Peer Revenue Equity in science

Papers

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

Mathematics

math Watch Mathematics (sign in to use)

Monday 9 March

Mathematical Physics

Random walks in finite Abelian groups with Birkhoff subpolytopes of doubly stochastic matrices and their physical implementation

Random walks in a finite Abelian group $G$ are studied. They use Markov chains with doubly stochastic transition matrices, in a Birkhoff subpolytope ${\cal B}(G)$ associated with the group $G$. It is shown that all future probability vectors belong to a polytope which does not depend on the transition matrices, and which shrinks during time evolution. Various quantities are used to describe the probability vectors: the majorization preorder, Lorenz values and the Gini index, entropic quantities, and the total variation distance. The general results are applied to the additive group ${\mathbb Z}(d)$, and to the Heisenberg-Weyl group $HW(d)/{\mathbb Z}(d)$. A physical implementation of random walks in ${\mathbb Z}(d)$ that involves a sequence of non-selective projective measurements, is discussed. A physical implementation of random walks in the Heisenberg-Weyl group $HW(d)/{\mathbb Z}(d)$ using a sequence of non-selective POVM measurements with coherent states, is also presented.
Read the full abstract

Information Theory

Nonlinearity Compensation for Coherent Optical Satellite Communications

Optical satellite uplinks rely on high-power optical amplifiers (HPOAs) to overcome free-space attenuation and enable long-distance transmission. However, at high power levels, fiber Kerr nonlinearity becomes significant and degrades system performance. In this work, we develop a realistic model for optical uplinks that accounts for nonlinear effects and analyze their impact, highlighting key differences from conventional longhaul fiber systems. We then introduce low-complexity digital signal processing techniques for nonlinearity compensation, based on constellation shaping via a look-up table (LUT) and a simple nonlinear phase rotation applied at the transmitter and/or receiver. The LUT also enables adaptive rate tuning according to channel conditions, enhancing robustness against link variations. Simulation results show that the proposed techniques increase the maximum acceptable link loss by up to 6 dB with negligible complexity. Finally, we show that, at the system level, propagation in the HPOA can be modeled as a simple nonlinear phase rotation, equivalent to propagation in a zero-dispersion noiseless fiber link, and fully characterized by a single parameter - the characteristic nonlinear power.
Read the full abstract

All authors Stella Civelli, Luca Potì, Enrico Forestieri, Marco Secondini

Numerical Analysis

Discontinuous Galerkin approximation of a nonlinear multiphysics problem arising in ultrasound-enhanced drug delivery

Motivated by simulations of ultrasound-enhanced drug delivery, this work presents the numerical analysis of a mathematical model that captures the influence of ultrasound waves on the diffusivity of the drug. The system under study consists of the Westervelt wave equation, accounting for the nonlinear propagation of ultrasound, coupled to a convection-diffusion equation modeling the drug concentration. In particular, drug delivery is affected by ultrasound through a pressure-dependent diffusion coefficient. The Westervelt equation is supplemented by linear absorbing boundary conditions as a means of reducing spurious reflections off the boundaries of computational domains. For spatial discretization of this multiphysics system, we employ a discontinuous Galerkin approach on simplicial meshes. Under suitable assumptions on the exact pressure and the mesh size, we first establish well-posedness, non-degeneracy, and optimal convergence rates in the energy norm for the semi-discrete pressure subproblem. The smallness of the semi-discrete pressure is then used to establish the well-posedness and convergence of the wave–convection-diffusion system under suitable regularity of the exact concentration. Finally, theoretical findings are illustrated through numerical experiments.
Read the full abstract

Information Theory

Optical Communications with Relative Intensity Noise: Channel Modeling and Information Rates

We consider optical communications with intensity modulation and direct detection affected by laser relative intensity noise (RIN). Starting from a continuous-time waveform model, we derive an equivalent discrete-time channel model. As a result of RIN, the resulting channel model exhibits signal-dependent noise with memory. Unlike the commonly-assumed model in the literature, the conditional variance of this noise term has a polynomial dependence on the symbol of interest. Finally, we study achievable information rates for this channel under practically-relevant system parameters. We take a mismatched decoding approach and compute the generalized mutual information (GMI) using a memoryless decoding metric. Our numerical results show that when the memory in the channel is ignored by the receiver, GMI saturates as the constellation size increases, and thus, dense constellations do not offer gains. We also show that this saturation results from nonsymmetric nonvanishing contributions of the symbols to the GMI.
Read the full abstract

