks instead induce singular or coherence-amplified distortions on positive coherent representations. On ImageNet and ImageNet-C, GCh consistently improves calibration and under shift also improves NLL at competitive accuracy.
Read the full abstract

Differential Geometry

Weakly stable irreducible Yang-Mills fields over $S^4$

Addressing Yau's conjecture (Problem 117) on $S^4$, we investigate the self-duality of weakly stable Yang-Mills fields under the assumption of irreducibility. For structure groups with a simple Lie algebra, we prove that any weakly stable irreducible connection must be either self-dual or anti-self-dual. Furthermore, we demonstrate that if the Lie algebra admits a non-trivial abelian center, no irreducible Yang-Mills fields can exist over $S^4$.
Read the full abstract

Analysis of PDEs

Stochastic Compressible Euler Equations with Frictional Damping: Existence of $L^\infty$ Martingale Solutions and Asymptotic Porous Medium-Like Behavior

We study the one-dimensional isentropic compressible Euler equations with linear (frictional) damping, subject to multiplicative, white-in-time stochastic forcing. The system is posed on a bounded interval with $L^\infty$ initial data and Dirichlet boundary conditions imposed on the momentum. We establish the global-in-time existence of $L^\infty$ martingale solutions that satisfy an appropriate entropy inequality. Then, we analyze the long-time behavior of these solutions and show that, under suitable assumptions on the noise, they converge almost surely and exponentially fast to a constant steady state of the system. The limiting density is well-approximated by the asymptotic solution of the deterministic porous medium equation, while the momentum exhibits the asymptotic behavior predicted by Darcy's law. The analysis in the stochastic setting is delicate, as temporal white-noise perturbations can significantly influence the long-time statistics of the solution. Our approach hinges on deriving sharp moment estimates for the entropy, which enable us to quantify and ultimately prove the decay of stochastic effects. To the best of our knowledge, this work provides the first rigorous pathwise convergence result for the long-time behavior of solutions to the stochastic isentropic compressible Euler equations with linear damping.
Read the full abstract

All authors Rongyi Dai, Jeffrey Kuan, Krutika Tawri, Sunčica Čanić, Konstantina Trivisa

Optimization and Control

Decision-Scaled Scenario Approach for Rare Chance-Constrained Optimization

Chance-constrained optimization is a suitable modeling framework for safety-critical applications where violating constraints is nearly unacceptable. The scenario approach is a popular solution method for these problems, due to its straightforward implementation and ability to preserve problem structure. However, in the rare-event regime where constraint violations must be kept extremely unlikely, the scenario approach becomes computationally infeasible due to the excessively large sample sizes it demands. We address this limitation with a new yet straightforward decision-scaling method that relies exclusively on original data samples and a single scalar hyperparameter that scales the constraints in a way amenable to standard solvers. Our method leverages large deviation principles under mild nonparametric assumptions satisfied by commonly used distribution families in practice. For a broad class of problems satisfying certain practically verifiable structural assumptions, the method achieves a polynomial reduction in sample size requirements compared to the classical scenario approach, while also guaranteeing asymptotic feasibility in the rare-event regime. Numerical experiments spanning finance and engineering applications show that our decision-scaling method significantly expands the scope of problems that can be solved both efficiently and reliably.
Read the full abstract

All authors Jaeseok Choi, Anand Deo, Constantino Lagoa, Anirudh Subramanyam

Mathematical Physics

On the conservation of specific energy and entropy in infinite anharmonic systems

We work with infinite, closed, translation-invariant, finite-range lattice systems with "unbounded classical spins", also known as anharmonic crystals, under assumptions close to those used by Lanford, Lebowitz and Lieb (J. Stat. Phys., 1977); among other conditions, the pinning dominates the interaction. In this context, we prove conservation of the specific energy and specific entropy under the time evolution, and we discuss their relation to approach to thermal equilibrium, paralleling known results in the theory of quantum spin systems, where noncommutativity, as opposed to lack of compactness, is the main source of difficulties.
Read the full abstract

Optimization and Control · Cross-listed from eess.SY

Distributed Equilibrium-Seeking in Target Coverage Games via Self-Configurable Networks under Limited Communication

