Peer Revenue Equity in science

Papers

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

Mathematics

math Watch Mathematics (sign in to use)

Saturday 21 February

Optimization and Control · Cross-listed from cs.LG

Exponential Convergence of (Stochastic) Gradient Descent for Separable Logistic Regression

Gradient descent and stochastic gradient descent are central to modern machine learning, yet their behavior under large step sizes remains theoretically unclear. Recent work suggests that acceleration often arises near the edge of stability, where optimization trajectories become unstable and difficult to analyze. Existing results for separable logistic regression achieve faster convergence by explicitly leveraging such unstable regimes through constant or adaptive large step sizes. In this paper, we show that instability is not inherent to acceleration. We prove that gradient descent with a simple, non-adaptive increasing step-size schedule achieves exponential convergence for separable logistic regression under a margin condition, while remaining entirely within a stable optimization regime. The resulting method is anytime and does not require prior knowledge of the optimization horizon or target accuracy. We also establish exponential convergence of stochastic gradient descent using a lightweight adaptive step-size rule that avoids line search and specialized procedures, improving upon existing polynomial-rate guarantees. Together, our results demonstrate that carefully structured step-size growth alone suffices to obtain exponential acceleration for both gradient descent and stochastic gradient descent.
Read the full abstract

All authors Sacchit Kale, Piyushi Manupriya, Pierre Marion, Francis Bach, Anant Raj

Algebraic Topology

Covering Maps with respect to Topologies on the Fundamental Group

In this paper, using the classical covering theory, we introduce a generalization of covering maps of a space $X$ with respect to a topology $\tau$ on the fundamental group of $X$. We show that the famous notions, covering, semicovering, generalized covering and fibration maps are of special cases of this new notion $\pi_1^{\tau}$-covering map. Moreover, among presenting some properties for this new notion, we compare $\pi_1^{\tau}$-covering maps of a space $X$ for several famous topologies on the fundamental group of $X$.
Read the full abstract

Numerical Analysis

Structure-preserving Krylov Subspace Approximations for the Matrix Exponential of Hamiltonian Matrices: A Comparative Study

We study structure-preserving Krylov subspace methods for approximating the matrix-vector products f(H)b, where H is a large Hamiltonian matrix and f denotes either the matrix exponential or the related phi-function. Such computations are central to exponential integrators for Hamiltonian systems. Standard Krylov methods generally destroy the Hamiltonian structure under projection, motivating the use of Krylov bases with J-orthogonal columns that yield Hamiltonian projected matrices and symplectic reduced exponentials. We compare several such structure-preserving Krylov methods on representative Hamiltonian test problems, focusing on accuracy, efficiency, and structure preservation, and briefly discuss adaptive strategies for selecting the Krylov subspace dimension.
Read the full abstract

Differential Geometry · Cross-listed from q-bio.MN

Convex Analysis of Relaxation Dynamics in Chemical Reaction Networks and Generalized Gradient Flows

We obtain bounds on the Kullback–Leibler divergence to equilibrium for mass-action chemical reaction networks (CRNs) with equilibrium. The associated trajectory-dependent bounds are characterized in terms of the singular values of the stoichiometric matrix, convexity parameters, and time-integrated activities via deformed-exponential-type functions. We further extend these bounds within a generalized gradient flow framework. We highlight the biological relevance of this framework: the resulting bounds apply to quasi-steady-state regimes, where long transients and plateau-like behavior are common and functionally important. We illustrate the framework using a catalytic CRN exhibiting plateaus, where the bounds provide an a posteriori characterization of slow relaxation induced by local convexity and provide a bound-based approach to quantifying relaxation in CRNs.
Read the full abstract

Probability

Geometric Large-Deviation-Type Principles for Mixed Measures

We study a geometric analogue of the large deviation principle for mixed measures associated with a class of $\log$-concave probability measures whose densities depend on the gauge of a convex body. For convex bodies in $\mathbb{R}^n$, we prove a geometric large-deviation-type asymptotic for first-order mixed measures, in which the decay under dilation is governed by a natural inradius associated with the measure. In the planar case, we derive an explicit representation and prove a genuine logarithmic limit for second-order mixed measures. As an application, we prove a comparison theorem showing that asymptotic dominance under dilation forces inclusion between convex bodies.
Read the full abstract

