Mathematical Physics · Cross-listed from hep-th
On a Deformed Holomorphic Chern-Simons Theory
Eirik Høgmoe Kjelsnes, Eirik Eik Svanes and Vegard Undheim
We deform classical holomorphic Chern–Simons theory on a Calabi–Yau three-fold $X$ by deforming the complex structure by a deformation parameter $h \in\mathscr{H}^{0,1}(T^{1,0}X)$. The corresponding equations of motion admit new "instanton solutions" which which are invariant under re-scalings of $h$, and are perhaps more reminiscent of $G_2$-instantons for $G_2$
manifolds. We give examples of such instantons. In particular, when $h$ has non-vanishing Yukawa coupling ${\rm Yuk}(h,h,h)\neq 0$, it may be used to define a connection on ${\rm End}(T^{1,0}X)$ solving the instanton constraint. Interestingly, this connection gives rise to a hermitian (self-adjoint) connection for a real gauge theory on the real bundle ${\rm End}(TX)$ for only specific directions in deformation space, which may be classified using Morse theory. We quantize the deformed theory around these instanton backgrounds, and derive explicit expressions for the partition function in the limit where the complex structure deformation is large. We study anomalies, and the $h$-dependece of the partition function. In particular, coupling the theory to additional gravitational degrees of freedom, we find that the special directions in deformation space give rise to novel anomaly free theories on ${\rm End}(T^{1,0}X)$.
Read the full abstract ▾
Optimization and Control
On the Connectedness of Sublevel Sets in Invex Optimization
Vinzenz Thoma, Zebang Shen and Niao He
Understanding the
topology of sublevel sets yields crucial insights into the optimization landscape of non-convex functions. If sublevel sets are connected, local search
algorithms are less likely to be trapped in isolated valleys, facilitating convergence to global minimizers. However, few results exist to establish connectedness in the nonconvex setting. In this work, we present a mathematical toolkit based on the topological mountain pass
theorem and use it to study invex functions, a class of functions that includes those satisfying the Polyak-Łojasiewicz inequality and generalizations thereof. We show that their sublevel sets are connected under mild assumptions. We further leverage our result to establish the connectedness of different solution sets for invex-incave minimax problems and incave games.
Read the full abstract ▾
Functional Analysis
Karhunen Loève Expansions of Hilbert Space-Valued Random Elements
Trajan Murphy
The Karhunen-Loève Expansion (KLE) of a
stochastic process is a well understood eigenfunction expansion used widely in time series analysis,
stochastic PDEs, and signal processing. Karhunen-Loève expansions have also been proven to exist for other types of
stochastic elements whose values lie in certain $L^2$ spaces. This article provides a concise proof about the necessary and sufficient conditions for a function $v$ defined on some sample space $\Omega$ and whose values lie in some
Hilbert space $\mathcal H$ to admit an eigenfunction expansion like the well-known KLE. We draw on the existing theory of Bochner spaces and
Hilbert-Schmidt spaces and construct an isomorphism between them. Furthermore, this isomorphism is natural, which has important computational consequences. Finally, we demonstrate with an example the computational advantages conferred by considering the KLE in this generalized setting.
Read the full abstract ▾
Analysis of PDEs · Cross-listed from stat.ML
On the continuum limit of t-SNE for data visualization
Jeff Calder, Zhonggan Huang, Ryan Murray et al. (+1 more)
This work is concerned with the continuum limit of a graph-based data visualization technique called the t-Distributed
Stochastic Neighbor
Embedding (t-SNE), which is widely used for visualizing data in a variety of applications, but is still poorly understood from a theoretical standpoint. The t-SNE
algorithm produces visualizations by minimizing the Kullback-Leibler divergence between similarity
matrices representing the high dimensional data and its low dimensional representation. We prove that as the number of data points $n \to \infty$, after a natural rescaling and in applicable parameter regimes, the Kullback-Leibler divergence is consistent as the number of data points $n \to \infty$ and the similarity graph remains sparse with a continuum variational problem that involves a non-convex gradient regularization term and a penalty on the magnitude of the probability density function in the visualization space. These two terms represent the continuum limits of the attraction and repulsion forces in the t-SNE
algorithm. Due to the lack of convexity in the continuum variational problem, the question of well-posedeness is only partially resolved. We show that when both dimensions are $1$, the problem admits a unique smooth minimizer, along with an infinite number of discontinuous minimizers (interpreted in a relaxed sense). This aligns well with the empirically observed ability of t-SNE to separate data in seemingly arbitrary ways in the visualization. The energy is also very closely related to the famously ill-posed Perona-Malik equation, which is used for denoising and simplifying images. We present numerical results validating the continuum limit, provide some preliminary results about the delicate nature of the limiting energetic problem in higher dimensions, and highlight several problems for future work.
