Peer Revenue Play credits. Real papers.

Papers

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

Mathematics

math Watch Mathematics (sign in to use)

Tuesday 7 April

General Topology

Topology, forcing, and graph colourings

We introduce a family of forcing notions that are helpful in showing that certain graphs do not have countable colourings of (additive) Borel class alpha. We construct graphs that are ”weakly minimal” for such colourings.
Read the full abstract

All authors Noam Greenberg, Dominique Lecomte, Dan Turetsky, Miroslav Zelen

Algebraic Geometry

Minimal-Degree Foliations on Cominuscule Grassmannians

Given $X$ a cominuscule Grassmannian (or irreducible Hermitian symmetric space) and an integer $p,$ we compute the minimum $l(p)$ such that $H^0 (\Omega^p_X (l(p)))$ is not 0. This allows us to conclude that any codimension-one foliation of degree zero on a cominuscule Grassmannian is a pencil of hyperplanes, improving a result of the first and third authors with D. Faenzi. We also deduce the structure of codimension-one foliations of degree one. Finally, we provide families of examples of high codimensional foliations of minimal degree on classical Grassmannians, Lagrangian Grassmannians, Spinor varieties, and the Cayley plane.
Read the full abstract

Complex Variables

Localization of Bergman Kernels and the Cheng-Yau Conjecture on Real Analytic Pseudoconvex Domains

In this paper, we establish the localization of Bergman kernels for unbounded pseudoconvex domains near boundary points of finite D'Angelo type. This result was proved by Engliš more than twenty years ago for bounded pseudoconvex domains and had remained as an open question in the unbounded setting. Related foundational work was carried out by Fefferman, Kerzman, Boutet de Monvel–Sjöstrand, Boas, Bell, and others. Combining our localization theorem with an extension theorem of Mir–Zaitsev, we prove that the Bergman metric of a bounded pseudoconvex domain with real-analytic boundary is Einstein if and only if the domain is biholomorphic to the unit ball. This result advances a longstanding conjecture of Cheng and Yau. A key step is to show that the Bergman metric of a smooth, possibly unbounded, pseudoconvex domain cannot be Kähler–Einstein if its boundary contains a non-strongly pseudoconvex \(h\)-extendible point. We further prove that a bounded weakly pseudoconvex real-analytic domain with a Kähler–Einstein Bergman metric must possess a weakly pseudoconvex \(h\)-extendible boundary point, thereby reducing the problem to the \(h\)-extendible setting. This paper draws deeply on, and reveals essential connections among many subfields of microlocal analysis, several complex variables, and complex geometry.
Read the full abstract

Algebraic Geometry

A Pardon Algebra for Zero-cycles

Recently, John Pardon proved the MNOP conjecture (on the GW-DT correspondence for CY3s) by introducing a new mathematical gadget, which we call the Pardon homology algebra of 1-cycles in 3-folds. We work out an analogous construction for 0-cycles in d-folds. This gives a new point of view on enumerative problems involving point-counting, such as, for example, the degree zero MNOP conjecture on the Hilbert scheme of points in projective 3-folds.
Read the full abstract

Optimization and Control

Optimal Centered Active Excitation in Linear System Identification

We propose an active learning algorithm for linear system identification with optimal centered noise excitation. Notably, our algorithm, based on ordinary least squares and semidefinite programming, attains the minimal sample complexity while allowing for efficient computation of an estimate of a system matrix. More specifically, we first establish lower bounds of the sample complexity for any active learning algorithm to attain the prescribed accuracy and confidence levels. Next, we derive a sample complexity upper bound of the proposed algorithm, which matches the lower bound for any algorithm up to universal factors. Our tight bounds are easy to interpret and explicitly show their dependence on the system parameters such as the state dimension.
Read the full abstract

Optimization and Control

A flatness proof of the exponential turnpike phenomenon for linear-quadratic optimal control problems