We study a target coverage problem in which a team of sensing agents, operating under limited communication, must collaboratively monitor targets that may be adaptively repositioned by an attacker. We model this interaction as a zero-sum game between the sensing team (known as the defender) and the attacker. However, computing an exact Nash equilibrium (NE) for this game is computationally prohibitive as the action space of the defender grows exponentially with the number of sensors and their possible orientations. Exploiting the submodularity property of the game's utility function, we propose a distributed framework that enables agents to self-configure their communication neighborhoods under bandwidth constraints and collaboratively maximize the target coverage. We establish theoretical guarantees showing that the resulting sensing strategies converge to an approximate NE of the game. To our knowledge, this is the first distributed, communication-aware approach that scales effectively for games with combinatorial action spaces while explicitly incorporating communication constraints. To this end, we leverage the distributed bandit-submodular optimization framework and the notion of Value of Coordination that were introduced in [1]. Through simulations, we show that our approach attains near-optimal game value and higher target coverage compared to baselines.
Read the full abstract

All authors Jayanth Bhargav, Zirui Xu, Vasileios Tzoumas, Mahsa Ghasemi, Shreyas Sundaram

Number Theory

Density results for $r$-gaps between zeros of the Riemann zeta-function

Let $0<\gamma_1\leq \gamma_2\leq \ldots$ denote the positive ordinates of the non-trivial zeros of the Riemann zeta-function. A result first announced by Selberg states that there exist absolute constants $\Theta, \vartheta>0$ such that for each $r\in \mathbb{N}$, \[ \limsup_{n\to \infty}\frac{\gamma_{n+r}-\gamma_n}{2\pi r/\log \gamma_n}\geq 1+\frac{\Theta}{r^\alpha} \qquad \text{and}\qquad \liminf_{n\to \infty}\frac{\gamma_{n+r}-\gamma_n}{2\pi r/\log \gamma_n}\leq 1-\frac{\vartheta}{r^\alpha} \] where $\alpha$ may be taken as $2/3$, or as $1/2$ if one assumes the Riemann hypothesis. This was recently proved by Conrey and Turnage-Butterbaugh under RH and by Inoue unconditionally. We prove that in fact a positive proportion of $r$-gaps are large (and small) to the above extent, and we provide explicit estimates for the sizes and proportions of these gaps. In the case $r=1$, this quantitatively improves an unconditional result of Simonič, Trudgian and Turnage-Butterbaugh.
Read the full abstract

Combinatorics

Identities that represent powers of positive integers using multinomial coefficients

In this paper, we show combinatorial identities that represent powers of positive integers using multinomial coefficients, which do not come from the multinomial theorem and the multinomial Vandermonde's convolution.

Statistics Theory · Cross-listed from stat.ME

Empirical Likelihood Inference for Sen and Sen–Shorrocks–Thon Indices

The Sen index and Sen-Shorrocks-Thon (SST) index are widely used measures of poverty indices. Developing reliable inference for these measures enables us to compare these measures in different populations of interest in an effective way. It is important to construct confidence intervals for the Sen index and SST index, which provide better coverage probability and shorter interval length. Motivated by this, we discuss empirical likelihood (EL) and jackknife empirical likelihood (JEL) based inference for the Sen index. To derive a JEL-based confidence interval for the Sen and SST indices, we propose a new estimator for the Sen index using the theory of U-statistics and examine its properties. The large sample properties of the EL and JEL ratio statistics are studied. We also discuss EL and JEL-based inference for the Sen-Shorrocks-Thon (SST) index. The finite sample performance of the EL and JEL-based confidence intervals of both Sen and SST indices is evaluated through a Monte Carlo simulation study. Finally, we illustrate our methods using individual-level data from the Panel Study of Income Dynamics (PSID) survey from the US as well as Indian household level income data for different states sourced from the Consumer Pyramids Household Survey (CPHS).
Read the full abstract

Analysis of PDEs

Reformulation and Interpretation of the Regularity Criterion for 3D NSE Based on Finitely Many Observations