Algebraic Topology

On the closed geodesics problem

Let $\bk $ be a field of characteristic $p\geq 0$ and $X$ a simply connected finite CW complex. In this text, we prove that: {\sl if the cohomology algebra $H^*(X;\bk)$ is generated, as an algebra, by at least two linearly independent elements, then the sequence of Betti numbers $ \left( \dim H^n(LX;\bk)\right)_{n\geq 1 }$ grows unbounded.} This provides a complete solution of the closed geodesics problem.
Read the full abstract

Probability

Boundedness of discounted branching random walks via generic chaining

Consider a discrete-time supercritical discounted branching random walk, in which increments at depth $k$ are independent and identically distributed with the same law as $m^{-kH}Y$, where $Y$ has a fixed law, $H>0$, and $m>1$ is the expected number of offspring at depth one. We provide a clean characterization of the boundedness of the discounted branching random walk: under mild conditions on the offspring distribution, the process is almost surely bounded if and only if $\mathbb{E}[|Y|^{1/H}]<\infty$. This extends results of Athreya (1985) and Aïdékon–Hu–Shi (2024), and provides a partial answer to Open Problem 31 of Aldous–Bandyopadhyay (2005).
Read the full abstract

Analysis of PDEs

A unified duality framework for barotropic, quantum and Korteweg fluids

We investigate a dual variational formulation, in the spirit of Brenier, for several compressible fluid models: the compressible barotropic Euler system, the quantum Euler system, and the Euler-Korteweg system. We identify a unified abstract framework encompassing all three systems, which enables a simultaneous analysis. By introducing time-adaptive weights, we establish the consistency of the duality scheme on large time intervals. We prove the existence of variational dual solutions to the corresponding Cauchy problems for continuous, vacuum-free initial data in spaces of finite Radon measures, and establish the absence of a duality gap. As an application, we formulate and prove a 'Dafermos principle' for these models: no subsolution can dissipate the total entropy earlier or at a faster rate than the corresponding strong solution on its interval of existence. We also discuss connections between our abstract consistency result and Brenier's shock-free substitutes for entropy solutions of Burgers' equation.
Read the full abstract

Probability

Stochastic invariance in infinite dimension beyond Lipschitz coefficients

We establish necessary and sufficient conditions for stochastic invariance of closed subsets in Hilbert spaces for solutions to infinite-dimensional stochastic differential equations (SDEs) under mild assumptions on the coefficients. Our first characterization is formulated in terms of certain normal vectors to the invariance set and requires differentiability only of the dispersion operator, but not of the diffusion coefficient itself. The condition involves a suitable corrected drift expressed through the dispersion operator and its Moore-Penrose pseudoinverse, extending the classical Stratonovich correction term to the present low-regularity setting. Our second characterization is given in terms of the positive maximum principle for the infinitesimal generator of the associated diffusion process. We illustrate our characterizations in the case of invariant manifolds.
Read the full abstract

Analysis of PDEs

Multiple positive bound states for NLS equations on noncompact metric graphs with potential

In this paper, we establish the existence of geometrically distinct bound states for the subcritical Nonlinear Schrödinger (NLS) equation on a noncompact metric graph, provided that the mass is sufficiently large. By bound states we mean critical points of the NLS energy functional constrained to the manifold of fixed $L^2$-norm. We demonstrate that the NLS equation admits at least as many positive bound states as the number of bounded edges of the graph. Our approach is purely variational and relies on a doubly constrained minimization procedure. It is worth noting that, compared with the case where the potential vanishes, the scaling structure of solitons is destroyed here. Moreover, the classical energy asymptotic behavior $\mathcal{E}(\phi_{\mu},\mathbb{R})=-\theta \mu^{2\beta+1}$ is no longer preserved.
Read the full abstract