We revisit finite-dimensional linear-quadratic optimal control from the viewpoint of differential flatness. If the pair (A, B) is controllable, then the linear control system is flat, and every trajectory can be parametrized by a flat output and finitely many of its derivatives. Once this parametrization is inserted into the quadratic functional, the Euler-Lagrange condition becomes a linear differential equation with constant coefficients, or more generally a polynomial matrix differential equation. After reduction to Smith normal form, this equation decouples into scalar constant-coefficient equations, and its solutions are exponentialpolynomials. This yields a viewpoint on the turnpike phenomenon that is quite different from the classical Hamiltonian-Riccati analysis: the turnpike mechanism appears directly from the stable-unstable splitting of the reduced flat equation. In particular, when the reduced Euler-Lagrange operator has no purely imaginary characteristic roots and when the endpoint constraints act nondegenerately on the stable and unstable modes, the optimal trajectory consists of a left-boundary layer, a right-boundary layer, and a long interior arc exponentially close to the static optimum. The same viewpoint also clarifies what changes when some weights are only semidefinite: the order of the reduced equation may drop, some endpoint conditions may become incompatible, and polynomial or oscillatory modes may destroy the exponential turnpike. It also gives a natural meaning to certain endpoint constraints on the control and on finitely many derivatives of the control: such traces are not defined on the ambient L 2 control space, but they are meaningful on the smooth extremals selected by the reduced Euler-Lagrange equation. We formulate this principle as a general theorem and illustrate it in detail on the double integrator.
Read the full abstract

Mathematical Physics · Cross-listed from cond-mat.stat-mech

Exact solution of three-point functions in critical loop models

We propose an exact formula for three-point functions on the sphere in critical loop models with primary fields $V_{(r,s)}$ characterized by $2r$ legs and a parameter \(s\) that describes diagonal fields for $r=0$ and the momentum of legs for $r>0$. We demonstrate its validity in three ways: the conformal bootstrap method for 4-point functions, a transfer-matrix study of the lattice model, and a probabilistic method based on conformal loop ensemble and Liouville quantum gravity. This work provides a crucial missing piece for solving critical loop models and reveals a deep unity between three fundamental approaches to 2D statistical physics: transfer matrix, conformal field theory, and probability theory.
Read the full abstract

All authors Morris Ang, Gefei Cai, Jesper Lykke Jacobsen, Rongvoram Nivesvivat, Paul Roux, Xin Sun, Baojun Wu

Information Theory · Cross-listed from cs.CG

Selecting a Maximum Solow-Polasky Diversity Subset in General Metric Spaces Is NP-hard

The Solow–Polasky diversity indicator (or magnitude) is a classical measure of diversity based on pairwise distances. It has applications in ecology, conservation planning, and, more recently, in algorithmic subset selection and diversity optimization. In this note, we investigate the computational complexity of selecting a subset of fixed cardinality from a finite set so as to maximize the Solow–Polasky diversity value. We prove that this problem is NP-hard in general metric spaces. The reduction is from the classical Independent Set problem and uses a simple metric construction containing only two non-zero distance values. Importantly, the hardness result holds for every fixed kernel parameter $\theta_0>0$; equivalently, by rescaling the metric, one may fix the parameter to $1$ without loss of generality. A central point is that this is not a boilerplate reduction: because the Solow–Polasky objective is defined through matrix inversion, it is a nontrivial nonlinear function of the distances. Accordingly, the proof requires a dedicated strict-monotonicity argument for the specific family of distance matrices arising in the reduction; this strict monotonicity is established here for that family, but it is not assumed to hold in full generality. We also explain how the proof connects to continuity and monotonicity considerations for diversity indicators.
Read the full abstract

Category Theory

Higher exact dg-categories

We introduce the notion of an $n$-exact dg-category. This notion provides a higher analogue of Chen's exact dg-category, in the sense that the case where $n$ equals 1 recovers exact dg-categories. We prove that, under a suitable vanishing condition on the cohomologies of $\mathrm{Hom}$-complexes of an $n$-exact dg-category $\mathscr{A}$, its homotopy category admits a natural $n$-exangulated structure. Thus $n$-exact dg-categories provide dg-enhancements of $n$-exangulated categories. At the same time, our framework can be regarded as a dg-categorical generalization of $n$-exangulated categories applicable even without the vanishing condition. In the latter part of the article, we show that an $n$-cluster tilting subcategory of an exact dg-category naturally carries the structure of an $n$-exact dg-category. This result indicates that $n$-exact dg-structures provide an intrinsic dg-categorical axiomatization of $n$-cluster tilting subcategories, highlighting the advantages of studying dg-generalizations of $n$-exangulated categories.
Read the full abstract

Combinatorics

A counterexample to the conjecture on Biclique Partition number of Split Graphs and related problems