Read the full abstract ▾
All authors Jeff Calder, Zhonggan Huang, Ryan Murray, Adam Pickarski
Combinatorics
Upper bounds for double Roman domination and $[k]$-Roman domination of cylindrical graphs $C_m \Box P_n$
Simon Brezovnik and Janez Žerovnik
Roman-type domination parameters form an important class of graph invariants that model protection and resource allocation problems on networks. Among them, $[k]$-Roman domination provides a unified framework that generalizes Roman, double Roman, and higher-order variants. In this paper we investigate the $[k]$-Roman domination number of cylindrical grids $C_m\Box P_n$ and derive several new constructive upper bounds. Our approach combines three complementary techniques: linear periodic constructions, uniform ceiling-type labelings, and packing-based refinements. We first analyze the case $C_9\Box P_n$, where these three families of bounds can be compared explicitly and their relative efficiency is shown to depend on the parameter $k$. We then extend the linear constructions to cylindrical grids whose circumference is a multiple of one of the values $3,\dots,9$, obtaining a unified family of upper bounds for $C_{rt}\Box P_n$. Motivated by the asymptotic behavior of these estimates, we further derive general upper bounds depending only on the residue class of $m$ modulo $5$, which apply to all cylindrical grids. As a consequence, we obtain explicit estimates for the double Roman domination number $\gamma_{[2]R}(C_m\Box P_n)$ and compare the resulting multiple-based constructions with the residue-class bounds. This comparison shows that the residue-class construction becomes asymptotically superior for all sufficiently large admissible circumferences, while several exceptional small cases remain better covered by tailored constructions.
Read the full abstract ▾
Geometric Topology · Cross-listed from cs.GR
Twisted Edges: A Unified Framework for Designing Linked Knot (LK) Structures Using Labeled Non-Manifold Surface Meshes
Tolga Talha Y{\i}ld{\i}z, Uğur Önal, Vinayak R. Krishnamurthy et al. (+1 more)
We present Twisted Edges, a unified framework for designing Linked
Knot (LK) structures using labeled non-
manifold surface meshes. While the concept of edge twists, originating in topological graph theory, is foundational to these designs, prior approaches have been strictly limited to binary states. We identify this restriction as a critical barrier; binary twisting fails to capture the full spectrum of topological possibilities, rendering a vast class of structural and dynamic behaviors inaccessible. To overcome this limitation, we generalize the twist formulation to support arbitrary integer twist labels. This expansion reveals that while zero twists may introduce disconnections, applying even twists to 2-
manifold meshes robustly preserves connectivity, transforming surfaces into fully connected, chainmail-like structures where faces form consistently linked cycles. Furthermore, we extend this framework to non-
manifold meshes, where specific integer assignments prevent cycle merging. This capability, unattainable with binary methods, enables the design of partial connectivity and functional hinges, supporting dynamic folding and articulation. Theoretically, we show that these integer-twisted meshes correspond to knotted surfaces in four dimensions, with LK structures arising as their immersions into $\mathbb{R}^3$. By breaking the binary constraint, this work establishes a coherent paradigm for the systematic exploration of previously unstudied woven and articulated structures.
Read the full abstract ▾
All authors Tolga Talha Y{\i}ld{\i}z, Uğur Önal, Vinayak R. Krishnamurthy, Ergun Akleman
Numerical Analysis · Cross-listed from physics.optics
And Yet Another FEM-Based Mode Solver for Dielectric Waveguides
Ergun Simsek
We present a full-vector finite element method (FEM) mode solver for dielectric waveguides based on a mixed Nedelec-Lagrange discretization of Maxwell's curl equations in the frequency domain. The formulation combines edge elements for transverse field components with nodal elements for the longitudinal component, enabling accurate modeling of hybrid modes while effectively suppressing spurious solutions. The solver is implemented in both MATLAB and Python with an emphasis on reproducibility, computational efficiency, and accessibility, including compatibility with cloud-based platforms. Numerical validation is performed on representative waveguide structures, demonstrating excellent agreement with COMSOL Multiphysics, with relative errors below 0.05%. Convergence studies confirm the expected accuracy trends with mesh refinement, while highlighting the trade-off between computational cost and precision. The proposed implementation provides a flexible and reliable open-source tool for integrated photonics research and education.