Differential Geometry

Instanton construction of the mapping cone Thom-Smale complex

The cup product structure on the topological side of the classical Thom-Smale complex leads to the topological side of the mapping cone Thom-Smale complex. Following the spirit of Witten's analytic Morse theory, we ask whether the mapping cone Thom-Smale complex has the analytic side. However, due to the missing cup product structure on the analytic side of the classical Thom-Smale complex, it seems that we can only have a hybrid analytic-topological construction of the mapping cone Thom-Smale complex. In this paper, we overcome the cup product issue and give the purely analytic construction of the mapping cone Thom-Smale complex. More precisely, for a Morse function with the transversality condition on a closed oriented Riemannian manifold, we construct an instanton cochain complex using the eigenspaces of the mapping cone Laplacian deformed by the Morse function and two parameters. One parameter gives the classical Witten deformation. The other parameter overcomes the cup product obstacle by suppressing the norm of the given differential form. As the main result, we prove that our instanton complex is cochain isomorphic to the topologically constructed mapping cone Thom-Smale complex, and therefore it is the purely analytic construction that we need.
Read the full abstract

Numerical Analysis

Optimal Local Error Estimates for Finite Element Methods with Measure-Valued Sources

We study finite element approximations of second-order elliptic problems with measure-valued right-hand sides supported on lower-dimensional sets. The exact solution generally lacks $H^1$-regularity due to the source singularity, which limits global convergence rates of numerical methods. Using a very weak solution framework, we establish well-posedness and global error estimates for standard Lagrange finite element methods on Lipschitz polyhedral/polygonal domains. By using interior estimates techniques, we prove optimal local $L^2$- and $H^1$-error estimates in subdomains that are strictly separated from the support of the measure. Extensive numerical experiments are provided to verify the theoretical results. These results show that for Lagrange FEMs solving elliptic problems with singular right-hand sides, the loss of global convergence is purely local, and that optimal convergence rates still hold away from the singular source.
Read the full abstract

Symplectic Geometry

A remark on monoidal structure and homological mirror symmetry

For a symplectic geometry $X$, suppose the (derived) Fukaya category $\mathrm{Fuk}(X)$ of $X$ is equipped with a monoidal structure. Then its Balmer spectrum recovers a mirror $Y$ of $X$ if there exists homological mirror symmetry $\mathrm{Fuk}(X)\cong D^b\mathrm{coh}(Y)$ and the monoidal structure is the mirror of the standard one of $D^b\mathrm{coh}(Y)$. In this short note, we fill one gap of this story in the literature: we show that the monoidal structure determines the homological mirror functor $\mathrm{Fuk}(X)\to D^b\mathrm{coh}(Y)$.
Read the full abstract

Analysis of PDEs

Flexibility of Codimension One $C^{1,\theta}$ Isometric Immersions

We study the problem of constructing $C^{1,\theta}$ isometric immersions of Riemannian metrics on $n$-dimensional domains into $\mathbb{R}^{n+1}$. While the classical Nash–Kuiper theorem established the flexibility of $C^1$ isometries, subsequent work has extended this to $C^{1,\theta}$ isometries for certain $\theta$, though the optimal exponent remains unknown. In this work we show that any short immersion can be uniformly approximated by $C^{1,\theta}$ isometric immersions for $\theta< 1/(1+2(n-1))$, improving upon the previously known exponent for $n\geq 3$. The improvement is obtained via a convex integration scheme incorporating a refined iterative integration by parts procedure resting on a detailed structural analysis of error terms and the interaction of multiple frequency scales.
Read the full abstract

Combinatorics

The triplication method for constructing strong starters

The triplication method for constructing strong starters in $Z_{3m}$ from starters in $Z_{m}$ (say, a starter of order 21 from a starter of order 7) was proposed by the authors in 2025. The method reduced construction of the particular combinatorial design (a strong starter in a cyclic group) to solving a Sudoku-type problem – an independent task with its own tools and techniques available. The Sudoku-type problem was formulated in terms of the so-called triplication table constructed from a starter of order $m$. The method was applicable for odd orders $m\ge 7$ not divisible by 3. In the present paper, our previous approach is developed in two directions: (1) the definition of the triplication table is generalized, which expands possibilities for its construction to include three base starters or even “pseudostarters”; (2) the formulation of the Sudoku-type problem is broadened to embrace various scenarios of “modular encoding” and reconstruction of strong starters from its solution. A theoretical gain of these developments consists in the improved understanding of the general structure of the triplication approach. A practical outcome is elimination of the requirement that $m$ be not divisible by 3. This leads to a broader scope of strong starters obtainable by triplication: any latent strong starter of odd order $3m$ can emerge this way.
Read the full abstract