The biclique partition number of a graph \(G\), denoted \( \operatorname{bp}(G)\), is the minimum number of biclique subgraphs needed to partition the edge set of $G$. Lyu and Hicks \cite{lyu2023finding} posed the open problem of whether \( \operatorname{bp}(G) = \operatorname{mc}(G^c) - 1 \) holds for every co-chordal graph or split graph, where \( \operatorname{mc}(G^c) \) denotes the number of maximal cliques in the complement of \( G \). Such a result would extend the celebrated Graham–Pollak theorem to a more general class of graphs. In this note, we answer this problem in the negative by providing a counterexample using a split graph. We also construct an infinite family of counterexamples and prove some structural properties of biclique partitions of split graphs. Finally, we solve an open problem posed by Siewert \cite{siewert2000biclique} on the existence of singular \(n\)-tournaments with binary rank \(n\).
Read the full abstract

Analysis of PDEs

Positive Solutions for a Mixed Local-Nonlocal Problem with Semipositone Nonlinearity

In this article, we prove the existence of at least one positive solution for the mixed local-nonlocal semipositone problem \begin{equation*} \left\{ \begin{aligned} -\Delta_p u+ (-\Delta)^s_p u &= \lambda f(u) && \text{in } \Omega, u &= 0 && \text{in } \mathbb{R}^N \setminus \Omega, \end{aligned} \right. \end{equation*} using mountain pass arguments, comparison principles and regularity principles.
Read the full abstract

Statistics Theory · Cross-listed from stat.ML

Hierarchical Contrastive Learning for Multimodal Data

Multimodal representation learning is commonly built on a shared-private decomposition, treating latent information as either common to all modalities or specific to one. This binary view is often inadequate: many factors are shared by only subsets of modalities, and ignoring such partial sharing can over-align unrelated signals and obscure complementary information. We propose Hierarchical Contrastive Learning (HCL), a framework that learns globally shared, partially shared, and modality-specific representations within a unified model. HCL combines a hierarchical latent-variable formulation with structural sparsity and a structure-aware contrastive objective that aligns only modalities that genuinely share a latent factor. Under uncorrelated latent variables, we prove identifiability of the hierarchical decomposition, establish recovery guarantees for the loading matrices, and derive parameter estimation and excess-risk bounds for downstream prediction. Simulations show accurate recovery of hierarchical structure and effective selection of task-relevant components. On multimodal electronic health records, HCL yields more informative representations and consistently improves predictive performance.
Read the full abstract

Number Theory

There are infinitely many Hilbert cubes of dimension 3 in the set of squares

For an integer tuple $(a_0,a_1,\ldots,a_d)$ with $a_i\neq0$ for $i=1,\ldots,d$, we define the associated Hilbert cube by \[ H(a_0;a_1,\ldots,a_d) = \left\{ a_0+\sum_{i=1}^d \varepsilon_i a_i: \varepsilon_i\in\{0,1\} \right\}. \] We call $(a_0,a_1,\ldots,a_d)$ a $d$-dimensional Hilbert-cube presentation. We prove that there are $\gg N^{1/4}$ reduced Hilbert-cube presentations $(a_0,a_1,a_2,a_3)$ with $a_0,a_1,a_2,a_3\in[0,N]$ such that \[ H(a_0;a_1,a_2,a_3)\subset \mathcal{S}, \] where $\mathcal{S}=\{m^2:\;m\in\Z\}$. Moreover, we prove that for each $i, j\in\{0, 1, 2, 3\}$ with $i<j$, the set $$ \left\{\frac{a_{i}}{a_{j}}:\;H(a_{0}; a_{1}, a_{2}, a_{3})\subset \cal{S} \;\text{and}\;a_{i}>0\;\text{for}\;i=0, 1, 2, 3\;\right\} $$ is dense in the set of positive real numbers (in the Euclidean topology).
Read the full abstract

Analysis of PDEs

Polynomial Stability of a Type II Porous Thermoelastic System with Local Memory Damping

This paper studies the asymptotic behavior of a one-dimensional Type II porous thermoelastic system with a conservative porous structure and local memory damping applied to the elastic component. Using frequency domain resolvent estimates, we prove polynomial decay of the associated semigroup. Our results clarify the effect of local memory damping and provide a unified framework for partially damped porous thermoelastic systems
Read the full abstract