We revisit and sharpen a recent observable regularity criterion for the three-dimensional Navier-Stokes equations on the periodic cube by requiring only finitely many measurements of the flow on a given time interval. Two data models are treated: (i) modal observations (a finite set of low Fourier modes), and (ii) nodal observations, i.e. values of the velocity field sampled at finitely many points on a uniform grid. The key upgrade is a piecewise linear interpolant built on a fixed five-tetrahedra subdivision of each grid cube, which removes the mollification step used previously and yields an explicit control of the derivatives of the interpolation operator purely in terms of the measured data. The criterion is also shown to be both necessary and sufficient for regularity.
Read the full abstract

Information Theory

Shannon meets Gödel-Tarski-Löb: Undecidability of Shannon Feedback Capacity for Finite-State Channels

We study the exact decision problem for feedback capacity of finite-state channels (FSCs). Given an encoding $e$ of a binary-input binary-output rational unifilar FSC with specified rational initial distribution, and a rational threshold $q$, we ask whether the feedback capacity satisfies $C_{fb}(W_e, \pi_{1,e}) \ge q$. We prove that this exact threshold problem is undecidable, even when restricted to a severely constrained class of rational unifilar FSCs with bounded state space. The reduction is effective and preserves rationality of all channel parameters. As a structural consequence, the exact threshold predicate does not lie in the existential theory of the reals ($\exists\mathbb{R}$), and therefore cannot admit a universal reduction to finite systems of polynomial equalities and inequalities over the real numbers. In particular, there is no algorithm deciding all instances of the exact feedback-capacity threshold problem within this class. These results do not preclude approximation schemes or solvability for special subclasses; rather, they establish a fundamental limitation for exact feedback-capacity reasoning in general finite-state settings. At the metatheoretic level, the undecidability result entails corresponding Gödel-Tarski-Löb incompleteness phenomena for sufficiently expressive formal theories capable of representing the threshold predicate.
Read the full abstract

Analysis of PDEs

Mean first escape times of Brownian motion on asymptotically hyperbolic and gas giant metric surfaces

This paper deals with the mean first escape time of Brownian motion on asymptotically hyperbolic and gas giant surfaces. We show that for a boundary defining function $\rho$, the mean first escape time $u_\epsilon(x)$ from the truncated Riemannian surface with an asymptotically hyperbolic metric $(M_\epsilon,\bar{g}/\rho^2) = (\{x\in M:\rho(x)\geq \epsilon\},\bar{g}/\rho^2) \subset (M,\bar{g}/\rho^2)$ satisfies the asymptotic expansion $u_\epsilon(x) = -\log \epsilon + \mathcal{O}(1)$ as $\epsilon\to 0 $. Furthermore, we show that in the case of a gas giant metric $g = \bar{g}/\rho^\alpha$, where $\alpha\in (0,2)$, the mean first escape time from the surface $(M_\epsilon,\bar{g}/\rho^\alpha)$ satisfies $u_\epsilon(x) = \mathcal{O}(1)$ as $\epsilon\to 0 $. Using techniques from the theory of polyhomogeneous conormal functions we explain this difference between in the mean first escape time on gas giant metric surfaces and asymptotically hyperbolic surfaces on the unit disc. Finally, we confirm these results using Monte Carlo simulations and finite difference methods on the disc.
Read the full abstract

All authors Jesse Gell-Redman, Emanuel József Godfried, Justin Tzou, Leo Tzou

Probability

On the structure of marginals in high dimensions

Let $G, G_1,\dots,G_N$ be independent copies of a standard gaussian random vector in $\mathbb{R}^d$ and denote by $\Gamma = \sum_{i=1}^N \langle G_i,\cdot\rangle e_i$ the standard gaussian ensemble. We show that, for any set $A\subset S^{d-1}$, with exponentially high probability, \[ \sup_{x\in A} \frac{1}{N}\sum_{i=1}^N \big| (\Gamma x)^\sharp_i - q_i\big| \le c \frac{ \mathbb{E} \sup_{x\in A} \langle G,x\rangle + \log^2N }{\sqrt N }. \] Here each $q_i$ is the $\frac{i}{N+1}$-quantile of the standard normal distribution and $(\Gamma x)^\sharp $ denotes the monotone increasing rearrangement of the vector $\Gamma x$. The estimate is sharp up to a possible logarithmic factor and significantly extends previously known bounds. Moreover, we show that similar estimates hold in much greater generality: after replacing the gaussian quantiles by the appropriate ones, the same phenomenon persists for a broad class of random vectors.
Read the full abstract