Dynamical Systems

On the minimum of $\sigma$-Brjuno functions

$\sigma$-Brjuno functions were introduced in \cite{MaMoYo_06} as an interesting variant of the classical Brjuno function, where one substitutes the $\log$ singularity at $x=0$ with the power law divergence $x^{-1/\sigma},$ $(\sigma>0).$ As in the classical case, $B_{\sigma}$ is a locally unbounded, highly irregular lower semi continuous function; from semi continuity property it easily follows that $B_{\sigma}$ admits a global minimum but to locate it is quite a challenging problem. We prove that for $\sigma=n \in \mathbb{N}$, the unique global minimum of $B_n$ is achieved at the fixed point $ [0; \overline{n+1}]$. Furthermore, we prove that these minimizers are locally stable, showing that the point of minimum remains constant for $\sigma$ in a neighborhood of $n$. Finally, we discuss the scaling behavior near these minima and we formulate a conjecture about the phase transitions for the location of the minimizer as $\sigma$ varies.
Read the full abstract

Analysis of PDEs

A note on the well-posedness of the quartic Zakharov-Kuznetsov equation on $\mathbb{R} \times \mathbb{T}$

By using a bilinear smoothing estimate recently developed in [12], together with several linear Strichartz-type estimates established therein, we improve the threshold for local well-posedness of the quartic Zakharov-Kuznetsov equation and prove that it is locally well-posed in $H^s(\mathbb{R} \times \mathbb{T})$ for all $s > \frac{1}{2}$.
Read the full abstract

Optimization and Control

Loopless Proximal Riemannian Gradient EXTRA for Distributed Optimization on Compact Manifolds

Distributed optimization has gained substantial interest in recent years due to its wide applications in machine learning. However, most of existing algorithms are designed for Euclidean spaces, leaving composite optimization on Riemannian manifolds largely unexplored. To bridge this gap, we propose the proximal Riemannian gradient EXTRA algorithm (PR-EXTRA) to solve distributed composite optimization problem with nonsmooth regularizer over compact manifolds. In each iteration, PR-EXTRA requires only a single round communication, coupled with local gradient evaluations and proximal mappings. Furthermore, a manifold projection operator is integrated to ensure the feasibility of all iterates throughout the optimization process. Theoretical analysis shows that with a constant stepsize, PR-EXTRA achieves a sublinear convergence rate of $\mathcal{O}(1/K)$ to a stationary point, matching the proximal gradient EXTRA algorithm in Euclidean spaces. Numerical experiments show the effectiveness of the proposed algorithm.
Read the full abstract

All authors Yongyang Xiong, Chen Ouyang, Keyou You, Yang Shi, Ligang Wu

Logic

On the elementary theory of the real exponential field

Assuming Schanuel's conjecture, we prove that the complete theory $T_{\exp}$ of the real exponential field is axiomatized by the axioms of definably complete exponential fields satisfying $\exp' = \exp$. This implies the result of Macintyre and Wilkie that, under the same conjecture, $T_{\exp}$ is decidable. Our approach is based on the model completeness of a similar set of axioms for the exponential function restricted to $(-1,1)$, which we prove unconditionally.
Read the full abstract

Spectral Theory

The Point Spectrum Of Periodic Quantum Trees

We study the point spectrum of a periodic quantum tree equipped with a Schrödinger type differential operator with delta-type vertex conditions, using subsets of the compact graph that generates the tree. We prove analogs of existing discrete results concerning the eigenvalues of such operators (see Aomoto, 1991 and see Banks, Garza-Vargas and Mukherjee, 2022). In particular, we define the density of states measure and find the measure of eigenvalues of the periodic tree. While most results carry over from the discrete case, a notable difference between the continuum and discrete cases is that a regular quantum periodic tree may have eigenvalues. We prove that after an arbitrarily small adjustment of edge lengths, the point spectrum of the universal cover of a compact quantum graph, with at least one cycle and the standard Schrödinger operator, is empty.
Read the full abstract

