Statistics Theory · Cross-listed from stat.ML
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
Saturday 21 March
We highlight a striking difference in behavior between two widely used variants of coordinate ascent variational inference: the sequential and parallel algorithms. While such differences were known in the numerical analysis literature in simpler settings, they remain largely unexplored in the optimization-focused literature on variational inference in more complex models. Focusing on the moderately high-dimensional linear regression problem, we show that the sequential algorithm, although typically slower, enjoys convergence guarantees under more relaxed conditions than the parallel variant, which is often employed to facilitate block-wise updates and improve computational efficiency.
Information Theory
Deep Adaptive Rate Allocation in Volatile Heterogeneous Wireless Networks
Modern multi-access 5G+ networks provide mobile terminals with additional capacity, improving network stability and performance. However, in highly mobile environments such as vehicular networks, supporting multi-access connectivity remains challenging. The rapid fluctuations of wireless link quality often outpace the responsiveness of existing multipath schedulers and transport-layer protocols. This paper addresses this challenge by integrating Transformer-based path state forecasting with a new multipath splitting scheduler called Deep Adaptive Rate Allocation (DARA). The proposed scheduler employs a deep reinforcement learning engine to dynamically compute optimal congestion window fractions on available paths, determining data allocation among them. A six-component normalised reward function with weight-mediated conflict resolution drives a DQN policy that eliminates the observation-reaction lag inherent in reactive schedulers. Performance evaluation uses a Mininet-based Multipath Datagram Congestion Control Protocol testbed with traces from mobile users in vehicular environments. Experimental results demonstrate that DARA achieves better file transfer time reductions compared to learning-based schedulers under moderate-volatility traces. For buffered video streaming, resolution improvements are maintained across all tested conditions. Under controlled burst scenarios with sub-second buffer constraints, DARA achieves substantial rebuffering improvements whilst state-of-the-art schedulers exhibit near-continuous stalling.
Read the full abstract
Combinatorics
Deligne–Lusztig varieties, toric orbifolds, and the $q$-Klyachko algebra
We investigate the geometry behind the $q$-Klyachko algebra, introduced by Nadeau–Tewari. When $q$ is a prime power, we show that the $q$-Klyachko algebra is the image of the pullback map on Chow rings $\mathrm{CH}(\mathrm{Fl}_{n+1})\to\mathrm{CH}(\mathrm{DL}_n)$, where $\mathrm{DL}_n\subseteq \mathrm{Fl}_n$ is a compactified Deligne–Lusztig variety inside the complete flag variety $\mathrm{Fl}_{n+1}$. When $q$ is a positive rational number, we establish a Kähler package for the $q$-Klyachko algebra through inputs from toric geometry.
Read the full abstract
Operator Algebras
$k$-graph algebras are iterated Cuntz-Pimsner algebras – from the bottom up
We introduce a new method of expressing a $k$-graph $C^*$-algebra as a Cuntz-Pimsner algebra. Kumjian, Pask, and Sims have done this directly, using a linking algebra approach and a $(k-1)$-graph algebra. This can be iterated downward. Our process, on the other hand, starts at the bottom, with Pimsner's theorem for graph algebras, and iterates upward. We actually work with product systems over $\mathbb N^k$, and the result for $k$-graphs is a special case. Our iteration step involves a “decategorization” of a recent theorem showing that the Cuntz-Pimsner construction is functorial at the level of “enchilada categories”.