Optimization and Control

Distributed Algorithm for the Global Optimal Controller of Nonlinear Multi-Agent Systems

In this paper, we investigate the distributed optimal control problem for a kind of nonlinear multi-agent systems. In particular,both the state and the system dynamic structures of each agent are private and can only be shared among communicating agents.This type of information structure is inevitable in fields such as collaborative control for industrial confidentiality, and renders traditional distributed control methods using all systems' dynamic structures ineffective. The primary contribution is the proposal of a distributed algorithm for the global optimal controller under such practical information structure via distributed approximation of the Hamilton-Jacobi-Bellman equation. Practical numerical simulation demonstrates the effectiveness of the proposed algorithm.
Read the full abstract

All authors Ruixue Li, Wenjing Yang, Zhaorong Zhang, Xun Li, Juanjuan Xu

Combinatorics

Generic Rigidity and the Grassmannian

We characterize the generic rigidity of bar-joint frameworks in all dimensions. The conditions are constructed explicitly from the graph alone, independently of any placement, and are checked by exchange rules on Young tableaux. By following paths in a directed acyclic subgraph of the line graph, we build multisets of red and blue tableaux, called balancing conditions. Then generic rigidity is determined by counting if every semistandard tableau appears the same number of times red as it does blue.
Read the full abstract

Analysis of PDEs

The Stability of a Coupled Degenerate Wave System Under Boundary Control

In this paper, we investigate a system composed of two degenerate wave equations which are connected at one point. By introducing some inequalities on the weighted spaces and employing the frequency domain method, we prove that the system is polynomially stable,which depends on the degree of the degeneracy.
Read the full abstract

Mathematical Physics · Cross-listed from physics.soc-ph

Scale-free congestion clusters in large-scale traffic networks: a continuum modeling study

Recent empirical studies have reported that spatiotemporal congestion clusters in urban traffic exhibit scale-free statistics, with cluster size following a power-law distribution. In this study, we address whether macroscopic continuum descriptions of traffic flow are capable of generating such scale-free spatiotemporal congestion patterns. To this end, we analyze the second-order Aw-Rascle-Zhang model on directed networks under junction coupling. The governing equations are solved by a high-order discontinuous Galerkin scheme, and junction fluxes are determined by an optimization-based coupling procedure enforcing conservation and admissibility at intersections. Congestion is defined by thresholding the road-averaged density, and spatiotemporal clusters are extracted as connected components in space and time. Numerical experiments on lattice networks of varying sizes reveal that the cluster size follows a robust power-law distribution. Moreover, when rescaled by the linear system size inherent to the two-dimensional network geometry, the distribution collapses onto an approximately universal curve, indicating finite-size scaling governed by the linear system size. The observed power-law statistics and finite-size scaling are reminiscent of scale-invariant dynamics characteristic of self-organized criticality. These results demonstrate that macroscopic continuum traffic models can reproduce large-scale statistical features observed in real urban congestion dynamics.
Read the full abstract

All authors Yuki Chiba, Norikazu Saito, Yuki Ueda, Hiroaki Yoshida

Commutative Algebra

A characterization of Cohen-Macaulay rings in terms of levels of perfect complexes

Let $R$ be a commutative noetherian ring, and let $C$ be a semidualizing $R$-module. In this paper, we study levels of bounded complexes of finitely generated $R$-modules with respect to the full subcategory $\mathsf{G}_{C}(R)$ consisting of Gorenstein $C$-projective $R$-modules. Our main result provides a characterization of the Cohen-Macaulayness of $R$ in terms of the finiteness of levels of perfect complexes with respect to $\mathsf{G}_{C}(R)$. This recovers a recent theorem of Christensen, Kekkou, Lyle and Soto Levins on the Gorensteinness of $R$.
Read the full abstract

Spectral Theory

Toda flow with unbounded initial data

A Toda flow is constructed starting from a certain class of unbounded initial conditions including sequences growing with power order of less than 1. Unbounded ergodic sequences are allowed, and especially e̱ta}-ensembles matrix models in random matrix theory can be an initial data and they yiled invariant measures for the flow.
Read the full abstract