Combinatorics

On the Dual Drazin Inverse of Adjacency Matrices of Dual-number-Weighted Digraphs

The motivation of this paper is to investigate the dual Drazin inverse of adjacency matrices arising from several classes of connected dual-number-weighted digraphs over the dual complex algebra. Explicit formulas for the dual Drazin inverse of dual complex anti-triangular block matrices are derived under suitable assumptions. These results are applied to DN-DS digraphs, DN-DLS digraphs, and DN-DW digraphs. In particular, the assumptions in \cite{AMPMJM2026} are weakened for DN-DS digraphs, an open problem in \cite{AMPMJM2026} for the case $BC=0$ is generalized and solved for DN-DLS digraphs. And the group inverse result in \cite{MNSEJAA2022} for bipartite block form adjacency matrices is extended to the dual Drazin inverse for DN-DW digraphs. We further derive explicit formulas for the dual group inverse and dual Drazin inverse of another adjacency matrix for DN-DW digraphs.
Read the full abstract

All authors Yue Zhao, Daochang Zhang, Zhongshan Li, Frank J. Hall

Statistics Theory

Limiting Spectral Distribution of moderately large Kendall's correlation matrix and its application

We establish the limiting spectral distribution of Kendall's correlation matrices in the moderate high-dimensional regime where the dimension grows slower than the sample size. Our framework allows observations to be independent but not necessarily identically distributed, and accommodates both discrete and continuous data. Unlike existing results developed under i.i.d. observations, our approach remains valid under substantial distributional heterogeneity and also covers certain i.i.d. models beyond previously studied settings. Under mild symmetry and convergence conditions on some traces, we prove that the empirical spectral distribution of a properly centered and scaled Kendall's correlation matrix converges weakly almost surely to a deterministic, generally model-dependent limit. The analysis clarifies how distributional heterogeneity influences the limiting spectrum. As an application, we propose a graphical tool for detecting dependence among components in high-dimensional data and show that ignoring heterogeneity may lead to spurious detection of dependence.
Read the full abstract

Dynamical Systems · Cross-listed from cond-mat.stat-mech

Synchronization of higher-dimensional Kuramoto oscillators on networks: from scalar to matrix-weighted couplings

The Kuramoto model is the paradigmatic model to study synchronization in coupled oscillator systems. In its classical formulation, the oscillators move on the unit circle, each characterized by a scalar phase and a natural frequency, by interacting through a sinusoidal coupling. In this work, we propose a d-dimensional generalization in which oscillators are represented as unit vectors on the (d-1)-sphere and interact through a matrix-weighted network (MWN), a recently introduced framework where links are endowed with a matrix weight instead of a scalar one. We derive necessary conditions for global synchronization via a Master Stability Function approach: the existence of a synchronous solution requires identical frequency matrices across nodes and, in the MWN case, a coherence condition on the network structure. Through a suitable change of variables, the stability analysis reduces the full Nd-dimensional problem to a family of d-dimensional eigenvalue problems, each one parametrized by the eigenvalue of a suitable scalar weighted Laplacian, showing that the synchronous solution is locally stable for any positive coupling strength K on any connected network. Analytical results are complemented by numerical simulations.
Read the full abstract

Analysis of PDEs

Barta Theorem for the $p$-Laplacian and Geometric Applications

In this article, we develop a Barta-type formulation for the $p$-Laplacian on Riemannian manifolds, extending the approach of Cheung-Leung and Bessa-Montenegro from the linear to the nonlinear setting. This framework yields sharp lower bounds for the $p$-fundamental tone without any assumptions on boundary regularity. As applications, we obtain nonlinear extensions of Cheng's eigenvalue comparison theorem and the Cheng-Li-Yau estimate for $p \geq 2$ in the context of minimal immersions. In particular, under the above assumptions, the domain $\Omega$ is $p$-stable for the Schrödinger-type operator associated with the potential $\mathcal{V} = \|A\|^{p}$, where $A$ denotes the second fundamental form of the minimal immersion. In addition, we establish a lower bound for the $p$-fundamental tone in the setting where the immersion has locally bounded mean curvature. Finally, we provide a Kazdan-Kramer type characterization of the $p$-fundamental tone, offering a unified and geometric perspective on spectral bounds for the operator $p$-Laplacian.
Read the full abstract