Read the full abstract
Probability · Cross-listed from q-bio.PE
Spectral Geometry and Heat Kernels on Phylogenetic Trees
We develop a unified spectral framework for finite ultrametric phylogenetic trees, grounding the analysis of phylogenetic structure in operator theory and stochastic dynamics in the finite setting. For a given finite ultrametric measure space $(X,d,m)$, we introduce the ultrametric Laplacian $L_X$ as the generator of a continuous time Markov chain with transition rate $q(x,y)=k(d(x,y))m(y)$. We establish its complete spectral theory, obtaining explicit closed-form eigenvalues and an eigenbasis supported on the clades of the tree. For phylogenetic applications, we associate to any ultrametric phylogenetic tree $\mathcal{T}$ a canonical operator $L_{\mathcal{T}}$, the ultrametric phylogenetic Laplacian, whose jump rates encode the temporal structure of evolutionary divergence. We show that the geometry and topology of the tree are explicitly encoded in the spectrum and eigenvectors of $L_{\mathcal{T}}$: eigenvalues aggregate branch lengths weighted by clade mass along ancestral paths, while eigenvectors are supported on the clades, with one eigenspace attached to each internal node. From this we derive three main contributions: a spectral reconstruction theorem with linear complexity $O(|X|)$; a rigorous geometric interpretation of the spectral gaps of $L_{\mathcal{T}}$ as detectors of distinct evolutionary modes, validated on an empirical primate phylogeny; an eigenmode decomposition of biological traits that resolves trait variance into contributions from individual splits of the phylogeny; and a closed-form centrality index for continuous-time Markov chains on ultrametric spaces, which we propose as a mathematically grounded measure of evolutionary distinctiveness. All results are exact and biologically interpretable, and are supported by numerical experiments on empirical primate data.
Read the full abstract
Optimization and Control
Mirror-Free Proximal Methods
We present a mirror-free mirror prox (MFMP) algorithm, which extends the classic approach of Nemirovski (2004) to allow for proximal-like updates without the explicit need for a mirror map. We further analyze the convergence of our method under suitable notions of relative smoothness and relative Lipschitzness, for which we introduce a relaxation of the standard Bregman divergence in terms of more general potential operators. Finally, we show how a strongly monotone variant of our method allows us to solve regularized Taylor-expansion subproblems that appear in both second- and third-order smooth min-max optimization.
Read the full abstract
Algebraic Geometry
Semiprojectivity and very stability in moduli of symplectic and orthogonal parabolic Higgs bundles
Let $X$ be a compact Riemann surface of genus $g \geq 2$, and let $D \subset X$ be a fixed finite subset. We prove the semiprojectivity of the moduli space of semistable symplectic or orthogonal parabolic Higgs bundles over $X$. We show that a stable symplectic parabolic bundle $E$ on $X$ is strongly very stable, meaning $E$ does not have any nonzero strongly parabolic nilpotent Higgs field, if and only if the symplectic parabolic Hitchin morphism induced on the affine space $$H^0(X,\mathrm{SPEnd}_\mathrm{Sp}(E) \otimes K(D))$$ is a proper morphism, where $\mathrm{SPEnd}_\mathrm{Sp}(E)$ denotes the set of symplectic strongly parabolic endomorphisms of $E$. We remark that the same criterion for very stability applies to the orthogonal case.
Read the full abstract
Mathematical Physics · Cross-listed from hep-th
Gaugings of Groupoids, Strings in Shadows, and Emergent Poisson $\sigma$-Models
The gauge principle is proposed for rigid Lie-groupoidal symmetries $G=>M$ of the Polyakov-Alvarez-Gawędzki 2$d$ non-linear $\sigma$-model with metric target $(M,g_M)$ and the WZ term given by a CS differential character coming from an abelian gerbe $\mathcal{G}$. The principle bases on the notion of principaloid bundle with connection $(P,\Theta)$, introduced by Strobl and the Author. The descent of the model to the shadow of $P$ is demonstrated to require a twisted $G$-equivariant structure on $\mathcal{G}$, prequantising a multiplicative extension $(H_M,\rho,0)$ of the gerbe's curvature $H_M$ to a 3-cocycle in the BSS cohomology of the groupoid's nerve. The descent is accompanied by a combined $g_M$-isometric/$\rho$-holonomic reduction of the structure group of $P$, and uses an augmentation of the original gerbe by a trivial one depending quadratically on $\Theta$. The latter couples to the field of the $\sigma$-model – in an extension of the scheme worked out for the action groupoid by Gawędzki, Waldorf and the Author – through the comomentum component of the Spencer pair of $\rho$, as described by Crainic et al. A fully fledged cohomological analysis of the relevant gauge anomaly and of inequivalent gaugings is presented. In the symplectic setting, the augmentation procedure is shown to lead to the emergence of the standard Poisson $\sigma$-model. Conversely, a coaugmentation of a (local) Poisson $\sigma$-model by a flat equivariant gerbe yields a novel field theory with the shadow of the underlying symplectic principaloid bundle as the configuration bundle, and with manifest lagrangean gauge symmetry. A simple Cartan-type associating mechanism is proposed to account for the reduction of the structure group of the principaloid bundle. The mechanism allows for the coupling of an arbitrary number of distinct charged matter-field species to a given gauge field.