Algebraic Topology

Topological shape transform for thymus structures

The Euler characteristic transform (ECT) is an emerging and powerful framework within topological data analysis for quantifying the geometry of shape. The applicability of ECT has been limited due to its sensitivity to noisy data. Here, we introduce SampEuler, a novel ECT-based shape descriptor designed to achieve enhanced robustness to perturbations. We provide a theoretical analysis establishing the stability of SampEuler and validate these properties empirically through pairwise similarity analyses on a benchmark dataset and showcase it on a thymus dataset. The thymus is a primary lymphoid organ that is essential for the maturation and selection of self-tolerant T cells, and within the thymus, thymic epithelial cells are organized in complex three-dimensional architectures, yet the principles governing their formation, functional organization, and remodeling during age-related involution remain poorly understood. Addressing these questions requires robust and informative shape descriptors capable of capturing subtle architectural changes across developmental stages. We develop and apply SampEuler to a newly generated two-dimensional imaging dataset of mouse thymi spanning multiple age groups, where SampEuler outperforms both persistent homology–based methods and deep learning models in detecting subtle, localized morphological differences associated with aging. To facilitate interpretation, we develop a vectorization and visualization framework for SampEuler, which preserves rich morphological information and enables identification of structural features that distinguish thymi across age groups. Collectively, our results demonstrate that SampEuler provides a robust and interpretable approach for quantifying thymic architecture and reveals age-dependent structural changes that offer new insights into thymic organization and involution.
Read the full abstract

All authors Haochen Yang, Vadim Lebovici, Andreas Tarcevski, Liliana Tchernev, Saulius Zuklys, Georg A. Holländer, Helen M. Byrne, Heather A. Harrington

Information Theory

Turbo Coded Single Sideband OFDM-OQAM Signaling through Frequency Selective Rayleigh Fading Channels

This work investigates the bit-error-rate (BER) performance of turbo coded orthogonal frequency division multiplexed - offset quadrature amplitude modulated (OFDM- OQAM) signals transmitted through frequency selective Rayleigh fading channels in the presence of carrier frequency offset (CFO) and additive white Gaussian noise (AWGN). The highlight of this work is to use the root raised cosine (RRC) pulse and its Hilbert transform as the complex-valued transmit filter and a simple matched filter at the receiver. The proposed system is similar to single sideband (SSB) modulation, that has roots in analog communications. Turbo code and subcarrier diversity is employed to improve the BER performance over that of an uncoded system. Discrete-time algorithms for frame detection, two-step CFO, channel and noise variance estimation have been proposed. A single transmit and receive antenna is assumed. Similar work has not been done earlier.
Read the full abstract

Operator Algebras

Banach manifolds of spectrally small quantum-group representations

We prove that finite-spectrum representations of compact quantum groups either in unital $C^*$-algebras $A$ or on Banach spaces $E$ exhibit the same Banach-space-modeled differential-geometric structure as their classical analogues: (a) they are Banach analytic manifolds; (b) locally homogeneous under conjugation by the pertinent Banach Lie group $U(A)$ or $GL(E)$; (c) with orbit maps fibering principally; (d) and hence with said orbit maps admitting local analytic splitting. We also identify the finite-spectrum unitary representations as precisely those that are norm-continuous in the appropriate sense when the compact quantum group has at least one classical point, again generalizing the classical parallel present in various forms in work of Kallman, Shtern and the author.
Read the full abstract

Number Theory

Solving equations of signature $(p,p,2)$ with coefficients over number fields

Using the modular method, we study solutions to the Diophantine equation $$Aa^p+Bb^p=Cc^2$$ over number fields. We first prove an asymptotic result for general number fields satisfying an appropriate $S$-unit condition by assuming some standard conjectures in the case of fields that are not totally real. Specifically, we verify that this condition holds for an infinite family of real quadratic fields. Outside the asymptotic setting, we also obtain effective results. In particular, for the equation $$a^p+db^p=c^2$$ over $K= \mathbb{Q}(\sqrt{-d})$ with $d \in \{3, 11, 19, 43, \}$ and $K= \mathbb{Q}(\sqrt d)$ with $d \in \{3, 5, 11, 13, 19, 29\}$, we find explicit bounds (depending on $d$) such that no non-trivial solutions of a certain type exist whenever $p$ exceeds these bounds.
Read the full abstract

