Analysis of PDEs
Unique continuation inequalities for the Dunkl-Schrödinger equation via uncertainty principles
Xingyu Zhao, Hui Xu and Zhiwen Duan
In this paper, we establish unique continuation inequalities at two time points for the Dunkl–Schrödinger equation. The proof is based on quantitative uncertainty principles for the Dunkl transform. In particular, we prove that pairs of (\varepsilon,k)-thin sets form strong annihilating pairs for the Dunkl transform, which yields quantitative unique continuation properties for solutions to the Dunkl–Schrödinger equation.
Read the full abstract ▾
Number Theory
On some constancy of Hecke eigensystems for Drinfeld cuspforms of finite slope
Shin Hattori
Let $p$ be a rational
prime, let $q>1$ be a $p$-power integer, let $\mathbb{F}_q$ be the field of $q$ elements and let $A=\mathbb{F}_q[t]$ be the
polynomial ring over $\mathbb{F}_q$. Let $\mathfrak{n}\in A$ be a nonzero element and let $\wp\in A$ be a monic irreducible
polynomial of positive degree. Let $k\geq 2$ and $r\geq 1$ be integers. Let $S_k(\Gamma_1(\mathfrak{n}\wp^r))$ be the space of Drinfeld cuspforms of level $\Gamma_1(\mathfrak{n}\wp^r)$ and weight $k$. In this paper, we prove that the multiplicity of a Hecke eigensystem of finite $\wp$-slope in $S_k(\Gamma_1(\mathfrak{n}\wp^r))$ is equal to $q^{(r-1)\mathrm{deg}(\wp)}$ times that in $S_k(\Gamma_1(\mathfrak{n}\wp))$. In particular, this shows that a Hecke eigensystem of finite $\wp$-slope appears in $S_k(\Gamma_1(\mathfrak{n}\wp^r))$ if and only if it appears in $S_k(\Gamma_1(\mathfrak{n}\wp))$.
Read the full abstract ▾
Analysis of PDEs
Global Regularity of Axisymmetric Navier-Stokes Equations with NHL Boundary Conditions under a Critical Smallness Condition
Tsz-Lik Chan
We investigate the global regularity problem for the three-dimensional incompressible Navier-Stokes equations restricted to axisymmetric flows in a finite cylinder $D = \{(r,\theta,x_3): 0 \le r \le 1, 0 \le \theta < 2\pi, 0 \le x_3 \le 1\}$, subject to the Navier-Hodge-Lions (NHL) boundary condition. While global existence of smooth solutions is known in the swirl-free case, the presence of swirl ($v_\theta \neq 0$) introduces
vortex stretching that may potentially lead to finite-time singularity formation. In this work, we prove that if the initial data satisfy a scaling-invariant smallness condition of the form \[ \frac{9C_1C_3^{1/2}}{4}\left(\frac{1}{2}\|V_0\|_{L^4}^4 + \|\Omega_0\|_{L^2}^2\right)^{1/4}\|\Gamma_0\|_{L^4} \le \frac{1}{4}, \] where $V = v_\theta/\sqrt{r}$, $\Omega = \omega_\theta/r$, $\Gamma = r v_\theta$, and $C_1, C_3$ are explicit constants given in this paper, then the solution remains globally regular for all time. The proof proceeds via a transformed system for $\Omega$ and $V$, leveraging a maximum principle for $\Gamma$, refined Agmon-type inequalities to control $\|v_r/r\|_{L^\infty}$, and delicate boundary analysis of the finite cylinder geometry. Key energy estimates yield $L^\infty_T L^4_x$ bounds for all velocity components, which fall within the regularity class, thereby precluding finite-time blow-up. The result extends the known criticality theory for axisymmetric Navier-Stokes flows to the setting of NHL boundary conditions, which are physically relevant for flows with stress-free or slip-type constraints on lateral and horizontal boundaries.
Read the full abstract ▾
Representation Theory
Representation theory of the real Gelfand order and real Harish-Chandra modules for $\mathsf{SL}_2(\mathbbm{R})$
Igor Burban and Yuriy Drozd
In this article we study the principal block of the category of real Harish-Chandra modules for the group $\mathsf{SL}_2(\RR)$ and relate it to the category of finite dimensional modules over the so-called real Gelfand order. We describe several distinguished classes of the corresponding indecomposable representations.