Read the full abstract
Logic
Combinatorial Properties Related to the Higher Baumgartner's Axiom
We isolate two combinatorial properties, each expressible by a $\Pi_2$-sentence over the structure $(H(\omega_3),\in,\omega_1,\omega_2,\text{NS}_{\omega_2})$, such that each property is consistent with CH, and their conjunction together with $2^\omega \le \omega_2$ and $2^{\omega_1} = 2^{\omega_2} = \omega_3$ implies the existence of a c.c.c. forcing which forces the higher Baumgartner's axiom.
Read the full abstract
Mathematical Physics · Cross-listed from stat.ME
Weak-Form Recovery of Stochastic Generators
Data-driven discovery of governing equations is well developed for deterministic systems but much less so for stochastic ones. We present a sparse-symbolic algorithm that recovers both the drift $b(x)$ and the diffusion $a(x)$ of a stochastic differential equation from a single discretely sampled trajectory. It replaces the time-indexed test functions of weak-form SINDy with state-dependent spatial kernels. Fixed-width time-window weighting induces a persistent endogeneity bias: for the Ornstein-Uhlenbeck process it erases the drift coefficient entirely, whatever the window shape. Spatial kernels avoid this through a martingale strong-law argument, linked to the estimating-function theory of Bibby and Sorensen. We prove consistency of the drift and diffusion estimators for the Euler-Maruyama coefficients, a $T^{-1/2}$ central limit theorem for the drift, a finite-step correction that links drift and diffusion estimation, and a correction for additive Gaussian observation noise. The theorems assume ergodicity of the Euler-Maruyama chain, which holds for the OU and multiplicative systems but not for the cubic double well, so the double-well results are empirical. The method is validated on three benchmark SDEs (Ornstein-Uhlenbeck, a double well, and multiplicative diffusion), on eight further systems identified from single trajectories with elastic-net-stabilized sparse regression, and in a support-recovery experiment. Exact support recovery from one trajectory needs a long record (about $T \gtrsim 200$ on the double well) and is not guaranteed for shorter records or weakly identified systems.
Read the full abstract
Optimization and Control
Unfolding with a Wasserstein Loss
Data unfolding – the removal of noise or artifacts from measurements – is a fundamental task across the experimental sciences. Of particular interest are applications in physics, where the dominant approach is Richardson-Lucy (RL) deconvolution. The classical RL approach aims to find denoised data that, once passed through the noise model, is as close as possible to the measured data in terms of Kullback-Leibler (KL) divergence. This requires that the support of the measured data overlaps with the output of the noise model, a hypothesis typically enforced by binning, which introduces numerical error. As a counterpoint, the present work studies an alternative formulation using a Wasserstein loss. We establish sharp conditions for existence and uniqueness of optimizers, answering open questions of Li, et al., regarding necessary conditions for uniqueness in the case of transport map noise models. We then develop a provably convergent generalized Sinkhorn algorithm to compute approximate optimizers. Our algorithm requires only empirical observations of the noise model and measured data and scales with the size of the data, rather than the ambient dimension. Numerical experiments on one- and two-dimensional problems inspired by jet mass unfolding in particle physics demonstrate that the optimal transport approach offers robust, accurate performance compared to classical RL deconvolution, particularly when binning artifacts are significant.