All authors Begum Gulsah Cakti, Erman Isik, Yasemin Kara, Ekin Ozman

Optimization and Control

Limits of Convergence-Rate Control for Open-Weight Safety

Open-weight foundation models can be fine-tuned for harmful purposes after release, yet no existing training resistance methods provide theoretical guarantees. Treating these interventions as convergence-rate control problems allows us to connect optimization speed to the spectral structure of model weights. We leverage this insight to develop a novel understanding of convergence rate control through spectral reparameterization and derive an algorithm, SpecDef, that can both provably and empirically slow first- and second-order optimization in non-adversarial settings. In adversarial settings, we establish a fundamental limit on a broad class of convergence rate control methods including our own: an attacker with sufficient knowledge can restore fast convergence at a linear increase in model size. In order to overcome this limitation, future works will need to investigate methods that are not equivalent to controlling convergence rate.
Read the full abstract

All authors Domenic Rosati, Xijie Zeng, Hong Huang, Sebastian Dionicio, Subhabrata Majumdar, Frank Rudzicz, Hassan Sajjad

Combinatorics

Constructions of two-dimensional optical orthogonal codes of weight three

The study of optical orthogonal codes has been motivated by an application in an optical code-division multiple access system. This paper focuses on optimal two-dimensional optical orthogonal codes with autocorrelation and cross-correlation both equal to $1$. By examining the structures of $n$-cyclic group divisible packings and semi-cyclic incomplete holey group divisible designs, we present new combinatorial constructions for two-dimensional $(m\times n,k,1)$-optical orthogonal codes. As a consequence, the exact number of codewords of an optimal two-dimensional $(m\times n,3,1)$-optical orthogonal code is determined for any positive integers $m$ and $n$.
Read the full abstract

All authors Xiuling Shan, Lidong Wang, Yanxun Chang, Xiaomiao Wang

Analysis of PDEs

Partial regularity in nonlocal systems II

Solutions to nonlinear nonlocal systems of order $2s>1$ in $\mathbb{R}^n$ are $C^{1,\alpha}$, for every $\alpha <2s-1$, outside a closed singular set whose Hausdorff dimension is less than $n-2$, and which is empty when $n=2$.

Analysis of PDEs

Slow-fast system in Rosales-Majda combustion model with fractional order kinetics

We consider traveling wave solutions of a one-dimensional model for detonation waves derived by Rosales and Majda, when the reaction order $\alpha$ belongs to $[0,1)$. The chemical kinetics is a simplified Arrhenius law or a Heaviside function. The model in the reaction zone is a slow-fast dynamical system for a vector representing temperature and mass fraction, which depends on the velocity $c$ and small viscosity $\beta$. Our goal in this paper is to study the bifurcation diagram in the $(\beta,c)$ parameter space and identify the nature of the trajectories corresponding to viscous shock waves. The demonstrations are based on a variety of techniques including the Poincaré-Bendixson theorem and the Fenichel theory. Theoretical results are confirmed by numerical computations.
Read the full abstract

Group Theory

Profinite groups with restricted centralizers of powers

A group $G$ is said to have restricted centralizers if for every $x\in G$ the centralizer $C_G(x)$ either is finite or has finite index in $G$. Shalev showed that a profinite group with restricted centralizers is virtually abelian. Here we take interest in profinite groups $G$ for which there is an integer $n$ such that $C_G(x^n)$ is either finite or open whenever $x\in G$. It is shown that such a group $G$ has an open normal subgroup $T$ with the property that $G/Z(T)$ has finite exponent.
Read the full abstract

Logic

Big Ramsey combinatorics of the Cantor set and a simple proof of Blass' perfect set theorem

In this paper we present a simple approach to big Ramsey combinatorics of the Cantor set $2^\omega$. Using Infinite Dual Ramsey Theorem of Carlson and Simpson, we show that $2^\omega$, viewed as a topological space, has finite big Ramsey degrees. We then examine several natural topological first-order structures arising from the Cantor set and prove that each of them inherits finite big Ramsey degrees. As a consequence, we obtain a simple proof of Blass' perfect set theorem, although our method does not recover the sharp bound $(n-1)!$ for the number of colors. We also show that the complete Boolean algebra on countably many atoms has finite big Ramsey degrees, in contrast with the recent result showing that the countable atomless Boolean algebra does not have big Ramsey degrees.
Read the full abstract

Group Theory

On the decidability of the integrability of finite groups

An integral of a group $G$ is a group $H$ whose commutator subgroup is isomorphic to $G$. In this paper, we prove that the integrability of a finite group is a decidable problem.