Read the full abstract ▾
Numerical Analysis
McLachlan-projected reduced dynamics for ill-posed Schrödingerized backward diffusion
Jeongbin Jo
Backward
diffusion is a prototype ill-posed
evolution: high spatial frequencies grow exponentially in time, so mesh-based time marching without explicit regularization is quickly overwhelmed by noise. Schrödingerization embeds the semidiscrete generator into Hermitian dynamics on an extended space; we ask whether McLachlan projection onto a fixed low-dimensional frame supplies a structured regularizer whose error budget can be read from a projection defect that separates full lifted propagation from the reduced trajectory. We prove uniqueness of the reduced flow, Gram-norm conservation, a lifted–reduced gap bound in terms of that defect, and perturbation estimates that highlight overlap-
matrix conditioning when
matrix elements are estimated statistically. We also spell out a fair classical baseline – spectral low-pass or Tikhonov filtering on the same semidiscrete model, with bandwidth or ridge strength matched to the information content of the chosen frame – so numerical contrasts isolate the Schrödingerized reduced pipeline rather than an unregularized Crank–Nicolson march that mainly showcases blow-up. A calibrated one-dimensional
benchmark pairs a spectrally truncated reference with snapshot-built subspace
evolution and finite-shot Qiskit Aer estimation, illustrating how lift, projection, and sampling layers contribute differently to the overall error.
Read the full abstract ▾
Rings and Algebras
Separability for relative extensions of object unital strongly groupoid graded rings
Zaqueu Cristiano and Patrik Lundström
We prove that if $R$ is a ring that is object unital and strongly graded by a groupoid $\Gamma$, and if $\Delta$ is a wide subgroupoid of $\Gamma$, then $R/R_\Delta$ is separable if and only if, for each $e \in \Gamma_0$, there exist $f \in [e]$ and $r \in C_{R_0}(R_\Lambda) := \{ x \in R_0 \mid xy = yx \text{ for all } y \in R_\Lambda \}$ with ${\rm tr}_{\Gamma/\Delta}^f(r) = 1_{R_f}$. Here, $\Gamma_0$ denotes the set of objects of $\Gamma$, $[e]$ the connected component of $\Gamma_0$ containing $e$, $\Lambda$ the isotropy groupoid of $\Delta$, and ${\rm tr}_{\Gamma/\Delta}^f$ the relative trace map at $f$. This result simultaneously generalizes earlier
theorems on separability for
matrix rings and group-graded rings due to DeMeyer-Ingraham, N{\v a}st{\v a}sescu, Van den Bergh, Van Oystaeyen, Miyashita, Theohari-Apostolidi, and Vavatsoulas, as well as results on groupoid-graded rings due to Cala, Lundström, and Pinedo. As an application, we consider separability for object crossed products, including object twisted groupoid rings, classical groupoid rings and
matrix rings, as well as crossed product
algebras defined by infinite separable field extensions.
Read the full abstract ▾
Optimization and Control
A Benders Decomposition Approach for the k-Defensive Domination Problem
Bilge Varol, T{\i}naz Ekim and Kübra Tan{\i}nm{\i}ş
The k-defensive domination problem is a powerful modeling tool for strategic decision-making in network security and disaster/emergency management, where multiple nodes may be simultaneously under attack. Despite its practical relevance, the problem has been poorly studied, largely due to its high computational difficulty. This study investigates the application of Benders decomposition to the k-defensive domination problem, aiming to improve computational efficiency over standard integer programming formulations. Several cut generation strategies, including a combinatorial approach and the simultaneous addition of multiple cuts, are proposed. Theoretical results on the strength of feasibility cuts are presented. In addition, two novel enhancement strategies are proposed: a clique-cover-based heuristic, the first feasible solution method in the literature for this problem, achieving up to 98% improvement over the trivial upper bound, and an initial cut generation heuristic that in some cases resolves the problem without further branching. Experiments on Erdos-Renyi, chordal, and Barabasi-Albert instances show that the
algorithm variant involving all of the proposed components is able to solve instances that remain out of reach for classical formulations and standard Benders decomposition.