Information Theory · Cross-listed from cs.ET

Multi-Axis Concentration Modulation for Mobile Molecular Communication Systems

Molecular communication (MC) is emerging paradigm that employs molecules as information carriers, inspired by biological signaling processes. Existing modulation schemes such as on-off keying (OOK), although simple to implement, suffer from high error probability in dynamic or hard-to-estimate channels due to their dependence on accurate channel information (CI). This work develops a unified MC constellation framework that allows higher order modulation across multiple dimensions and designs efficient constellation for dynamic MC. We propose a general multi-axis concentration modulation (MAxCM(K,M)) of modulation order M, utilizing K-dimensional constellation space with each axis corresponding to a particular molecular type, and information is jointly encoded in their concentrations. The corresponding ML decoders are derived for both static and dynamic MC under exact and partial CI. We show that the use of MAxCM can provide improvements in spectral efficiency (SE) and error rate. We then focus on a special subclass, namely multiple-axis ratio shift keying (MAxRSK), that encodes information into the concentration ratios. Its ML decoder is shown to be a weighted combiner, and design constraints are derived to enable channel-independent decoding. We study one such example, symmetric binary RSK (SBRSK), to show its robustness in dynamic channel conditions compared to OOK. Numerical investigations show significant performance gains over OOK and provide insights into optimal constellation design and receiver configurations.
Read the full abstract

K-Theory and Homology

A remark on the invariance of $K$-theory under duality

In this short remark, we explain that two examples of invariance under duality for a localizing invariant $F$ hold purely formally when $F$ is $K$-theory, whereas the general statement for arbitrary localizing invariants does not reduce to a formal statement. We record a counterexample to the claim that the universal localizing invariant is invariant under the operation of taking opposite categories, originally due to Tabuada.
Read the full abstract

Differential Geometry

Horizontal curvatures of surfaces in 3D contact sub-Riemannian Lie groups

In this paper we study horizontal curvatures for surfaces embedded in three-dimensional contact sub-Riemannian Lie groups. Using a Riemannian approximation scheme, we derive explicit formulas for horizontal Gauss curvature, horizontal mean curvature, and symplectic distortion for surfaces embedded in three dimensional Lie groups with a sub-Riemannian structure obtained by a contact form. We focus on two primary examples: the Heisenberg group and the affine-additive group. We classify surfaces of revolution within these groups that exhibit constant horizontal curvatures, often expressing their profiles through elementary or elliptic integrals.
Read the full abstract

All authors Elia Bubani, Andrea Pinamonti, Ioannis D. Platis, Dimitrios Tsolis

Information Theory

GRAND for Gaussian Intersymbol Interference Channels

Channel decoding is challenging in communication channels exhibiting memory effects. In this work, we extend the recently proposed decoding paradigm of guessing random additive noise decoding (GRAND) to channels with memory, focusing on linear Gaussian intersymbol interference (ISI) channels. To characterize error patterns (EPs) under channel memory, we introduce the concept of error bursts and define sequence reliability to quantify the likelihood of an EP. Based on sequence reliability, we develop soft GRAND for ISI channels, termed SGRAND-ISI, which generalizes SGRAND to linear Gaussian ISI channels and is equivalent to maximum-likelihood (ML) decoding. We then develop ordered reliability bit GRAND (ORBGRAND) variants based on SGRAND-ISI that replace exact sequence reliabilities with rank-based approximations to reduce online computation. The proposed framework is first established for binary-input first-order ISI channels and subsequently extended to higher-order channel memory and higher-order modulation. Numerical experiments over first- and second-order ISI channels show that the proposed algorithms achieve multiple-dB gains over GRAND algorithms that ignore channel memory and can operate within $0.1$-$0.2$ dB of the ML benchmark. Compared with the recently proposed ORBGRAND-Approximate Independence algorithm, CDF-ORBGRAND-ISI provides a gain of at least $0.5$ dB at a block error rate (BLER) of $10^{-3}$. We further evaluate the online decoding complexity of the proposed GRAND-ISI algorithms using a module-wise complexity analysis.
Read the full abstract

Combinatorics · Cross-listed from cs.AI

Agentic Neurosymbolic Collaboration for Mathematical Discovery: A Case Study in Combinatorial Design