Read the full abstract
Logic
An Alternative Approach to Formal Mathematics that Focuses on Communication and Accessibility
Formal mathematics is mathematics done within the framework of a formal logic. It offers major benefits to mathematicians as well as to computing professionals, engineers, and scientists who use mathematics in their work. The standard approach to formal mathematics, in which mathematics is done with the help of a proof assistant and all details are formally proved and mechanically checked, achieves these benefits and offers a very high level of assurance that the results produced are correct. However, since the main goal of the standard approach is certification, the proof assistants supporting the standard approach are generally complex, based on unfamiliar logics, difficult to learn how to use, and far removed from mathematical practice. Thus the standard approach does not adequately serve mathematics practitioners who are more interested in communicating mathematical ideas than in formally certifying their correctness or who prefer not to make the investment needed to gain proficiency in the use of a proof assistant. This paper presents an alternative to the standard approach that focuses on communication and accessibility, the two weaknesses of the standard approach. It is called the free approach to formal mathematics since it is free of the obligation to formally prove and mechanically check all details of a mathematical development. The paper argues that the free approach would serve the needs of the average mathematics practitioner much better than the standard approach. It describes an implementation of the free approach based on a logic named Alonzo, a practice-oriented version of Alonzo Church's formulation of simple type theory. And it calls for the mathematics community to develop logics, software, and libraries of formal mathematical knowledge to support the free approach and to train mathematics practitioners to use them.
Read the full abstract
Optimization and Control
Implicit Regularization of Large Neural Networks via Mean-Field Formulation
We propose a mathematical framework to explain implicit regularization from early stopping during the training of overparametrized neural networks. In the mean-field limit, the parameter distribution evolves according to a gradient flow on the space of probability measures. We show that these dynamics admit an equivalent McKean-Vlasov stochastic control formulation through the corresponding Hamilton-Jacobi-Bellman (HJB) equation. The control viewpoint yields a Dynamic Programming Principle (DPP), which we use to define a new metric on probability measures. This metric can be viewed as a mean-field generalization of the control representation of the Wasserstein-2 distance, and it naturally appears as a regularization term selected by early stopping. We further obtain non-asymptotic bounds describing how the induced regularization depends on the stopping time.
Read the full abstract
Dynamical Systems · Cross-listed from stat.ML
Auto-differentiable data assimilation: Co-learning of states, dynamics, and filtering algorithms
Data assimilation algorithms estimate the state of a dynamical system from partial observations, where the successful performance of these algorithms hinges on costly parameter tuning and on employing an accurate model for the dynamics. This paper introduces a framework for jointly learning the state, dynamics, and parameters of filtering algorithms in data assimilation through a process we refer to as auto-differentiable filtering. The framework leverages a theoretically motivated loss function that enables learning from partial, noisy observations via gradient-based optimization using auto-differentiation. We further demonstrate how several well-known data assimilation methods can be learned or tuned within this framework. To underscore the versatility of auto-differentiable filtering, we perform experiments on dynamical systems spanning multiple scientific domains, such as the Clohessy-Wiltshire equations from aerospace engineering, the Lorenz-96 system from atmospheric science, the generalized Lotka-Volterra equations from systems biology, and the two-layer quasi-geostrophic model from geophysical fluid dynamics. Finally, we provide guidelines for practitioners to customize our framework according to their observation model, accuracy requirements, and computational budget.
Read the full abstract
Differential Geometry
The first eigenvalue of embedded minimal hypersurfaces in the unit sphere
In this article, we prove that for an embedded minimal hypersurface $\Sigma^{m}$ in $S^{m+1}$, the first eigenvalue $\lambda_1$ of the Laplacian operator on $\Sigma$ satisfies: $$\lambda_1> \frac{m}{2}+G(m, |A|_{\max}, |A|_{\min} ) ,$$ where $|A|_{\max}$ and $|A|_{\min}$ denote the maximum and minimum of the norm of the second fundamental form on $\Sigma$, respectively; $G(m, |A|_{\max}, |A|_{\min} )$ is a positive constant that depends only on $m,|A|_{\max}, |A|_{\min}$. In particular, when the norm $|A|$ of the second fundamental form is constant, we can obtain a gap depending only on $m$, i.e., $$\lambda_1>\left(\frac{1}{2}+ c \right)m ,$$ where $c$ is a positive absolute constant. This improves Choi and Wang's previous result \cite{chw1983first} that $\lambda_1\geq \frac{m}{2}$. Our result shows that one can improve Choi and Wang's result directly without proving Chern's conjecture. This also generalizes Tang and Yan's work \cite{tangyan2013isoparametric}. Based on the proof of the result above, using the lower bound of the first Steklov eigenvalue, we prove that if the norm $|A|$ of the second fundamental form is constant, then $$|A| \leq \frac{C(m)\textup{Volume}(\Sigma)}{\textup{Volume}(S^m)},$$ where $C(m)$ is a constant that depends only on $m$. This provides a uniform estimate for the scalar curvature of embedded minimal hypersurfaces with constant norm of the second fundamental form. Moreover, this may be useful for Chern's problem.