Read the full abstract ▾
Algebraic Geometry
Duality for dormant opers of classical types B and C
Yasuhiro Wakabayashi
A $\mathfrak{g}$-oper for a simple Lie
algebra $\mathfrak{g}$ is a specific type of flat principal bundle on an algebraic curve. When the base field is of
prime characteristic $p$, those with vanishing $p$-
curvature are called dormant $\mathfrak{g}$-opers, and they form finite and geometrically meaningful
moduli spaces. In earlier work, a canonical duality was established between dormant $\mathfrak{sl}_n$-opers and dormant $\mathfrak{sl}_{p-n}$-opers. This duality has provided effective tools for the study of higher-rank cases, as well as for the computation and structural understanding of the associated enumerative invariants. The main result of this paper extends this duality phenomenon to classical Lie
algebras of type B and C. More precisely, under the numerical condition $p-1 = 2 (\ell +m)$, we construct a canonical isomorphism between the
moduli spaces of dormant $\mathfrak{so}_{2\ell +1}$-opers and dormant and $\mathfrak{sp}_{2m}$-opers with prescribed symmetric radii.
Read the full abstract ▾
Optimization and Control · Cross-listed from cs.AI
Unleashing LLMs in Bayesian Optimization: Preference-Guided Framework for Scientific Discovery
Xinzhe Yuan, Zhuo Chen, Jianshu Zhang et al. (+4 more)
Scientific discovery is increasingly constrained by costly experiments and limited resources, underscoring the need for efficient optimization in AI for science.
Bayesian Optimization (BO), though widely adopted for balancing exploration and exploitation, often exhibits slow cold-start performance and poor scalability in high-dimensional settings, limiting its applicability in real-world scientific problems. To overcome these challenges, we propose LLM-Guided
Bayesian Optimization (LGBO), the first LLM preference-guided BO framework that continuously integrates the semantic
reasoning of large
language models (LLMs) into the optimization loop. Unlike prior works that use LLMs only for warm-start initialization or candidate generation, LGBO introduces a region-lifted preference mechanism that embeds LLM-driven preferences into every iteration, shifting the surrogate mean in a stable and controllable way. Theoretically, we prove that LGBO does not perform significantly worse than standard BO in the worst case, while achieving significantly faster convergence when preferences align with the objective. Empirically, LGBO consistently outperforms existing methods across diverse dry
benchmarks in physics, chemistry, biology, and materials science. Most notably, in a new wet-lab optimization of Fe-Cr battery electrolytes, LGBO attains 90% of the best observed value within 6 iterations, whereas standard BO and existing LLM-augmented baselines require more than 10. Together, these results suggest that LGBO offers a promising direction for integrating LLMs into scientific optimization workflows.
Read the full abstract ▾
All authors Xinzhe Yuan, Zhuo Chen, Jianshu Zhang, Huan Xiong, Nanyang Ye, Yuqiang Li, Qinying Gu
Algebraic Geometry
Genus formulas for dormant modular curves and asymptotic behavior of their function fields
Kohei Aoyama, Youhei Morita and Yasuhiro Wakabayashi
Towers of algebraic function fields over finite fields play a fundamental role in arithmetic geometry and coding theory. Classical examples arising from modular and Drinfeld modular curves exhibit asymptotically good behavior. In this paper, we introduce an analogous construction derived from the
moduli spaces of higher-level dormant $\mathrm{PGL}_2$-opers of prescribed radii on $4$-pointed stable curves of genus $0$. These spaces, which we refer to as dormant modular curves, form projective systems under level reduction. Building on previous results in the
moduli theory of dormant opers, we establish an explicit formula for computing the genera of these curves. This formula allows us to study the asymptotic behavior of the corresponding towers of function fields and to compare them with the classical modular and Drinfeld modular cases.
Read the full abstract ▾
Number Theory
Decay of weighted cusp counts for congruence subgroups of $SL_2$ over number fields
Shengyuan Zhao
For congruence subgroups commensurable with $\operatorname{SL}_2$ over number fields, we study cusp counts with arithmetic multiplicities. We prove that the ratio of the total weighted cusp count to the group index is bounded by a negative power of the norm of the congruence level, with an exponent that can be chosen explicitly in terms of the degree of the number field. This is a generalization of a
theorem of Cox–Parry over rational numbers. The main input is an explicit local orbit estimate for arbitrary subgroups of exact level in $\operatorname{SL}_2(\mathcal O_K/\mathfrak p^e)$, uniform over all number fields of fixed degree. This estimate is covered by a general result of Finis–Lapid. We give a new explicit proof in our setting based on an analysis reminiscent of additive combinatorics.