Differential Geometry

A brief introduction to Poisson geometry

These notes are based on an introductory minicourse on Poisson geometry given at CRM, Barcelona, in July 2022. They mostly contain foundational material, including motivating questions and key examples of Poisson structures, and highlight some recent tools and developments. The text is largely structured through exercises that lead readers to the desired conclusions, so it serves as a guided introduction to the basics of Poisson geometry.
Read the full abstract

Algebraic Geometry

Certifying Galois/monodromy Actions via Homotopy Graphs

We develop a certified numerical algorithm for computing Galois/monodromy groups of parametrized polynomial systems. Our approach employs certified homotopy path tracking to guarantee the correctness of the monodromy action produced by the algorithm, and builds on previous “homotopy graph" frameworks. We conduct extensive experiments with an implementation of this algorithm, which we have used to certify properties of several notable Galois/monodromy groups which arise in several examples drawn from pure and applied mathematics.
Read the full abstract

Combinatorics

Forest webs and pattern avoidance

In a recent preprint, Mike Cummings showed that the smooth components of suitably parametrized Springer fibers are in bijection with contracted, fully reduced Plücker degree-two $\mathfrak{sl}_r$-webs of standard type and that are forests. He showed these are enumerated by sequence A116731 in the OEIS, which is equinumerous with permutations avoiding the patterns {321,2143,3124}. Cummings posed the problem of strengthening this enumerative result by finding a bijection between these webs and a collection of pattern-avoiding permutations. Here we solve this problem, although notably not with the collection of patterns that Cummings had proposed. Rather, we give a bijection between this class of webs and permutations avoiding the patterns {132,4321,3214}.
Read the full abstract

Number Theory

Counting subalgebras of $\mathfrak{o}^n$

Let $\mathfrak{o}$ be a compact discrete valuation ring with residue field $\mathbb{F}_q$. We study the cotype zeta function of finite-index subalgebras of $\mathfrak{o}^n$ by adapting an approach developed by Voll for more general algebras. For each $\emptyset\neq I\subseteq\{1,\ldots,n-1\}$, we express the generating series counting subalgebras whose elementary divisor valuations, listed in decreasing order, have jumps precisely at the positions in $I$ as a finite weighted sum of $\mathfrak{o}$-adic integrals. We compute these integrals in several cases, obtaining explicit formulas for the generating series associated with every singleton set $I$ in arbitrary rank. We also obtain the cotype zeta function of unital subalgebras of $\mathfrak{o}^4$. As a by-product, we obtain a universal lower bound for the abscissa of convergence of the subalgebra zeta function of any $\mathfrak{o}$-algebra that is free of rank $n$ as an $\mathfrak{o}$-module. We also establish several results concerning the uniformity problem for the subalgebra zeta function of \(\mathfrak{o}^n\). To do so, we study two systems of polynomial equations over the residue field and apply the results to derive the number of irreducible unital subalgebras of $\mathfrak{o}^n$ of index $q^e$, for $e\leq n+3$. Finally, we use the decomposition of subalgebras into local factors to obtain the number of unital subalgebras of $\mathfrak{o}^n$ of index $q^e$ for $e\leq 9$. For each fixed $n$, these subalgebra counts are universal polynomials in $q$.
Read the full abstract

Functional Analysis

Hilbertian Hardy–Sobolev Spaces on Tube Domains over Convex Cones

We introduce Hilbertian Hardy–Sobolev spaces on tube domains over convex cones and develop their structural theory from a Fourier-analytic point of view. We first establish a Paley–Wiener type representation, which identifies these spaces with weighted $L^2$ spaces on the dual cone and reveals their intrinsic Fourier structure. This representation leads naturally to a Hardy–Sobolev decomposition theorem for boundary Sobolev spaces on $\mathbb{R}^d$. Building on these structural results, we derive explicit reproducing kernels and characterize Carleson measures for the Hilbertian Hardy–Sobolev spaces. As a preliminary operator-theoretic application, we also derive basic consequences for multipliers and weighted composition operators on these spaces.
Read the full abstract