Read the full abstract
Numerical Analysis · Cross-listed from cs.MS
Implementation of QR factorization of tall and very skinny matrices on current GPUs
We consider the problem of computing a QR (or QZ) decomposition of a real, dense, tall and very skinny matrix. That is, the number of columns is tiny compared to the number of rows, rendering most computations completely or partially memory-bandwidth limited. The paper focuses on recent NVIDIA GPGPUs still supporting 64-bit floating-point arithmetic, but the findings carry over to AMD GPUs as well. We discuss two basic algorithms: Methods based on the normal equations (Gram matrix), in particular Cholesky-QR2 and SVQB, and the "tall-skinny QR" (TSQR), based on Householder transformations in a tree-reduction scheme. We propose two primary optimization techniques: Avoiding the write-back of the Q factor ("Q-less QR"), and exploiting fast local memory (shared memory on GPUs). We compare a straight-forward implementation of Gramian-based methods, and a more sophisticated TSQR implementation, in terms of performance achieved, time-to-solution, and implementation complexity. By performance modelling and numerical experiments with our own code and a vendor-optimized library routine, we demonstrate the crucial need for specialized methods and implementations in this memory-bound to transitional (memory/compute-bound) regime, and that TSQR is competitive in terms of time-to-solution, but at the cost of an investment in low-level code optimization.
Read the full abstract
Number Theory
A refined non-vanishing of the $p$-adic logarithm of a rational point on an abelian variety
Inspired by a beautiful formula of Bertolini, Darmon, and Prasanna – the oft-termed BDP formula – we address questions about the non-vanishing of non-torsion points under $p$-adic logarithms of abelian varieties. We largely consider situations most applicable to ${\mathrm GL}_2$-type abelian varieties associated with Hilbert modular newforms and Heegner points. Not surprisingly, the main tool employed is the $p$-adic analytic subgroup theorem.
Read the full abstract
Complex Variables
A Solution to a Problem of Rubel on Two-Parameter Normal Families of Entire Functions
We construct an entire function $ F(z,a,b)\in \mathcal{O}(\mathbb{C}^3) $ such that the family $$ \{F(\,\cdot\,,a,b):a,b\in\mathbb{C}\} $$ of entire functions of \(z\) is normal on \(\mathbb{C}\), while \(F\) does not factor through a single entire parameter. This solves a problem of L.~A.~Rubel concerning Liouville-type rigidity. In fact, our example satisfies the stronger condition $$ F_bF_{a,z}-F_aF_{b,z}\neq 0 \qquad\text{on }\mathbb{C}^3. $$ The geometric core of the construction is a Fatou–Bieberbach domain contained in the thin region $$ \{(u,v)\in\mathbb{C}^2:|u-v^2|<1+|v|\}. $$ We obtain this domain from the basin of attraction of an explicit polynomial automorphism of \(\mathbb{C}^2\), together with the theorem of Rosay and Rudin on attracting basins.