Read the full abstract ▾
Classical Analysis and ODEs
Gabor unconditional bases and frames in $L^p(\mathbb{R})$
Nir Lev and Anton Tselishchev
We consider the following problem: given a set $\Lambda \subset \mathbb{R} \times \mathbb{R}$ and $p \neq 2$, does there exist a function $g \in L^p(\mathbb{R})$ such that the Gabor system $\{g(x-t) e^{2 \pi isx}\}$, $(t,s) \in \Lambda$, consisting of time-frequency shifts of $g$, forms an unconditional basis or unconditional Schauder frame in the space $L^p(\mathbb{R})$? We completely resolve this question for $p>2$; in particular, we characterize the sets $\Lambda$ such that an unconditional Schauder frame of this form exists. We also prove a Balian-Low type result, showing that the window function $g$ cannot enjoy mild continuity and decay conditions. For $1<p<2$, we prove that a Gabor system cannot form an unconditional basis or unconditional Schauder frame in $L^p(\mathbb{R})$ if the set $\Lambda$ satisfies a natural separation condition.
Read the full abstract ▾
Optimization and Control
Control-Channel Informativity for Koopman EDMDc under Behavior-Policy Data
Yue Wu
Extended dynamic mode decomposition with control (EDMDc) is often trained from trajectories generated by a behavior policy or a pre-existing feedback controller. Such data can predict the observed behavior accurately while failing to identify how new input commands change the lifted state. This paper studies that failure as a control-channel informativity problem. We introduce a conditional intervention certificate, defined as the residual input covariance after projecting the input data away from the active lifted-state feature span. The certificate is the Schur complement of the lifted-state block in the EDMDc information
matrix. We prove that its strict positivity is necessary and sufficient for finite-sample sample- identifiability of the lifted control-channel block. If the certificate vanishes, distinct lifted models agree on every collected transition but disagree under counterfactual inputs. We then give a closed-loop statistical bound using predictable regressors, conditionally sub-Gaussian transition noise, and a regularized Schur complement. A scalar feedback example shows the unavoidable scaling: under dithered feedback, residual intervention information grows quadratically with the dither amplitude and the control-channel error decreases with the inverse intervention signal-to-noise scale. New experiments verify these scalings exactly in a linear system and diagnostically in controlled Duffing and Van der Pol
benchmarks. A larger EDMDc acquisition grid further shows that state coverage, joint regression conditioning, and intervention excitation are complementary diagnostics rather than interchangeable performance score.
Read the full abstract ▾
Optimization and Control
Wasserstein Saddle-Free Newton: Saddle Escape and Local Spectral Convergence
Razvan-Andrei Lascu and Taiji Suzuki
Extending saddle-free Newton methods to Wasserstein space requires controlling Hessian dependent operators along nonlinear transport trajectories and accounting for a Hessian spectrum that may accumulate at zero. We address these challenges through Wasserstein Saddle-Free Newton (WSFN), a second-order method for nonconvex optimization over Wasserstein space that preconditions the Wasserstein gradient by a regularized inverse square root Hessian operator. The dynamics exploit
curvature magnitude while maintaining negative
curvature repulsion and positive
curvature attraction, avoiding key Wasserstein Newton limitations. Whereas previous Wasserstein saddle escape results rely on perturbed first-order dynamics, WSFN uses
curvature both to modify the deterministic transport and to construct the perturbation mechanism. Our analysis combines stability estimates for Hessian operators with a perturbative saddle escape argument to establish
polynomial time convergence to approximate second-order stationarity under regularity of landscape assumptions. The contribution of the saddle
curvature parameter $\delta$ improves from $\tilde O(\delta^{-4})$ for prior perturbed first-order Wasserstein methods to $\tilde O(\delta^{-3})$. For benign landscapes, this implies proximity to global minimizers. The cubic dependence corresponds to the scale of the objective decrease that can be guaranteed from negative
curvature under Lipschitz Hessian regularity. Near a minimizer, the possible absence of a uniform spectral gap prevents a uniform contraction result and instead calls for a spectral characterization. We prove
geometric contraction of the linearized dynamics on positive
curvature spectral bands bounded away from zero and
polynomial decay under spectral source conditions. Together, these results connect quantitative saddle escape with the local spectral behavior of second-order Wasserstein dynamics.