Combinatorics

Enumerating Prime Patterns in Juggling Variations

Juggling patterns can be mathematically modeled as closed walks within directed state graphs. In this paper, we present a unified framework of unbounded juggling patterns and its variations (including multiplex, colored, and passing) primarily through the formalism of the juggling state. By extending this state-based approach and utilizing combinatorial tools such as set partitions and filled Ferrers diagrams, we find and prove a new lower bound on the number of $b$-ball prime patterns with period $n$. Further, we determine exact counts for 2-ball multiplex, 1-ball passing, and 2-ball colored juggling patterns, as well as a lower bound for 2-ball passing. We also provide an extensive analysis of the asymptotic growth rates for these pattern counts. Finally, we formalize the infinite state graph, $G_\infty$, and utilize flip-reverse involutions to establish bijections between classes of prime patterns, exploring how fixing a specific state influences the enumeration of prime walks.
Read the full abstract

All authors Steve Butler, Vera Choi, Joel Jeffries, Nina McCambridge, Asia Morgenstern, Samuel Orellana Mateo

Group Theory

On the triviality and non-triviality of the automorphism group of a skew brace

It is a simple fact that a group has a trivial automorphism group if and only if it is of order $1$ or $2$. We prove that the same holds for certain families of skew braces, and given any odd prime $p$, we construct a skew brace of order $2p^3$ that has a trivial automorphism group.
Read the full abstract

Representation Theory

Tangent spaces of spherical Schubert varieties and counterexamples to the reducedness conjecture

Given a simply-connected simple algebraic group $G$, we determine the tangent space of any Finkelberg-Mirković Schubert scheme at the base point of the affine Grassmannian of $G$. As a consequence, we exhibit non-reduced Finkelberg-Mirković Schubert schemes when $G$ is of type $E_6,E_7$ and $E_8$.
Read the full abstract

Analysis of PDEs

Asymptotic linear stability of columnar vortices driven by Coriolis force

In this paper, we establish the asymptotic linear stability of a class of Coriolis-driven columnar vortices for the 3-D axisymmetric Euler equations. This result represents a critical step toward proving the nonlinear asymptotic stability of such vortices. The key and widely applicable strategy is to construct a distorted Fourier basis, which is achieved by solving a two-parameter $(c, \xi)$-dependent Schrödinger equation associated with the linearized operator of the system. To capture the precise asymptotic behavior of the solution, we decompose the $c-\xi$ plane into distinct regions, with the partitioning guided by the leading-order profiles of the Schrödinger equation across different parameter regimes.
Read the full abstract

Information Theory

Asymptotically ideal Disjunctive Hierarchical Secret Sharing Scheme with an Explicit Construction

Disjunctive Hierarchical Secret Sharing (DHSS) scheme is a secret sharing scheme in which the set of all participants is partitioned into disjoint subsets. Each disjoint subset is said to be a level, and different levels have different degrees of trust and different thresholds. If the number of cooperating participants from a given level falls to meet its threshold, the shortfall can be compensated by participants from higher levels. Many ideal DHSS schemes have been proposed, but they often suffer from big share sizes. Conversely, existing non-ideal DHSS schemes achieve small share sizes, yet they fail to be both secure and asymptotically ideal simultaneously. In this work, we present an explicit construct of an asymptotically ideal DHSS scheme by using a polynomial, multiple linear homogeneous recurrence relations and one-way functions. Although our scheme has computational security and many public values, it has a small share size and the dealer is required polynomial time.
Read the full abstract

All authors Jian Ding, Cheng Wang, Haifeng Yu, Hongju Li, Cheng Shu

Optimization and Control

On the sensitivity of the subspace predictor to behavioral perturbations

Behavioral systems define discrete-time LTI systems in terms of a set of trajectories, which forms a linear subspace. This subspace underlies the subspace predictor used in data-driven prediction and control. In practice, such subspaces are typically represented through data matrices. For robustness certification and uncertainty quantification, however, these matrix representations are coordinate-dependent and therefore do not provide a coordinate-free way to quantify uncertainty. In this work, we derive an explicit prediction error bound in terms of behavioral distance between the true subspace and an estimate, showing that the predictor is locally Lipschitz with respect to behavioral perturbations. We also present a one-step prediction error bound that is relevant for receding-horizon implementations, which becomes computable when combined with existing behavioral-distance certificates. Numerical studies show that our bound is tighter than an existing data-matrix perturbation bound, and remains computable, though more conservative, when combined with an existing behavioral distance certificate.
Read the full abstract