Read the full abstract
Numerical Analysis
Parallel-in-iteration optimization using multigrid reduction-in-time
Standard gradient-based iteration algorithms for optimization, such as gradient descent and its various proximal-based extensions to nonsmooth problems, are known to converge slowly for ill-conditioned problems, sometimes requiring many tens of thousands of iterations in practice. Since these iterations are computed sequentially, they may present a computational bottleneck in large-scale parallel simulations. In this work, we present a "parallel-in-iteration" framework that allows one to parallelize across these iterations using multiple processors with the objective of reducing the wall-clock time needed to solve the underlying optimization problem. Our methodology is based on re-purposing parallel time integration algorithms for time-dependent differential equations, motivated by the fact that optimization algorithms often have interpretations as discretizations of time-dependent differential equations (such as gradient flow). Specifically in this work, we use the parallel-in-time method of multigrid reduction-in-time (MGRIT), but note that our approach permits in principle the use of any other parallel-in-time method. We numerically demonstrate the efficacy of our approach on two different model problems, including a standard convex quadratic problem and the nonsmooth elastic obstacle problem in one and two spatial dimensions. For our model problems, we observe fast MGRIT convergence analogous to its prototypical performance on partial differential equations of diffusion type. Some theory is presented to connect the convergence of MGRIT to the convergence of the underlying optimization algorithm. Theoretically predicted parallel speedup results are also provided.
Read the full abstract
Optimization and Control
On hyperbolic PDEs, filtered feedback control laws, and fractal-like stability crossing curves
The paper addresses the boundary control of a class of hyperbolic PDEs, based on an equivalent representation in terms of an integral-difference equation. The situation is considered where direct compensation of reflection terms induces a fragile closed-loop system, in the sense of lack of strong stability. This is theoretically resolved by adding a low-pass filter to the control law, but the choice of its cut-off frequency is crucial in balancing robustness at high frequencies and performance at low frequencies. First, the maximum stability interval in parameter $T$ is determined, with $T$ the inverse of the filter's cutoff frequency. Next, model mismatch on the PDE parameters is considered and a sufficient stability condition is derived in terms of allowable mismatch and cut-off frequency, satisfied in a region in the combined parameter space with a conic shape around $T=0$. Finally, this qualitative behavior is confirmed by exact stability charts for a special case where all model mismatch is contained into one parameter. It is highlighted that the set of stability crossing curves exhibits a fractal-like structure, which is explained using a limit system with discrete delays.
Read the full abstract
Number Theory
Upper and lower estimates for integer complexity
Let $\|n\|$ stand for the integer complexity of the number $n$, i.e. for the least number of $1$'s needed to write $n$ using arbitrary many additions, multiplications, and parentheses. The two-sided inequality $3\log_3 n\leq\|n\|\leq 3\log_2 n$ for all $n$ is well known and reveals the logarithmic behaviour of the complexity function $\|n\|$. While the lower bound $3\log_3 n$ is attained infinitely many times at powers of $3$, the best upper estimate is still unknown, although there are some improvements of the trivial bound $3\log_2 n$. Besides, for $``$typical$"$ numbers, i.e. for almost all numbers $n$, the better inequality $\|n\|\leq C_{avg}\log n$ holds, where, importantly, $C_{avg}\approx 3.236<\sup_{n} \frac{\|n\|}{\log n}$. We show that in fact $\|n\|\leq C_{avg}\log n+o(\log n)$ as $n\to\infty$, which, in particular, yields that $\limsup\limits_{n\to\infty}\frac{\|n\|}{\log n}\leq C_{avg}$. We also obtain the first nontrivial lower bound $\|n\|\geq 3.06\log_3 n$ for almost all numbers $n$.
Read the full abstract
Optimization and Control · Cross-listed from cs.LG
Incentive-Aware Federated Averaging with Performance Guarantees under Strategic Participation
Federated learning (FL) is a communication-efficient collaborative learning framework that enables model training across multiple agents with private local datasets. While the benefits of FL in improving global model performance are well established, individual agents may behave strategically, balancing the learning payoff against the cost of contributing their local data. Motivated by the need for FL frameworks that successfully retain participating agents, we propose an incentive-aware federated averaging method in which, at each communication round, clients transmit both their local model parameters and their updated training dataset sizes to the server. The dataset sizes are dynamically adjusted via a Nash equilibrium (NE)-seeking update rule that captures strategic data participation. We analyze the proposed method under convex and nonconvex global objective settings and establish performance guarantees for the resulting incentive-aware FL algorithm. Furthermore, under a merely monotone game setting, we consider a welfare loss minimization framework and establish asymptotic convergence of the scheme. Numerical experiments on the MNIST and CIFAR-10 datasets demonstrate that agents achieve competitive global model performance while converging to stable data participation strategies.