Read the full abstract ▾
Combinatorics
Note on the codegree version of the Erdős–Ko–Rado theorem
Luyining Gan, Jie Han and Seonghyuk Im
Kupavskii proved a codegree version of the Erdős–Ko–Rado
theorem by showing that for an intersecting family $\mathcal{F} \subseteq \binom{[n]}{k}$ with $n \geq 2k +3d/(1-d/k)$, the minimum $d$-degree of $\mathcal{F}$ is at most $\binom{n-d-1}{k-d-1}$. Huang and Zhang improved the bound on $n$ to $n \geq 2k+2d-3$. In this short note, we prove that if $d = k-1$, then the bound on $n$ can be improved to $2k + \sqrt{2k} + O(1)$. In addition, we extend our method to show that the bound on $n$ can be improved to $2k + 7k^{2/3}+O(k^{1/3})$ when $d=k-2$.
Read the full abstract ▾
Optimization and Control
Quantitative Fredholm backstepping and rapid stabilization
Ludovick Gagnon, Amaury Hayat, Swann Marx et al. (+2 more)
In this paper, we address the existence of Fredholm backstepping transformations for self-adjoint and skew-adjoint operators $A$. Under suitable assumptions on the operator $A$ and the possibly unbounded control operator $B$, we prove the existence of a Fredholm backstepping transformation for operators of order strictly greater than $1$. This work overcomes two major limitations of the classical Fredholm backstepping framework. One of the main contributions is the explicit identification of the underlying isomorphism used in the construction of the transformation $T$, thereby bypassing the compactness arguments and Riesz basis mechanisms traditionally used in the literature. This explicit structure enables us to derive quantitative and sharp estimates for $\|T\|_{\mathcal{L}(H;H)}$ and $\|T^{-1}\|_{\mathcal{L}(H;H)}$ with respect to the decay rate $\lambda$. As a consequence, we obtain quantitative rapid stabilization and small-time null controllability results for a broad class of operators.
Read the full abstract ▾
All authors Ludovick Gagnon, Amaury Hayat, Swann Marx, Shengquan Xiang, Christophe Zhang
Statistics Theory · Cross-listed from stat.ME
Conditional Predictive Inference for General Structured Data with Group Symmetries
Yichen Shen and Mengxin Yu
We study distribution-free predictive inference for data with group
symmetries, aiming to establish near-conditional coverage guarantees beyond exchangeability for structured data. While many predictive inference methods achieve a target coverage level, most provide marginal coverage. In practice, conditional predictive inference is often preferred, as it quantifies uncertainty for black-box predictions given observed attributes, thereby accommodating heterogeneity. Although many efforts have pursued efficient conditional coverage, existing methods rely on the i.i.d. or exchangeable assumption, often violated in structured settings such as networks, clusters, and imaging data. Recently, SymmPI introduced a unified approach to predictive inference under group
symmetries beyond exchangeability; nevertheless, its guarantees remain marginal and do not account for population heterogeneity. To bridge this gap, we introduce C-SymmPI, a framework that achieves near-conditional coverage under general data structures with group
symmetries, extending beyond exchangeability to cover networks, cluster-level data, and related structures. Inspired by relaxed multi-accuracy, our approach reformulates conditional coverage as miscoverage error over a user-specified function class. We establish theoretical guarantees under distributional invariance and distribution shift, and derive convergence rates for linear and RKHS function classes, recovering state-of-the-art results in the exchangeable setting as special cases. For computational efficiency, we develop two variants: a projection-based
algorithm for high-dimensional observations, and a sampling-based
algorithm for large or infinite groups. We demonstrate effectiveness on hierarchical and network data. Empirical results show that C-SymmPI delivers more informative and stable conditional coverage with improved accuracy compared to existing methods.