Read the full abstract ▾
Complex Variables
The Schwarzian Derivative for Convex Holomorphic Mappings in Several Complex Variables
Rodrigo Hernández
We obtain upper bounds for the norm of the Schwarzian derivative of convex holomorphic mappings defined on the polydisk and the unit ball in $\mathbb{C}^n$. For coordinate-wise convex mappings on the polydisk, we derive a sharp estimate extending the classical one-variable result of Chuaqui–Duren–Osgood to higher dimensions. For the Roper–Suffridge extension operator in the unit ball, we obtain an explicit bound that represents the best available estimate in this setting.
Read the full abstract ▾
Optimization and Control · Cross-listed from cs.LG
Offline-Online Reinforcement Learning for Linear Mixture MDPs
Zhongjun Zhang and Sean R. Sinclair
We study offline-online reinforcement learning in linear mixture Markov decision processes (MDPs) under environment shift. In the offline phase, data are collected by an unknown behavior policy and may come from a mismatched environment, while in the online phase the learner interacts with the target environment. We propose an
algorithm that adaptively leverages offline data. When the offline data are informative, either due to sufficient coverage or small environment shift, the
algorithm provably improves over purely online learning. When the offline data are uninformative, it safely ignores them and matches the online-only performance. We establish regret upper bounds that explicitly characterize when offline data are beneficial, together with nearly matching lower bounds. Numerical experiments further corroborate our theoretical findings.
Read the full abstract ▾
Mathematical Physics · Cross-listed from gr-qc
Asymptotic Theorems and Averaging in Scalar Field Cosmology
Genly Leon, Aleksander Kozak and Claudio Michea
We present a hybrid study that combines a concise review of scalar-field cosmology with new analytic developments that integrate averaging reductions for oscillatory regimes with dynamical-systems techniques. For oscillatory fields, we derive an averaging reduction that yields an effective slow system whose time averages control dissipation; introducing uniform derivative bounds, Barbalat/LaSalle arguments, and a finite-dimensional center/stable
manifold reduction, we carry out late-time analysis of the models. We prove persistence of
equilibria, decay estimates, and local invariant
manifolds under small $C^k$ perturbations of $\chi(\phi)$ and $G(a)$, quantify how averaged dissipation lifts to the full oscillatory dynamics with an $\mathcal{O}(H)$ error, and provide numerical examples. In addition to asymptotic reductions, we obtain exact quadrature solutions in general relativistic, anisotropic, and brane-world settings, yielding closed-form expressions for $t(a)$, $\phi(a)$, and $H(a)$ and enabling analytic computation of inflationary observables.
Read the full abstract ▾
Dynamical Systems
Quadratic differentials and random walks on the dual graph of a pants decomposition
Charles Bordenave, Xinlong Dong and Dragomir Šarić
Let X be an infinite Riemann surface with an upper-bounded geodesic pants decomposition. The vertices of the corresponding dual graph G are pairs of pants and edges are cuffs with conductances equal to their lengths. We prove that the geodesic flow on X is ergodic if and only if the random walk on G is recurrent. This yields explicit criteria for deciding, in terms of cuff-length growth, whether the geodesic flow is ergodic. We provide concrete and new families of Riemann surfaces with an explicit understanding of the phase transitions from recurrent to non-recurrent geodesic flows. In addition, we show that rough isometry of surfaces does not preserve the ergodicity of the geodesic flow while rough isometry of their dual graphs does. The above equivalence result uses a characterization of the measured geodesic laminations on X that arise as straightened horizontal foliations of finite-area holomorphic quadratic differentials. The conditions on the measured laminations are translated into the conditions on the existence of a square summable flow function on G.