Read the full abstract
Combinatorics
Neutral-Fermion constructions of factorial $gp$-and $gq$-Functions
We develop neutral-fermionic constructions for the factorial $gp$-and $gq$-functions introduced by Nakagawa and Naruse, which are respectively dual to the factorial $GQ$- and $GP$-functions of Ikeda and Naruse. In particular, we realize the factorial $GP$-, $GQ$- and $gq$-functions as vacuum expectation values. As applications, we obtain, Jacobi–Trudi type determinantal formulas for the transition coefficients between functions with different equivariant parameters for $gq$ and its dual $GP$, as well as a Pfaffian formula for the factorial $gq$-functions. We further prove a remarkable coincidence among the transition coefficients for parameter changes for $gp$, $gq$, $GQ$, and $GP$. These coefficients admit a description in terms of factorial Grothendieck polynomials of type A.
Read the full abstract
Differential Geometry
A sharp inequality relating scalar curvature, bottom spectrum, and the relative $\widehat{A}$-cowaist on complete manifolds
We introduce a refined notion of relative $\widehat{A}$-cowaist, extending the framework of Cecchini–Zeidler to complete manifolds, which may be noncompact and have compact boundary. We establish a sharp inequality relating this invariant to the scalar curvature and the bottom spectrum of the Laplacian. As a consequence, we partially answer a question raised by Shi, and we obtain, in all dimensions, sharp upper bounds for the bottom spectrum under scalar curvature lower bounds. In particular, this removes the bounded geometry assumption from a result of Wang–Zhu. Our approach is based on deformed Dirac operators.
Read the full abstract
Algebraic Topology
Discrete Coefficients and Open Invariant Covers in the Homology of Ample Groupoids
The homology of an ample groupoid is computed from the complex of compactly supported continuous functions on the nerve. Two hypotheses routinely imposed on this complex behave in opposite ways. We show that the comparison map from the integral chain complex tensored with the coefficient group to the chain complex with coefficients is always injective, and that it is surjective if and only if every compactly supported continuous function into the coefficient group is locally constant. The universal coefficient theorem for ample groupoids is therefore a discrete coefficient statement, and it already fails for the real numbers on the Cantor set. We then show that a cover of the unit space by two clopen invariant subsets forces the groupoid to split as a disjoint union of three clopen invariant reductions, so that the associated Mayer-Vietoris sequence has vanishing connecting maps and computes nothing. Dropping closedness repairs this, since two open invariant subsets covering the unit space give a Mayer-Vietoris sequence for every topological abelian coefficient group. For an integer action on a two point compactification of the integers we compute that sequence and find that its connecting map is an isomorphism of infinite cyclic groups, which no cover of a unit space by clopen invariant subsets can achieve.
Read the full abstract
Numerical Analysis
Accelerated Nehari manifold optimization methods for computing ground state solutions of coupled elliptic systems
This paper presents a class of efficient algorithms for computing the ground state solutions of a semilinear elliptic system, which are unstable saddle points of the variational functional. Variational arguments show that these unstable saddle points can be characterized as the local minimizers of the variational functional constrained to the Nehari manifold $\mathcal{N}$. The Nehari manifold optimization method (NMOM) proposed in [Z. Chen, W. Liu, Z. Xie, and W. Yi. SIAM J. Sci. Comput., 47(4): A2098-A2126, 2025] provides a Riemannian gradient descent framework on $\mathcal{N}$ for such constrained minimization problems. To deal with both the intrinsic instability of the solutions and the increased computational complexity introduced by the coupling between components, we combine the ideas from the NMOM and the Nesterov-type acceleration to develop a new efficient Riemannian accelerated gradient algorithm on $\mathcal{N}$ (RAG-$\mathcal{N}$). The key idea is to perform an easy-to-implement nonlinear extrapolation step on $\mathcal{N}$, followed by a Riemannian steepest-descent update at the extrapolated point. To enhance the robustness, we further incorporate a nonmonotone step-size search strategy into the RAG-$\mathcal{N}$ algorithm, obtaining a variant with improved stability. Numerical experiments show that the RAG-$\mathcal{N}$ algorithms substantially reduce the number of iterations compared with the Riemannian steepest descent algorithm of NMOM. Finally, we apply the RAG-$\mathcal{N}$ algorithms to compute the ground state solutions of semilinear elliptic systems with two, three and four components, and investigate their behavior under different coupling coefficients and various settings. Specifically, a phase diagram separating fully nontrivial and semi-trivial ground state solutions in the coupling-parameter space is numerically obtained.