Read the full abstract ▾
Algebraic Geometry
Logarithmic Hilbert schemes of curves as weighted blow-ups and their integral Chow rings
Veronica Arena and Terry Dekun Song
The logarithmic
Hilbert scheme of a logarithmic curve parametrizes subschemes on the expanded degenerations of the curve that are transverse to the boundary. We prove that the logarithmic
Hilbert scheme of points on a smooth pointed curve is an iterated weighted blow-up of the symmetric product of the underlying curve. In doing so, we explicitly identify the blow-up centers, weights, and give them modular interpretations. As applications, we calculate their integral Chow rings in terms of those of the symmetric products. Key ingredients in our work include two recent results: the integral Chow ring formula of weighted blow-ups and a weighted analogue of Castelnuovo's criterion for blow-downs. We recover the folklore result the logarithmic
Hilbert scheme of toric $\mathbb{P}^1$ is a toric stack, and the Appendix by Dhruv Ranganathan outlines a complementary approach using Chow quotients.
Read the full abstract ▾
Mathematical Physics · Cross-listed from hep-th
Superform Approach to Equivariant Localization in Supergravity
Michele Galli, Christian Kennedy, Parth Raina et al. (+1 more)
We identify a superspace mechanism behind equivariant localization in supergravity. We show that closed superforms generate, on supersymmetric backgrounds, equivariantly closed polyforms. After presenting the general mechanism, we construct such polyforms for vector and linear multiplets, and for
chiral and BF action principles, in off-shell $4d$ $\mathcal{N}=2$ conformal supergravity, reproducing and extending recent results. Our construction provides a
geometric first step toward equivariant localization of BPS observables in supergravity, including higher-derivative theories, and holography.
Read the full abstract ▾
All authors Michele Galli, Christian Kennedy, Parth Raina, Gabriele Tartaglino-Mazzucchelli
Optimization and Control
A Differentiable Interior-Point Method in Single Precision
Jon Arrizabalaga, Kevin Tracy and Zachary Manchester
Primal-dual interior-point methods solve constrained convex optimization problems to tight tolerances with speed and robustness. Their solutions are also efficiently differentiable with respect to the problem data through the implicit function
theorem. However, the standard treatment of primal-dual complementarity makes the underlying linear systems increasingly ill-conditioned near the solution. While this ill-conditioning is often benign in double precision, it can be catastrophic in single precision, preventing interior-point methods from fully exploiting the accelerated hardware that underpins modern machine learning. This paper introduces a differentiable interior-point method designed for low-precision arithmetic. By using an alternative complementarity representation, we ensure that the underlying linear systems remain spectrally bounded – even near the solution – a property that is essential for computing accurate gradients and avoiding arithmetic exceptions. As a result, our method enables interior-point solvers to reliably solve and differentiate optimization problems in single precision that were previously confined to double precision. We demonstrate the approach through an ablation study against the standard interior-point formulation and applications in bilevel and end-to-end learning settings where differentiating through constrained optimization is essential. The source code is available at https://github.com/qpax-solver/qpax.
Read the full abstract ▾
Representation Theory
Quasi-admissible, raisable nilpotent orbits and covering Barbasch-Vogan duality
Yi-Yang Zhang
For simply-connected Lie groups of type E over \( p \)-adic local field \( F \), we determine the degree of the cover required for a given \( F \)-split nilpotent orbit to be quasi-admissible or raisable, respectively. Combining this result with the previously computed data for other types by Gao-Liu-Tsai, we prove that all \( F \)-split nilpotent orbits whose geometry type contained in the image of the covering Barbasch-Vogan duality map \( d_{\mathrm{BV},G}^{(n)} \) of almost-simple Lie groups \( G \) in each Cartan type are always \( \overline{G}^{(n)} \)-quasi-admissible.
Read the full abstract ▾
Optimization and Control · Cross-listed from cs.PL
Optimizing Optimizations, Declaratively: Optimizing the Higher-Order Functions in Mathematical Optimization with egglog
Hiromi Ishii
We present two applications of egglog to mathematical optimization in JijModeling 2, a mathematical modeller whose internal representation is based on simply typed $\lambda$-calculus. First, we use egglog to improve $\LaTeX$ output for mathematical models expressed with higher-order functions. Python comprehensions are desugared into stream operations such as $\textsf{map}$, $\textsf{flat_map}$, and $\textsf{filter}$; emitting these terms directly produces unnatural mathematical notation. We reconstruct comprehension syntax by ensugaring higher-order terms and use equality saturation with a custom cost model to minimize temporary variable rebindings. Second, we use egglog as a declarative engine for constraint detection, extending the previous egg-based approach presented at EGRAPHS '25. Egglog's datalog-style rules let us express multi-step detection logic directly, without external Rust orchestration code. We encode parametrized constraints using Henkin-like constants and propagate side conditions on subterms and indices through egglog facts. Finally, we show that the same ensugaring procedure also reduces large domain-set conditions before saturation, turning a problematic detection case from minutes or nontermination into a few seconds. Through these topics, we want to provide an example of an industrial application of egglog, demonstrate the trick to propagate the constraints using the ideas from mathematical logic, and show the importance of optimizing premises of egglog rules to get practical performance in egglog programs.