We study mathematical discovery through the lens of neurosymbolic reasoning, where an AI agent powered by a large language model (LLM), coupled with symbolic computation tools, and human strategic direction, jointly produced a new result in combinatorial design theory. The main result of this human-AI collaboration is a tight lower bound on the imbalance of Latin squares for the notoriously difficult case $n \equiv 1 \pmod{3}$. We reconstruct the discovery process from detailed interaction logs spanning multiple sessions over several days and identify the distinct cognitive contributions of each component. The AI agent proved effective at uncovering hidden structure and generating hypotheses. The symbolic component consists of computer algebra, constraint solvers, and simulated annealing, which provides rigorous verification and exhaustive enumeration. Human steering supplied the critical research pivot that transformed a dead end into a productive inquiry. Our analysis reveals that multi-model deliberation among frontier LLMs proved reliable for criticism and error detection but unreliable for constructive claims. The resulting human-AI mathematical contribution, a tight lower bound of $4n(n{-}1)/9$, is achieved via a novel class of near-perfect permutations. The bound was formally verified in Lean 4. Our experiments show that neurosymbolic systems can indeed produce genuine discoveries in pure mathematics.
Read the full abstract

All authors Hai Xia, Carla P. Gomes, Bart Selman, Stefan Szeider

Statistics Theory

Size-Location Correlation for Set-Valued Processes: Theory, Estimation, and Laws of Large Numbers under $\rho$-Mixing

We propose a variational framework for analyzing dependence structures of convex compact random sets based on their support functions. The approach relies on the canonical even–odd decomposition on the unit sphere, which separates size-related and location-related components and induces an exact orthogonality in the sphere $L^2(\sigma)$ space. This decomposition yields an additive variance–covariance structure that is intrinsic to set-valued data and cannot be recovered from point-based or selection-based representations. Within this framework, we introduce size, location, and total covariance and correlation indices for random sets, together with compatible $\rho$-mixing coefficients for set-valued processes. The resulting dependence measures are geometrically interpretable, invariant under translations, and free of degeneracies that arise for centrally symmetric sets under classical approaches. Weak and strong laws of large numbers are established under weak stationarity, providing asymptotic stability of Minkowski averages in the $L^2(\sigma)$ support-function norm. The proposed quantities admit natural numerical realizations via directional Monte Carlo and spherical designs. Applications to interval-valued and convex-valued data, including regression with set-valued responses, illustrate how the even–odd decomposition disentangles directional location dependence from size effects beyond what can be captured by finite-dimensional summaries such as the Steiner point.
Read the full abstract

Statistics Theory

Sign Identifiability of Causal Effects in Stationary Stochastic Dynamical Systems

We study identifiability in continuous-time linear stationary stochastic differential equations with a known causal structure. Unlike existing approaches, we relax the assumption of a known diffusion matrix, thereby respecting the model's intrinsic scale invariance. Therefore, rather than recovering drift coefficients themselves, we introduce edge-sign identifiability: for a given causal structure, we ask whether the sign of a given drift entry is uniquely determined across all observational covariance matrices induced by parametrisations compatible with that structure. This leads to a trichotomy of edge-sign identifiability: identifiable, non-identifiable, and partially identifiable. This trichotomy introduces the new notion of partial identifiability to the literature, which we show is a genuine category in our setting. Under a notion of faithfulness, we derive criteria to identify membership of each category for general graphs. Applying our criteria to specific causal structures, both analogous to classical causal settings (e.g., instrumental variables) and novel cyclic settings, we determine their edge-sign identifiability and, in some cases, obtain explicit expressions for the sign of a target edge in terms of the observational covariance matrix.
Read the full abstract

Statistics Theory

Weighted Chernoff information and optimal loss exponent in context-sensitive hypothesis testing

We study binary hypothesis testing for i.i.d. observations under a multiplicative context weight. For the optimal weighted total loss, defined as the sum of weighted type-I and type-II losses, we prove the logarithmic asymptotic $$ L_n^* = \exp\{-n D_C^{\mathrm{w}}(\mathbb{P}, \mathbb{Q}) + o(n)\}, \quad n \to \infty, $$ where $D_C^{\mathrm{w}}$ is the weighted Chernoff information. The single-letter form of the exponent relies on a structural assumption that the weight factorises across observations, $\varphi(x_1^n) = \prod_{i=1}^n \varphi(x_i)$; this restriction is essential for the single-letter representation and should be distinguished from the weaker qualitative description "multiplicative context weight". The proof embeds the weighted geometric mixtures $\varphi p^\alpha q^{1-\alpha}$ into a likelihood-ratio exponential family and identifies the rate through its log-normaliser. We also derive concentration bounds for the tilted weighted log-likelihood, obtain closed forms for Gaussian, Poisson, and exponential models, and extend the exponent characterisation to finitely many hypotheses.
Read the full abstract