Read the full abstract
Logic
Inquisitive first-order logic is neither compact nor recursively axiomatizable
Inquisitive first-order logic is an extension of classical first-order logic with formulas regimenting first-order questions, such as "whether all objects are P", "which objects are P", and "what is one object that is P". Since it was first developed in 2009, two major meta-theoretical questions about this logic have remained open, in spite of significant efforts. The first concerns compactness: if a conclusion follows from a set of premises, does it always follow from some finite subset? The second concerns the computational status of validity: is the set of validities recursively enumerable, or equivalently, does the logic admit a recursive axiomatization? We settle both questions in the negative, showing that inquisitive first-order logic is neither compact nor recursively axiomatizable. Furthermore, we prove that it violates another signature property of first-order logic, namely, Craig interpolation. We discuss the significance of our results, and show how to extend them to a closely related logic, viz., inquisitive team logic.
Read the full abstract
Numerical Analysis
Asymptotic error distribution of Mittag–Leffler Euler method for a fractional stochastic differential equation
In this paper, we investigate the asymptotic distribution of the normalized error for the Mittag–Leffler Euler (MLE) method applied to a class of multidimensional fractional stochastic differential equations. These equations are reformulated as stochastic Volterra equations (SVEs) featuring a non-diagonal, matrix-valued kernel $K(u)=u^{\alpha-1}E_{\alpha,\alpha}(Au^{\alpha})$ with singular exponent $\alpha \in (\frac{1}{2}, 1)$. To enhance computational efficiency, the singular kernel is discretized using the left-rectangle rule, posing technical challenges for the theoretical analysis. To address this, we introduce an auxiliary $K$-undiscretized scheme to bridge the gap between the exact solution and the MLE method, integrating Jacod's stable convergence theory for conditional Gaussian martingales with methodologies developed for SVEs. To the best of our knowledge, this is the first work to establish the asymptotic error distribution for numerical methods incorporating non-diagonal matrix-valued kernels.
Read the full abstract
Analysis of PDEs
Generalized Reducibility and Growth of Sobolev Norms
We introduce the concept of generalized reducibility, which provides a flexible framework for analyzing the long-time behavior of solutions to quadratic quantum Hamiltonians. As an application of this notion, for many prescribed sub-exponential growth rates $f(t)$, either monotone or oscillatory, we explicitly construct time-decaying perturbations of the one-dimensional quantum harmonic oscillator such that the Sobolev norms of solutions grow at the rate $f(t)$.
Read the full abstract
Dynamical Systems
Symplectic blenders near whiskered tori and persistence of saddle-center homoclinics
A blender is a hyperbolic basic set such that the projection of its stable/unstable set onto some center subspace has a higher topological dimension than the set itself. We prove that, for any $C^s$ symplectic diffeomorphism (where $s=2,\dots\infty,\omega$), if it has a one-dimensional whiskered torus with a homoclinic orbit, then a symplectic blender can be created by an arbitrarily $C^s$-small perturbation. Using this result, we show that the non-transverse homoclinic intersection between the invariant manifolds of a saddle-center periodic point is persistent, in the sense that the original system lies in the $C^s$-closure of a $C^1$-open set of symplectic diffeomorphisms where those having saddle-center homoclinics are dense. Our results also hold in the corresponding continuous-time settings.