Read the full abstract ▾
Optimization and Control
On the convergence of doubly stochastic Primal-Dual Hybrid Gradient Method
Yiheng Xiao and Huikang Liu
We study a block-structured class of convex-concave saddle-point problems in which both the primal and dual variables admit natural separable decompositions. Motivated by large-scale applications where a full update on either side can be computationally expensive, we propose a doubly
stochastic primal–dual hybrid gradient method (DSPDHG) that performs randomized block updates on both primal and dual variables.The method extends classical PDHG and
stochastic PDHG (SPDHG) schemes in a unified manner:it reduces to deterministic PDHG when all blocks are selected and to one-sided
stochastic variants when only one side is randomized. For the general convex setting, we establish an $\mathcal{O}(1/K)$ ergodic convergence rate for the expected restricted primal–dual gap under suitable blockwise step-size conditions. We further analyze a restarted variant of DSPDHG under a quadratic growth condition in terms of the smoothed primal-dual gap. Under this regularity assumption, we prove linear convergence of the restarted outer iterates. Numerical evidence is provided to show that restarted DSPDHG with standard step sizes demonstrates competitive practical performance compared with PDHG, SPDHG, and their restarted variants.
Read the full abstract ▾
Combinatorics
Thrall's problem for two rows
JiSun Huh, Woo-Seok Jung, Jang Soo Kim et al. (+1 more)
In this paper, we study Thrall's problem for the higher Lie modules $L_\lambda$. Our main result provides a tableau-theoretic description of the Schur expansion of the character of $L_\lambda$ when $\lambda$ has two rows, thereby solving Thrall's problem in this case. This formula is expressed in terms of standard Young tableaux with major index congruence conditions and a
spin-parity condition defined through bijections with Yamanouchi domino tableaux. We also obtain tableau formulas for hook shapes and partitions with distinct parts, and these results extend to all partitions in which each part greater than $2$ occurs at most twice.
Read the full abstract ▾
All authors JiSun Huh, Woo-Seok Jung, Jang Soo Kim, Meesue Yoo
Differential Geometry
Minimal Lagrangian surfaces in the two-dimensional complex hyperbolic quadric via the loop group method
Shimpei Kobayashi and Sihao Zeng
We study minimal Lagrangian surfaces in the complex
hyperbolic quadric. We show that minimality of a Lagrangian surface is characterized by a loop of flat connections, which yields an associated $\mathbb S^1$-family of isometric deformations. We also establish a correspondence with spacelike maximal surfaces in anti-de Sitter $3$-space via the Gauss map. Using the resulting harmonic map into the
hyperbolic two-space, we develop a DPW-type representation and construct explicit examples, including $\mathbb{R}$-equivariant and radially symmetric surfaces. In particular, under suitable conditions, the $\mathbb{R}$-equivariant family contains catenoid-type examples.
Read the full abstract ▾
Geometric Topology
Realizing crosscap transpositions as monodromies of singular fibrations
Kenta Hayano
We introduce a new type of singularity for smooth maps from $4$-
manifolds to surfaces, called an $M$-singularity, whose critical locus is a circle contained in a single fiber. We show that the monodromy around an $M$-singularity is a crosscap transposition in the mapping class group of a non-orientable surface. We also introduce $M$-fibrations, namely smooth maps whose singularities consist only of $M$-singularities, and prove that relations among crosscap transpositions give rise to such fibrations on non-orientable $4$-
manifolds. We then study handle decompositions associated with $M$-fibrations and their orientation double coverings. In particular, we describe the attaching circles and framings of the two $2$-handles arising from the orientation double cover of an $M$-singularity. Using this description, we construct a closed non-orientable $4$-
manifold which admits an $M$-fibration but admits no Lefschetz fibration. We further discuss singularity-theoretic properties of the local model of an $M$-singularity, namely its infinite $\mathcal{A}_e$-codimension and an explicit stable perturbation.