General Topology

Countability Properties of Weakly Compact Sets in Asymmetric Locally Convex Spaces

A dual pair formulation for asymmetric locally convex spaces is introduced that strictly generalises the ordinary vector space setting. The concept of a polar topology carries over to the asymmetric case and some familiar results are reproduced or generalised such as the bipolar theorem and a Mackey-Arens type theorem. The implications of weak compactness and countability properties are studied and appear intimately connected to separation properties. An asymmetric analogue $C_a(K)$ of the well-known space $C_p(K)$ is introduced and the properties of (relatively) countably respectively sequentially) compact subspaces are investigated. In particular, it is shown that $C_a([0,1])$ is not angelic. However, for Hausdorff subspaces, the different compactness conditions are equivalent and imply the Fréchet-Urysohn property. Moreover, an analogue of the Eberlein-Šmulian theorem holds in asymmetrically normed spaces.
Read the full abstract

Analysis of PDEs

Steady states and dynamics of a higher dimensional thin film equation

We study a higher-dimensional thin film equation that incorporates competitive effects between aggregation and repulsion, where repulsion is modeled by fourth-order diffusion and aggregation by backward second-order degenerate diffusion, with a degenerate diffusion exponent $m>0$. We first conduct a systematic analysis of the existence and geometric properties of steady-state solutions for all $m>0$, revealing a critical threshold $m^*=(d+2)/(d-2)$ for variational compactness and solution structure. For $0 < m < m^*$, we then prove that, under natural regularity constraints, radially decreasing steady states coincide with both the extremals of the Gagliardo-Nirenberg-Sobolev inequality and the global minimizers of the free energy. Moreover, we establish the uniqueness of such steady states for $m \neq 1 + 2/d$. Furthermore, in the supercritical regime $1 + 2/d < m < m^*$, we identify a sharp threshold given by the $L^{m+1}$ norm of the unique radial steady-state solution, which distinguishes between global existence for initial data below the threshold and finite-time blow-up for initial data above the threshold. The main contribution of this work is to use steady-state solutions as a theoretical pivot to construct a unified analytical framework that connects parameter classification, variational structure and dynamic behavior. This framework elucidates how the regularity barrier prevents infinite energy descent and selects stable equilibrium states, and thus predicts the global evolution of the system, thereby providing a unified variational principle for understanding steady-state selection and dynamic bifurcations in such higher-order degenerate diffusion equations.
Read the full abstract

Representation Theory

Coxeter planes as fixed points of Verlinde fusion rings

For the Coxeter groups of ADE type, we provide a construction of their Coxeter planes as fixed points of actions of hypergroups associated to Verlinde fusion rings. This builds upon the well-known ADE classification of $\mathbb{Z}_+$-modules over these fusion rings.

Representation Theory

Balanced pairs, virtually Gorenstein rings, and cotorsion torsion triples

For any ring $R$, we investigate balanced pairs of classes of modules and their relations to cotorsion triples. We characterize the special balanced pairs that fit into complete hereditary cotorsion triples. As an application, we prove that Gorenstein projective and Gorenstein injective modules form a balanced pair, if and only if $R$ is right virtually Gorenstein. We also characterize the tilting and cotilting cotorsion pairs arising from balanced pairs. In [4], cotorsion torsion triples in abelian categories were employed in the representation theory of rectangular grids occurring in persistent homology theory. For module categories, we use infinite dimensional tilting theory to completely classify all cotorsion torsion triples by means of $1$-resolving subcategories of $\rfmod R$, and to give an explicit 1-1 correspondence between the formally dual notions of cotorsion torsion triples of right $R$-modules and torsion cotorsion triples of left $R$-modules. This correspondence is bijective in case the underlying ring $R$ is left noetherian, but not in general.
Read the full abstract