Read the full abstract ▾
Dynamical Systems
Mean dimension of general iterated function systems
Welington Cordeiro, Maria José Pacifico and Xuan Zhang
In this paper, we introduce and investigate the notions of Mean Dimension and Metric Mean Dimension for generalized iterated function systems (IFS). We establish basic properties of these invariants and prove that Mean Dimension is always bounded above by the lower Metric Mean Dimension and the upper Metric Mean Dimension in this setting. We further show that generalized iterated function systems with the Small Boundary Property have zero Mean Dimension. Finally, we introduce a Gluing Orbit Property for generalized iterated function systems and prove that, under suitable transitivity and non-
rigidity assumptions, it guarantees positive topological
entropy.
Read the full abstract ▾
Mathematical Physics
Effective Dynamics for the Bose Polaron in the Large-Volume Mean-Field Limit
Jonas Lampart, Peter Pickl and Siegfried Spruck
We consider the dynamics of the Bose polaron system, a dense
quantum gas consisting of $N$
bosons evolving in $\mathbb{R}^3$ in the presence of an
impurity particle. The system is studied in the mean-field scaling with initially high density $\rho$ and large volume $\Lambda$ of the gas. In the initial state, almost all
bosons are in the Bose-Einstein condensate, with a few excitations. We derive from the microscopic dynamics, in the joint limit of large densities and volumes, with the constraint $\Lambda^3 \ll \rho$, the effective description by the translation-invariant Bogoliubov-Fröhlich
Hamiltonian, which couples the
quantum field of excitations linearly to the
impurity particle.
Read the full abstract ▾
Analysis of PDEs
Sharp Makai-type inequalities for the best Poincaré-Sobolev constants
Giovanni Pisante and Francesca Prinari
Given a bounded convex open set $\Omega\subseteq \mathbb R^N$, we prove that the Poincaré-Sobolev constants $\lambda_{p,q}(\Omega)$ can be bounded from below by the $p$-power of the ratio between the perimeter of $\Omega$ and a suitable power of its volume, with an optimal constant which is explicitly given. This generalises an old result for torsional
rigidity due to Makai when $N=2$. The proof relies on new
geometric optimal bounds for the Lebesgue norms of the distance function from the boundary which are of independent interest. These results allow us to give a complete picture of the sharp inequalities for $\lambda_{p,q}(\Omega)$ in terms of suitable powers of perimeter, inradius and volume of $\Omega$.
Read the full abstract ▾
Algebraic Geometry
Affine Springer fiber and the small quantum group
Roman Bezrukavnikov, Pablo Boixeda Alvarez, Michael McBreen et al. (+1 more)
We find a new
geometric incarnation for the principal block in the category of modules over a
quantum group at a root of unity, realizing it as a full subcategory of microsheaves on a certain affine Springer fiber. We also prove a related
geometric Langlands type equivalence with wild ramification, identifying the latter category with a category of coherent sheaves on the Springer resolution for the dual group. This can also be viewed as a version of homological mirror
symmetry for the Springer resolution.
Read the full abstract ▾
All authors Roman Bezrukavnikov, Pablo Boixeda Alvarez, Michael McBreen, Zhiwei Yun
Analysis of PDEs
On the heat equation with singular drift
N.V. Krylov
We prove the maximum modulus estimates in terms of the $L_{q,p}$-norm of the free term for solutions of the heat equation with Morrey drift for any $q,p$ satisfying $d/p+2/q<2$ and any order of integration in the definition of the $L_{q,p}$-norm. An application to the case of $b$ satisfying the Ladyzhenskaya-Prodi-Serrin condition is given. The technique is easily adaptable to equations with
Laplacians of order $\geq 1$.
Read the full abstract ▾
Algebraic Geometry
Cycles in the universal moduli stack of bundles of rank two over genus two curves
Shubham Saha
We present a
conjecture for the Chow ring of the universal
moduli stack of bundles over hyperelliptic curves and prove it for rank and genus two. Consequently, we obtain explicit generators and relations to conclude that the Chow ring is tautological. In addition, we compute the Chow rings of products of universal Jacobians over genus two curves.