Numerical Analysis

Graph-Instructed Neural Networks for parametric problems with varying boundary conditions

This work addresses the accurate and efficient simulation of physical phenomena governed by parametric Partial Differential Equations (PDEs) characterized by varying boundary conditions, where parametric instances modify not only the physics of the problem but also the imposition of boundary constraints on the computational domain. In such scenarios, classical Galerkin projection-based reduced order techniques encounter a fundamental bottleneck. Parametric boundaries typically necessitate a re-formulation of the discrete problem for each new configuration, and often, these approaches are unsuitable for real-time applications. To overcome these limitations, we propose a novel methodology based on Graph-Instructed Neural Networks (GINNs). The GINN framework effectively learns the mapping between the parametric description of the computational domain and the corresponding PDE solution. Our results demonstrate that the proposed GINN-based models, can efficiently represent highly complex parametric PDEs, serving as a robust and scalable asset for several applied-oriented settings when compared with fully connected architectures.
Read the full abstract

Analysis of PDEs

Inverse boundary value problems for certain doubly nonlinear parabolic and elliptic equations

We consider an inverse boundary value problem for the doubly nonlinear parabolic equation \[ \epsilon(x)\partial_t u^m-\nabla\cdot\bigl(\gamma(x)|\nabla u|^{p-2}\nabla u\bigr)=0 \quad\text{in }(0,T)\times\Omega, \] where $p\in(1,\infty)\setminus\{2\}$, $m>0$, and the coefficients $\epsilon$ and $\gamma$ are positive. Our first main result shows that when $m>p-1$, the lateral Cauchy data determine both coefficients. The proof proceeds by reducing the parabolic inverse problem to an inverse problem for the nonlinear elliptic equation \[ -\nabla\cdot\bigl(\gamma|\nabla w|^{p-2}\nabla w\bigr)+Vw^m=0 \quad\text{in }\Omega. \] Our second main result establishes uniqueness for the pair $(\gamma,V)$ from the nonlinear Dirichlet-to-Neumann map of this elliptic equation. The argument has two steps. First, asymptotic expansions of the elliptic Dirichlet-to-Neumann map recover the weighted $p$-Laplacian Dirichlet-to-Neumann map, and and from it the coefficient $\gamma$. Second, once $\gamma$ is known, linearization at a noncritical background solution yields recovery of $V$. In dimension two we work under a simply connectedness assumption on the domain, while in dimensions $n\ge 3$ we assume that the conductivity is invariant in one known direction.
Read the full abstract

Optimization and Control · Cross-listed from cs.LG

Learning efficient representations of complex constraints for scalable optimization

Complex constraints often make real-world optimization computationally prohibitive at the scale and speed required for operational decision-making. Here we introduce PolyFormer, a PIML framework that learns compact polytopic representations of the geometry induced by complex constraints. PolyFormer captures constraint-induced geometry and transforms it into efficient polytopic reformulations, reducing the complexity of downstream optimization and enabling the use of off-the-shelf solvers. Neural parameterizations further enable rapid adaptation to varying operating conditions without retraining. Through evaluations across three important problems, i.e., large-scale resource aggregation, network-constrained optimization, and optimization under uncertainty, PolyFormer achieves online solver speedups of up to 6,400-fold and memory reductions of up to 99.87%, while maintaining small feasibility and objective errors. Together, these results establish learned geometric constraint representations as an effective and scalable route to prescriptive optimization under diverse forms of constraint complexity.
Read the full abstract

All authors Yilin Wen, Yi Guo, Bo Zhao, Wei Qi, Zechun Hu, Colin Jones, Jian Sun

Representation Theory

Gordan-Rankin-Cohen operators on the spaces of weighted densities in superdimension $1\vert 1$

The modular forms and weighted densities over the 1-dimensional manifold $M$ are transformed “alike" under the group of linear fractional changes of coordinates, so the classifications of differential operators between spaces of (A) modular forms and (B) weighted densities are sometimes identified, although they are different. Here, we solve problem B for superstrings in superdimension $(1\vert 1)$ – superizations of the result of arXiv:2404.18222. Open problems are offered.
Read the full abstract