Read the full abstract ▾
Number Theory
A note on the $q$-adic valuation of $\sigma_k(n)$
Olivier Bordellès
In this note, we obtain an exact formula for the $q$-adic valuation of $\sigma_k(n)$ where $q$ is an odd
prime, allowing us to derive an explicit upper bound which is asymptotically better than the previous bound obtained by Zhao when $n$ is large and $k \geqslant q-2$. The key parts are played by the LTE lemma and the use of cyclotomic
polynomials.
Read the full abstract ▾
Functional Analysis
Unbounded Toeplitz operators and finite rank de Branges-Rovnyak spaces
Soumitra Ghara, MD Ramiz Reza and Chaman Kumar Sahu
Motivated by the recent developments of de Branges-Rovnyak spaces, we investigate the function theoretic aspects of finite rank de Branges-Rovnyak spaces $H(B)$ generated by row-valued Schur functions $B$. We provide a generalization of Sarason's fundamental work by characterizing finite rank $H(B)$-spaces as the domain of the adjoint of the Toeplitz operators $T_\varphi^*$ with symbol $\varphi = BA^{-1}$, where $A$ is an
matrix-valued outer function satisfying $A^*A+B^*B = I$ a.e. on the unit circle. We derive a norm formula for functions in $H(B)$-space and provide a concrete realization of this norm in terms of the Taylor coefficients of the function and the symbol $\varphi$. As an application, we characterize all symbols $B$ for which $H^\infty \subseteq H(B)$ in terms of the boundary behavior of $I-BB^*$, thereby extending Sarason's criterion for the classical de Branges-Rovnyak spaces.
Read the full abstract ▾
Information Theory
Transformer-Based Hybrid Beamforming with Reconfigurable Pixel Antenna for HAPS Communications
Ruiqi Wang, Ziwei Wan, Keke Ying et al. (+1 more)
This paper proposes a
Transformer-based hybrid beamforming framework for reconfigurable pixel
antenna (RPA)-equipped massive multiple-input multiple-output (
MIMO) in high-altitude platform station (HAPS) communications. The proposed pattern reconfigurable hybrid beamforming network (PR-HBFNet) comprises two key components: 1) a pattern reconfigurable network that leverages a
Transformer encoder to determine the radiation pattern for each
antenna element, and 2) a hybrid beamforming network that employs model-driven residual learning to compute analog and digital precoders over SVD-based initializations. Simulation results demonstrate that the proposed PR-HBFNet closely approaches the spectral efficiency of a greedy
benchmark while significantly reducing computational complexity.
Read the full abstract ▾
All authors Ruiqi Wang, Ziwei Wan, Keke Ying, Zhen Gao
Numerical Analysis · Cross-listed from stat.ML
Simple Approximation and Derivative Free Inference-Time Scaling for Diffusion Models via Sequential Monte Carlo on Path Measures
Chenyang Wang, Weizhong Wang, Yinuo Ren et al. (+2 more)
iffusion-based generative models increasingly rely on inference-time guidance, adding a drift term or reweighting mixture of experts, to improve sample quality on task-specific objectives. However, most existing techniques require repeated score or gradient evaluations, introducing bias, high computational overhead, or both. We introduce URGE, Unbiased Resampling via Girsanov Estimation, a derivative-free inference-time scaling
algorithm that performs path-wise importance reweighting via a Girsanov change of measure. Instead of computing gradient-based particle weights in previous work, URGE attaches a simple multiplicative weight to each simulated trajectory and periodically resamples. No score, no Hessian, and no PDE evaluation is required. We establish an equivalence between path-wise and particle-wise SMC: the Girsanov path weight admits a backward conditional expectation that recovers the previous particle-level weights, guaranteeing that both schemes produce the same unbiased terminal law. Empirically, URGE outperforms existing inference-time guidance baselines on synthetic tests and
diffusion-model
benchmarks, achieving better generation quality, while being significantly simpler to implement and fully gradient-free.
Read the full abstract ▾
All authors Chenyang Wang, Weizhong Wang, Yinuo Ren, Jose Blanchet, Yiping Lu