Read the full abstract ▾
Combinatorics
Base Modulus for Matroid Truncation, Strength, and Fractional Arboricity
Huy Truong and Pietro Poggi-Corradini
In [27], we provided results on the $p$-modulus of the family of all bases of matroids and showed that it recovers various concepts in matroid theory, including strength, fractional arboricity, and principal partitions. In particular, the unique optimal density $\eta^*$ that arises for $p$-modulus, which we will refer to as universal density from now on, was shown to recover the concept of lexicographical base in polymatroids. Since truncation is a fundamental operation in matroid theory, it is natural to ask how the universal density behaves under matroid truncation. In this paper, we first provide the universal density of every truncation of a given matroid; equivalently, we determine the principal partition for every matroid truncation. Next, we give a new characterization of the universal density using the Kullback–Leibler divergence. Furthermore, we study the notion of strictly homogeneous matroids, generalizing the corresponding notion in graphs from [6]. We also offer several insights related to strength, fractional arboricity, and give the set of probability mass functions (pmfs) for bases that induce the universal density in a simple case. Finally, this paper also addresses two optimization problems for graph structures, particularly those involving edge-disjoint spanning trees and forest edge-coverings.
Read the full abstract ▾
Number Theory
Simultaneous non-vanishing of Dirichlet L-functions
Hung M. Bui, Alexandra Florea and Micah B. Milinovich
In this paper, we prove the simultaneous non-vanishing of four Dirichlet $L$-functions at any point on the critical line. More precisely, let $\chi_1,\ldots,\chi_4$ be even Dirichlet characters modulo $D_1,\ldots, D_4$ respectively, where the $D_j$ are pairwise co-
prime and square-free integers. Under the Generalized Riemann Hypothesis, we prove that $\prod_{j=1}^4 L(1/2+it,\chi \chi_j) \neq 0$ for a positive proportion of Dirichlet characters $\chi \pmod q$, with $q$
prime and sufficiently large in terms of the $D_j$ and $t$ (and with an explicit relationship between $D_j, t$ and $q$). Unconditionally, we also prove a simultaneous non-vanishing result for four Dirichlet $L$-functions for infinitely many characters $\chi \pmod q$, though in this case the proportion tends to zero as $q \to \infty$.
Read the full abstract ▾
Combinatorics
Degree sequences realizing labelled $h$-factors
Zhen Liu and Qinghou Zeng
For a positive integer \( k \), let \( [k] = \{1, 2, \ldots, k\} \). Let \( h \) be a non-negative integer, and let \( n \) be a multiple of \( h + 1 \). Define \( H \) as the disjoint union of \( n/(h+1) \) cliques (each of size \( h + 1 \)) with vertex sets \( V_1, \ldots, V_{n/(h+1)} \), where \( V_i = \{ v_j \mid j = (i-1)(h+1) + k, k \in [h+1] \} \) for \( i \in [n/(h+1)] \). A non-increasing integer sequence \( (d_1, \ldots, d_n) \) is \( H \)-realizable if there exists a graph \( G \) with \( V(G) = V(H) = \{ v_i \mid i \in [n] \} \), \( d_G(v_i) = d_i \) for all \( i\in [n] \), and \( G \) contains \( H \) as a spanning subgraph. If \( h = 0 \), then a non-increasing integer sequence \( (d_1, \ldots, d_n) \) is \( H \)-realizable if and only if there exists a graph \( G \) with degree sequence \( (d_1, d_2, \dots, d_n) \); Erdős and Gallai established a necessary and sufficient condition for this property. Recently, Briggs, McDonald, and Shan extended their result to the case \( h = 1 \). In this paper, we establish a necessary and sufficient condition for a sequence \( (d_1, d_2, \dots, d_n) \) to be \( H \)-realizable for any non-negative integer \( h \), thereby confirming a
conjecture due to Briggs, McDonald and Shan.