Mathematical Physics

On the adiabatic invariance of the action of a trapped wave

Recently, it has been shown (Gavrilov et al., Nonlinear Dyn, 112, 2024) that in a linear solid discrete-continuous system with several slowly time-varying parameters, the amplitude of a strongly localized mode (a trapped wave) can be calculated as a function of current parameter values and does not depend on the history of the parameter change. This result allows us to introduce the adiabatic invariant for such a system according to the general definition as a quantity that remains approximately constant if the parameters vary slowly. In this paper, we show that, defined in this manner, the adiabatic invariant can be calculated as the ratio of the total energy of the trapped wave to its frequency. This yields a significantly simplified approach to solving a class of problems concerning localized oscillation of continuous systems with discrete inclusions, although the definition of the wave energy can be ambiguous. Thus, we can consider the newly introduced adiabatic invariant as a straightforward generalization of the concept known to Hamiltonian systems. Finally, we introduce an effective Hamiltonian system, which is characterized by the same adiabatic invariant as the trapped wave. This yields another highly straightforward approach to deriving the amplitude evolution law, although further investigation is required.
Read the full abstract

Analysis of PDEs

Heisenberg Uncertainty Principle on half spaces and Orthants: Best constants, Optimizers and Stability

Though the sharp Heisenberg Uncertainty Principle has been extensively studied in the entire Euclidean spaces, the counterpart on the half spaces or more general orthants has been missing in the literature. We investigate the sharp Heisenberg Uncertainty Principle on orthants by computing explicitly the optimal constant and determining all possible extremal functions. Moreover, we establish several stability estimates of the Heisenberg Uncertainty Principle on the half spaces and orthants.
Read the full abstract

All authors Nguyen Lam, Yukta Lodha, Guozhen Lu, Ambar N. Sengupta

Probability

Orthogonal polynomials on path-space

We consider the orthogonalisation of the signature of a stochastic process as the analogue of orthogonal polynomials on path-space. Under an infinite radius of convergence assumption, we prove density of linear functions on the signature in $L^p$ functions on grouplike elements, making it possible to represent a square-integrable function on (rough) paths as an $L^2$-convergent series. By viewing the shuffle algebra as commutative polynomials on the free Lie algebra, we revisit much of the theory of classical orthogonal polynomials in several variables, such as the recurrence relation and Favard's theorem. Finally, we restrict our attention to the case of Brownian motion with and without drift, and prove that dimension-independent orthogonal signature exists with drift but not without. We end with numerical examples of how orthogonal signature polynomials of Brownian motion can be applied for the approximation of functions on paths sampled from the Wiener measure.
Read the full abstract

All authors Ilya Chevyrev, Emilio Ferrucci, Darrick Lee, Terry Lyons, Harald Oberhauser, Nikolas Tapia

Rings and Algebras

Finite basis problem for varieties of algebraic systems

This is a survey on the finite basis problem for varieties of algebraic systems. Our exposition is in two directions: (i) We give numerous examples of varieties which are not finitely based. (ii) We give examples of important varieties with the property that they and their subvarieties are finitely based. A special attention is paid on the varieties of semigroups, groups, associative, Lie and other nonassociative algebras.
Read the full abstract

Commutative Algebra

Locally prime modules

For a commutative unital ring $R$ with fixed ideals $I$ and $J$, we introduce and study $I$-prime $R$-modules and $(I, J)$-prime $R$-modules together with their duals $I$-coprime $R$-modules and $(I,J)$-coprime $R$-modules respectively. We employ category-theoretic techniques to reveal their structural properties. Our main results are versions of the Greenlees-May Duality and the Matlis-Greenlees-May Equivalence to the setting of these prime and coprime modules. This generalizes work on $I$-reduced modules and $I$-coreduced modules. We demonstrate that these “locally prime" modules serve as a tool for studying the classical “globally prime" modules, creating a bridge between local and global primality.
Read the full abstract