Differential Geometry

On vector-valued multisymplectic forms

We obtain a standard local presentation for a vector-valued multisymplectic form on a smooth manifold, generalizing the known proof for polysymplectic forms. We show that vector-valued multisymplectic forms on a finite-dimensional real vector space form a non-unital operad. We prove an entropy inequality for partial compositions.
Read the full abstract

Algebraic Geometry

$I$-Maximal Variation of Hodge Structure and Jacobian Rings

We investigate higher–order variation of Hodge structure for families of smooth hypersurfaces and complete intersections through the notion of $I$–maximal variation. Using Griffiths' description of primitive cohomology, we interpret the infinitesimal variation of Hodge structure and the $n$–fold Yukawa coupling as graded multiplication maps in the Jacobian ring. Our main result shows that the Strong Lefschetz property of the Jacobian ring provides the algebraic mechanism ensuring $I$–maximal variation. In particular, we prove that smooth hypersurfaces of degree $d\ge n+2$ and smooth complete intersections with $\kappa>0$ exhibit $I$–maximal variation. We further establish that for complete intersections of general type the infinitesimal Torelli property is equivalent to the nondegeneracy of the Yukawa coupling. Finally, we analyze degenerations and show that the failure of the Strong Lefschetz property leads to degeneration of the Yukawa coupling and the loss of $I$–maximal variation. These results identify the Lefschetz property of the Jacobian ring as the fundamental algebraic structure governing maximal variation of Hodge structure.
Read the full abstract

Algebraic Geometry

Maximal Infinitesimal Variation of Hodge Structure for Singular Curves

We study the infinitesimal variation of Hodge structure for families of algebraic curves and extend the classical theory from smooth curves to singular and non–planar settings. Using the deformation space $\mathrm{Ext}^1(\Omega_X,\mathcal O_X)$ and the dualizing sheaf, we define a singular analogue of maximal infinitesimal variation. For equisingular families of plane curves with planar Gorenstein singularities, we prove that the infinitesimal variation attains maximal rank equal to the arithmetic genus. We show that the rank decomposes into a geometric contribution from the normalization and a singular contribution measured by the $\delta$–invariants. For non–equisingular degenerations, the rank defect equals the drop of the total $\delta$–invariant and admits an interpretation in terms of vanishing cycles and mixed Hodge structures. We further extend the results to non–planar curves under suitable Petri and deformation conditions.
Read the full abstract

Tuesday 17 March

Information Theory · Cross-listed from cs.LG

A General Deep Learning Framework for Wireless Resource Allocation under Discrete Constraints

While deep learning (DL)-based methods have achieved remarkable success in continuous wireless resource allocation, efficient solutions for problems involving discrete variables remain challenging. This is primarily due to the zero-gradient issue in backpropagation, the difficulty of enforcing intricate constraints with discrete variables, and the inability in generating solutions with non-same-parameter-same-decision (non-SPSD) property. To address these challenges, this paper proposes a general DL framework by introducing the support set to represent the discrete variables. We model the elements of the support set as random variables and learn their joint probability distribution. By factorizing the joint probability as the product of conditional probabilities, each conditional probability is sequentially learned. This probabilistic modeling directly tackles all the aforementioned challenges of DL for handling discrete variables. By operating on probability distributions instead of hard binary decisions, the framework naturally avoids the zero-gradient issue. During the learning of the conditional probabilities, discrete constraints can be seamlessly enforced by masking out infeasible solutions. Moreover, with a dynamic context embedding that captures the evolving discrete solutions, the non-SPSD property is inherently provided by the proposed framework. We apply the proposed framework to two representative mixed-discrete wireless resource allocation problems: (a) joint user association and beamforming in cell-free systems, and (b) joint antenna positioning and beamforming in movable antenna-aided systems. Simulation results demonstrate that the proposed DL framework consistently outperforms existing baselines in terms of both system performance and computational efficiency.
Read the full abstract

All authors Yikun Wang, Yang Li, Yik-Chung Wu, Rui Zhang