Read the full abstract ▾
Probability · Cross-listed from cs.DS
Sampling Colorings Close to the Maximum Degree: Non-Markovian Coupling and Local Uniformity
Vishesh Jain, Clayton Mizgerd and Eric Vigoda
Sampling graph colorings via local Markov chains is a central problem in approximate counting and Markov chain Monte Carlo (MCMC). We address the problem of sampling a random $k$-coloring of a graph with maximum degree $\Delta$. The simplest algorithmic approach is to establish rapid mixing of the single-site update chain known as the Metropolis Glauber dynamics, which at each step chooses a random vertex $v$ and proposes a random color $c$, recoloring $v$ to $c$ if the resulting coloring remains proper. It is a long-standing open problem to prove that the Glauber dynamics has
polynomial mixing time on all graphs whenever $k\geq\Delta+2$. We prove that for every $\delta>0$ and all $\Delta \geq \Delta_0(\delta)$, if $k\ge (1+\delta)\Delta$ then the Glauber dynamics has optimal mixing time of $O_{\delta}(|V| \log |V|)$ on any graph of girth $\geq 7$ and maximum degree $\Delta$. Our approach builds on a non-Markovian coupling introduced by Hayes and Vigoda (2003) for the large-degree regime $\Delta=\Omega(\log n)$ and girth $11$, in which updates at time $t$ may depend on and modify proposed updates at future times. A complete analysis of this framework requires resolving substantial technical obstacles that remain in the original argument, and extending it to the constant-degree regime introduces further difficulties, since non-Markovian updates may fail with constant probability. We overcome these obstacles by developing and analyzing a refined local non-Markovian coupling, and by establishing new local-uniformity results for the Metropolis dynamics, extending prior results for the heat-bath chain due to Hayes (2013). Together, these ingredients provide a complete analysis of the non-Markovian coupling framework in the large-degree regime, while simultaneously strengthening it substantially to obtain optimal mixing all the way down to the constant-degree setting.
Read the full abstract ▾
Combinatorics
On the Ramsey numbers of wheels, cycles, and stars
Louis DeBiasio and Tucker Wimbish
The wheel $W_{k}$ is the graph on $k+1$ vertices consisting of a vertex joined to a cycle of length $k$, and we say that $W_k$ is an even wheel if $k$ is even. Mao, Wang, Magnant, Schiermeyer proved that the Ramsey number of $W_{2n}$ is between $4n+1$ and $12n-2$. We improve both of these bounds, showing that $5n-\frac{1+(-1)^{n}}{2}\leq R(W_{2n})\leq 8n+664$ for all integers $n\geq 2$. The main focus of the paper concerns two general results on the Ramsey numbers of
stars versus even wheels and even cycles versus even wheels, from which the above bounds are obtained as a corollary. That is, we asymptotically determine $R(K_{1,m}, W_{2n})$ and $R(C_{2m}, W_{2n})$ for all sufficiently large $m$ and $n$, both of which were open problems for most regimes. As for odd wheels, we note that the analogous values for
stars versus odd wheels and odd cycles versus odd wheels were already known exactly, from which it follows that $6n+4=R(K_{1,2n+1}, W_{2n+1})\leq R(W_{2n+1})\leq 2\cdot R(C_{2n+1}, W_{2n+1})=12n+2$. Very recently, Zhang and Chen improved the upper bound to $R(W_{2n+1})\leq \frac{32n}{3}+O(1)$. We are able to refine their proof, further improving the upper bound to $R(W_{2n+1})\leq 10n+O(1)$.
Read the full abstract ▾
Information Theory · Cross-listed from eess.SY
Quantized Online LQR
Barron Han, Victoria Kostina and Babak Hassibi
We study online linear-quadratic regulation (LQR) with unknown dynamics under communication rate constraints. Classical networked control quantizes the plant state at every time step, requiring $O(T)$ total bits while injecting persistent quantization noise that limits control performance. We consider a setting where the plant observes its state locally and can estimate system dynamics via ordinary least squares, while a remote controller possesses knowledge of the control cost. Rather than quantizing the raw state, the plant transmits learned dynamics estimates over a rate-limited uplink, and the controller returns the optimal control policy so that the plant can compute actions locally using its superior state knowledge. We first prove a fundamental information-theoretic lower bound: any scheme achieving $O(T^\alpha)$ regret for $\alpha \in [1/2,1)$ compared to the optimal infinite horizon LQR controller that knows the true system dynamics must transmit at least $\Omega(\log T)$ bits. We then design the Quantized Certainty Equivalent (QCE-LQR)
algorithm, which matches this bound. The resulting regret bound contains inflation factors $Q_{\mathrm{slow}}(\varrho)$ and $Q_{\mathrm{fast}}(\varrho)$ that vanish as the codebook resolution increases, smoothly recovering the unquantized baseline regret. Numerical experiments on four
benchmark systems – from a scalar unstable plant to a 24-parameter Boeing 747 lateral model – confirm that a variant of QCE-LQR achieves regret comparable to an unquantized certainty equivalent controller over a horizon of $T=10{,}000$ steps.