Dynamical Systems

A graph based advection framework for climate-driven species distribution

Climate change is reshaping species interactions and movement across fragmented landscapes. Despite this, most mathematical models assume random diffusion, overlooking the influence of directed movement. Here, we develop a graph based reaction-diffusion-advection framework explicitly incorporating directional movement induced by environmental gradients. Our results show while diffusion promotes overall population persistence across the network, advective movement induces asymmetric flows. It create population hotspots by directing individuals toward optimal niches, often associated with nodes of high in-degree. We demonstrate the interplay between advection strength and network topology in determining species persistence. Strong advection increase local extinction risk by accumulating populations toward favorable nodes. Additionally, loss of ecological corridors can disrupt directed flow within the network, thereby restricting species from favorable patches. We found that this disruption might not cause immediate extinction, rather forcing species to spread to the suboptimal patches. Our advection framework therefore efficiently captures how directional movement interacting with network topology governs species redistribution, hotspot formation, and predict extinction risk under environmental change.
Read the full abstract

Representation Theory

A Generalized Fourier Transform and a Smooth Analogue of Dunkl Operators

We introduce a deformation of the Fourier transform on $\mathbb{R}^N$ arising from a representation-theoretic construction associated with $\widetilde{SL}(2,\mathbb{R}) \times O(N)$ that still admits an underlying degree-one operator structure. More precisely, we construct a generalized Fourier transform $\mathcal{F}_b$, a non-local deformation $H_b$ of the Laplacian $\Delta$, and operators $D_{b,n}$ deforming the partial derivatives $\frac{\partial}{\partial x_n}$. We show that the operators $D_{b,n}$ and $x_n$ are compatible with the $\widetilde{SL}(2,\mathbb{R})$-representation in a way parallel to the classical case: for each $n$, the space spanned by $x_n$ and $D_{b,n}$ carries the standard representation of $\widetilde{SL}(2,\mathbb{R})$; in particular, the generalized Fourier transform $\mathcal{F}_b$ interchanges $D_{b,n}$ and $x_n$, and the $\mathfrak{sl}_2$-triple is recovered from quadratic expressions in these operators. We also establish the inversion formula for $\mathcal{F}_b$ and give explicit formulas for both $\mathcal{F}_b$ and $D_{b,n}$. In particular, $\mathcal{F}_b$ admits an explicit integral kernel representation, and $D_{b,n}$ is expressed as the sum of a differential term and a spherical integral term. Our construction might be viewed as a continuous analogue of Dunkl theory, with $O(N)$ playing the role of a reflection group.
Read the full abstract

Algebraic Geometry

Direct image and pullback of Parabolic vector bundles

Niels Borne established a natural correspondence between the parabolic vector bundles on curves and vector bundles on root stacks. The notions of direct image of parabolic vector bundles and pullback of parabolic vector bundles were studied in \cite{Alfaya_Biswas}. We show that these two notions correspond to the notions direct image of vector bundles on root stacks and pullback of vector bundles on root stacks respectively. Some applications of this correspondence are given.
Read the full abstract

Number Theory

Proof of a conjecture of Banerjee,Bringmann and Bachraoui on infinite families of congruences

Recently, Andrews and Bachraoui investigated congruences for certain restricted two-color partitions. They made two conjectures for Ramanujan type congruences and a vanishing identity for the limiting sequence. Very recently, Banerjee, Bringmann and Bachraoui confirmed these three conjectures by relating the corresponding generating function to modular forms and mock theta functions. At the end of their paper, they posed a conjecture on infinite families of congruences modulo 4 and 8 for the limiting sequence. The Banerjee-Bringmann-Bachraoui's conjecture implies the two conjectures given by Andrews and Bachraoui. In this note, we settle Banerjee-Bringmann-Bachraoui's conjecture on infinite famlies of congruences based on Banerjee-Bringmann-Bachraoui's results and an identity due to Waston.
Read the full abstract

Analysis of PDEs

Polynomial Stability of an Elastic Thin Plate on Non-Smooth Domain

This paper studies the polynomial stabilization of an elastic plate with dynamical boundary conditions on a non-smooth domain. To deal with the possible loss of solution regularity induced by boundary singularities, we formulate the problem as a precise variational framework. We prove that for domains with sufficiently small corner angles, the system retains the polynomial decay rate under standard geometric control conditions. In cases where larger corner angles lead to a significant regularity loss, we show that polynomial stability is recovered by introducing a feedback control at the corners.
Read the full abstract

Optimization and Control

A Continuous-Time Actor-Critic Framework for Jump-Diffusion Controls with Normalizing Flows