Read the full abstract ▾
Dynamical Systems · Cross-listed from cs.LG
Fast and principled equation discovery from chaos to climate
Yuzheng Zhang, Weizhen Li and Rui Carvalho
Our ability to predict, control, and ultimately understand complex systems rests on discovering the equations that govern their dynamics. Identifying these equations directly from noisy, limited observations has therefore become a central challenge in data-driven science, yet existing library-based sparse regression methods force a compromise between automation, statistical rigor, and computational efficiency. Here we develop
Bayesian-ARGOS, a hybrid framework that reconciles these demands by combining rapid frequentist screening with focused
Bayesian inference, enabling automated equation discovery with principled uncertainty quantification at a fraction of the computational cost of existing methods. Tested on seven chaotic systems under varying data scarcity and noise levels,
Bayesian-ARGOS outperforms two state-of-the-art methods in most scenarios. It surpasses SINDy in data efficiency for all systems and noise tolerance for six out of the seven, with a two-order-of-magnitude reduction in computational cost compared to bootstrap-based ARGOS. The probabilistic formulation additionally enables a suite of standard statistical diagnostics, including influence analysis and multicollinearity detection that expose failure modes otherwise opaque. When integrated with representation learning (SINDy-SHRED) for high dimensional sea surface temperature reconstruction,
Bayesian-ARGOS increases the yield of valid latent equations with significantly improved long horizon stability.
Bayesian-ARGOS thus provides a principled, automated, and computationally efficient route from scarce and noisy observations to interpretable governing equations, offering a practical framework for equation discovery across scales, from
benchmark chaotic systems to the latent dynamics underlying global
climate patterns.
Read the full abstract ▾
Probability
Spectral Structure in Finite Free Information Inequalities and $p$-Stam Phase Transitions
Baran Hashemi
Using FlowBoost, a closed-loop deep generative optimization framework for extremal structure discovery, we investigate $\ell^p$-generalizations of the finite free Stam inequality for real-rooted
polynomials under finite free additive convolution $\boxplus_n$. At $p=2$, FlowBoost finds the Hermite pair as the unique equality case and reveals the spectral structure of the linearized convolution map at this extremal point. As a result, we
conjecture that the singular values of the doubly
stochastic coupling
matrix $E_n$ on the mean-zero subspace are ${2^{-k/2}:k=1,\ldots,n-1}$, independent of $n$. Conditional on this
conjecture, we obtain a sharp local stability constant and the finite free CLT convergence rate, both uniform in $n$. We introduce a one-parameter family of $p$-Stam inequalities using $\ell^p$-Fisher information and prove that the Hermite pair itself violates the inequality for every $p>2$, with the sign of the deficit governed by the $\ell^p$-contraction ratio of $E_n$. Systematic computation via FlowBoost supports the
conjecture that $p^*\!=2$ is the sharp critical exponent. For $p<2$, the extremal configurations undergo a bifurcation, meaning that they become non-matching pairs with bimodal root structure, converging back to the Hermite diagonal only as $p\to 2^-$. Our findings demonstrate that FlowBoost, can be an effective tool of mathematical discovery in infinite-dimensional extremal problems.
Read the full abstract ▾
Probability
Replica symmetry up to the de Almeida-Thouless line in the Sherrington-Kirkpatrick model
Patrick Lopatto
We show that in the Sherrington-Kirkpatrick model at inverse temperature $\beta$ with uniform external field $h>0$, replica
symmetry holds for the set of pairs $(\beta, h)$ predicted by de Almeida and Thouless (1978). The proof proceeds by a direct analysis of the Parisi measure using the characterization provided by Jagannath and Tobasco (2017).
Read the full abstract ▾
Mathematical Physics · Cross-listed from hep-th
Universal formulae for correlators of a broad class of models
Clifford V. Johnson
A simple method is presented for deriving universal formulae for the correlators, frequently denoted $W_{g,n}(\{z_i\}), i=1,..n$, of a wide range of models of physical and mathematical interest. While many alternative methods exist for constructing such correlators, these formulae can be simply written in terms of one defining function and its derivatives. The method has been applied to the Airy and Bessel models, various minimal string and superstring theories, and their associated intersection theory settings, ordinary and various kinds of supersymmetric Weil-Petersson volumes, and more besides. For all these cases, their $W_{g,n}(\{z_i\})$ are just all specializations of the same universal formulae. A special variant of the method useful for ${N}{=}1$ supersymmetric cases is also presented. It allows for swift derivations of Norbury's three closed-form formulae for the volumes $V_{g,n}$ ($g{=}1,2,3$) of ${ N}{=}1$ supersymmetric Weil-Petersson volumes, and generalizations of them to a wider set of models. Moreover a new closed-form formula for the genus 4 case $V_{4,n}$ is derived. The straightforward method for how to derive such formulae for $g{>}4$ cases is described. Throughout, crucial roles are played by the underlying integrable KdV flows, as well as the Gel'fand-Dikii equation.
Read the full abstract ▾
Numerical Analysis · Cross-listed from physics.flu-dyn
RAPRAL v1.0: RAdiation Prediction using RAy tracing and Line-by-line methods for hypersonic air flows
Yuzhe Zhang, Qizhen Hong, Xiaoyong Wang et al. (+1 more)
A new radiation solver, RAPRAL (RAdiation Prediction based on RAy tracing and Line-by-line) implemented in C++, is developed for simulating high-temperature thermochemical nonequilibrium radiative processes. RAPRAL integrates detailed line-by-line spectral modeling with a ray-tracing solution of the radiative transfer equation, enabling accurate resolution of both spectral features and spatial radiation transport. The adopted methods and their implementation are described in detail. To assess the overall capability and accuracy of RAPRAL, we first focus on the computation of atomic and molecular bulk spectral coefficients. Through comparison with the established code in the literature, RAPRAL demonstrates its ability to accurately capture key spectral features across a wide range of conditions. Moreover, RAPRAL is applied to predict afterbody radiative heating in the Fire II flight experiment, based on a two-temperature, 11-species air flowfield. The results demonstrate that the present approach provides reliable predictions of radiative heat flux and effectively captures the dominant radiation mechanisms. Overall, the presented results demonstrate that RAPRAL is a robust tool for simulating radiative processes in hypersonic air flows, and future versions will extend its capabilities to include species relevant to planetary atmospheres.
Read the full abstract ▾
All authors Yuzhe Zhang, Qizhen Hong, Xiaoyong Wang, Quanhua Sun
Information Theory
Channel-Aware Preemptive Scheduling for Semantic Communication with Truncated Diffusion and Path Compensation
Chengyang Liang and Dong Li
Semantic communication (SemCom) presents a transformative paradigm for alleviating bandwidth limitations in mobile networks by transmitting task-relevant semantic features instead of raw data bits. While SemCom systems utilizing
diffusion models achieve superior generation quality, existing research treats semantic generation and wireless transmission as temporally independent processes. This separation neglects the intrinsic conflict between the multi-step iterative delays inherent in
diffusion models and the time-varying fading characteristics of wireless channels. To address this discrepancy, this paper proposes a channel-aware preemptive scheduling with truncated
diffusion and path compensation (CAPS-TDPC) framework. Contrary to conventional methods that require completion of the generation phase prior to transmission, the proposed framework implements a channel-driven scheduling mechanism: each user maintains a countdown inversely proportional to its instantaneous channel gain, and the user with the shortest countdown transmits immediately, regardless of whether its
diffusion process has completed. This design permits the interruption of the forward
diffusion process to enable early transmission under favorable channel conditions. In addition, a receiver-side compensation mechanism grounded in path dynamics is developed to mitigate the semantic loss resulting from such interruptions. A path deficit metric is proposed at the receiver to quantify the recovery difficulty of distinct image blocks by incorporating the velocity field of the inverse dynamics model, which allows for adaptive weighted inverse sampling. Experimental evaluations demonstrate that the proposed framework substantially reduces the end-to-end latency while maintaining the high-fidelity semantic reconstruction, thereby enhancing the system robustness in fast fading channel environments.
Read the full abstract ▾
Algebraic Geometry
Curves on the product of two $K-$trivial surfaces
Federico Moretti and Giovanni Passeri
We study curves on the product of two $K$-trivial surfaces. In the case of the product of two very general abelian surfaces $A_1\times A_2$, we prove that the minimal genus of a non-trivial curve on $A_1\times A_2$ is $6$.