Continuous-time stochastic control with time-inhomogeneous jump-diffusion dynamics is central in finance and economics, but computing optimal policies is difficult under explicit time dependence, discontinuous shocks, and high dimensionality. We propose an actor-critic framework that serves as a mesh-free solver for entropy-regularized control problems and stochastic games with jumps. The approach is built on a time-inhomogeneous little q-function and an appropriate occupation measure, yielding a policy-gradient representation that accommodates time-dependent drift, volatility, and jump terms. To represent expressive stochastic policies in continuous-action spaces, we parameterize the actor using conditional normalizing flows, enabling flexible non-Gaussian policies while retaining exact likelihood evaluation for entropy regularization and policy optimization. We validate the method on time-inhomogeneous linear-quadratic control, Merton portfolio optimization, and a multi-agent portfolio game, using explicit solutions or high-accuracy benchmarks. Numerical results demonstrate stable learning under jump discontinuities, accurate approximation of optimal stochastic policies, and favorable scaling with respect to dimension and number of agents.
Read the full abstract

All authors Liya Guo, Ruimeng Hu, Xu Yang, Yi Zhu

Optimization and Control

Distributed Load Frequency Control of Multi-Area Smart Grid

In this paper, we investigate the distributed load frequency control problem in a multi-area smart grid under external load disturbances and measurement noise. The novelty lies in that the information privacy is fully taken into account, that is, the internal structural parameters and operational states of each area are not shared with non-neighboring areas, which makes traditional distributed optimal control methods ineffective. The main contribution is to propose a distributed algorithm for the global optimal power regulation command under information privacy constraints via distributed approximation of the control Riccati equation, the estimation Riccati equation, and the state estimation. Simulation results show that the proposed algorithm can approximate the performance of centralized optimal control, and the performance index under the proposed distributed controller is smaller than that under the commonly used distributed control.
Read the full abstract

All authors Wenjing Yang, Zhaorong Zhang, Xun Li, Juanjuan Xu

Numerical Analysis

HDG Methods for the two-dimensional Vector Laplacian

We introduce new hybridizable discontinuous Galerkin (HDG) methods for solving the two-dimensional vector Laplacian equation under three types of boundary conditions: electric, magnetic, and Dirichlet. The method is formulated on a first-order system form of the equations, in which the rotational and divergence of the electric field are introduced as auxiliary variables. We study the well-posedness of the method and prove that, when using piecewise polynomial approximations of degree $k \geq 0$, the error in the $L^2$ norm of the electric field converges at the optimal rate of $k+1$. Additionally, we prove that the $L^2$-errors of the auxiliary variables, the rotational and divergence, converge with order $k + 1/2$. We also show that the methods can be implemented in three different forms, corresponding to three distinct hybridizations based on the choice of the globally coupled unknowns among the numerical traces defined on the mesh skeleton. Finally, we provide numerical tests that not only validate the theoretical convergence rates but also consistently showcase the optimal convergence across all variables.
Read the full abstract

Optimization and Control

From Nonsmooth Minima to Smooth Branches via Heat Kernel Regularization

Many optimization problems in science and engineering involve objective functions that are nonsmooth at their minimizers. A common strategy is to trace a branch of minimizers of a regularized objective as the smoothing scale tends to zero; however, for nonsmooth functions, it is generally unclear whether such a branch can be continued and whether the associated continuation equation remains locally solvable. We study heat-kernel regularization and the resulting continuation equation along a local minimizing branch connected to a minimizer of the original objective. Under a global growth condition and a local leading-order description of the form $|x|^a$ with $1 \le a \le 2$, we first show that the regularized objective admits global minimizers and that any such minimizing branch localizes at the natural heat scale $O(\sqrt{t})$. We then prove that the asymptotic behavior of the regularized Hessian is determined by the local profile of the original objective: it remains uniformly positive definite in the quadratic case $a=2$, while in the subquadratic regime $1 \le a < 2$ its smallest eigenvalue grows at the controlled rate $t^{(a-2)/2}$. Consequently, the regularized Hessian remains asymptotically nondegenerate for all sufficiently small $t>0$, and the continuation equation remains locally solvable, even when the original objective does not admit a classical Hessian at the minimizer. Our results provide a rigorous second-order framework for continuation-based analysis in nonsmooth optimization by showing how heat regularization restores nondegeneracy near singular minimizers.
Read the full abstract