Peer Revenue Returns subject to peer review

Papers

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

Mathematics

math Watch Mathematics (sign in to use)

Sunday 22 March

Classical Analysis and ODEs

Frames and Bases of Translates of Signals on Undirected Graphs

We study a shift invariant space on an undirected graphs $G$ having $N$ vertices. We obtain a characterization theorem for a system of generalized translates $\{T_{i}g : 1\leq i\leq N\}$, for $g\in C^N$, to form an orthonormal basis. Moreover, we find a necessary and sufficient condition for the system $\{T_{i}g : 1\leq i\leq m\}$, $m\leq N$, to form a linearly independent set and an orthonormal set. Further, we obtain a characterization result for a system of generalized translates which is generated by multiple generators $g_{1},...,g_{M}$ to form a frame for $C^N$. In particular, we deduce similar results for the system $\{T_{i}M_{s}g : 1\leq i,s\leq N\}$ with modulation $M_{s}$ and the spectral graph wavelet system. We also provide an illustration for the spectral graph wavelet system.
Read the full abstract

Combinatorics

More minor summation formulae

We prove determinantal-Pfaffian formulae that simultaneously generalise the Pfaffian minor summation formula of Ishikawa and Wakayama and Byun's recent minor summation formula. These formulae are based on factorisation formulae for the determinant of the sum of a skew-symmetric matrix and a rank-1 matrix. Applications include a Cauchy-type identity for skew Schur functions.
Read the full abstract

All authors Shane Chern, Theresia Eisenkölbl, Ilse Fischer, Moritz Gangl, Mona Gatzweiler, Álvaro Gutiérrez, Christian Krattenthaler, Nishu Kumari, Markus Reibnegger, Marcus Schönfelder, Atsuro Yoshida

Representation Theory · Cross-listed from quant-ph

Point-group filtering in unitary coupled-cluster ansätze: exact criticality under the Abelian filter, freeness measured near equilibrium under the full group

Filtering a unitary coupled-cluster ansatz by molecular point-group symmetry has no settled variational cost for groups with degenerate irreducible representations. We prove that the symmetry-adapted subvariety of the amplitude space is a critical subvariety of the energy, for irreducible representations of any dimension. The gradient along every removed direction vanishes there, so a quasi-Newton optimisation with exact gradients started at the reference determinant cannot distinguish the filtered from the unfiltered ansatz. Every operator invariant under the full point group already lies in the span of the pool retained by an Abelian subgroup, so the filtered pool reaches at every geometry an energy no higher than the fully invariant one reaches. A Hamiltonian-informed pool lies inside the set the equivariant pool touches, and in a symmetry-adapted basis inside the symmetry-filtered pool, which makes a published pair of parameter counts a certificate for the orbital basis. Across six molecules and two basis sets, the full-group filter changes the classical coupled-cluster correlation energy by at most $3.4 \times 10^{-12}$ millihartree in the larger basis where the point group is finite. The parameter count of the unitary ansatz falls from 75 to 30 for ammonia and from 65 to 21 for methane, measured against the Abelian filter current practice uses. The freeness measured for the full-group filter holds near equilibrium and degrades into the static-correlation regime.
Read the full abstract

2603.21009 · 22 Mar · 28 pages, 5 figures, 5 tables. v3 supersedes v1 and v2 and reverses their conclusion. The inference of variational incompleteness from a restricted dynamical Lie algebra does not follow, and the 21.8 mHa deficit reported for NH3 in STO-3G was an artifact of an incomplete excitation pool in our own code, not a cost of the filter. Title changed. One author added

Number Theory

Ties in Function Field Prime Races

The function field analogue of Chebyshev's bias was first studied by Cha. In this paper, we study *ties* in this race, namely collections of distinct congruence classes $c_1, \dots, c_k \in (\mathbb{F}_q[T] / m)^\times$ for which $$\pi(N; m, c_1) = \pi(N; m, c_2) = \dots = \pi(N; m, c_k)$$ holds for infinitely many $N$. We provide infinitely many examples of $(m, c_1, \dots, c_k)$ for which the tie holds whenever $N$ satisfies certain congruence conditions. We give two different proofs: first, via the explicit formula for prime counts in terms of $L$-functions together with a matrix analogue of Möbius inversion, where exceptional pairs of Galois-conjugate elements in the corresponding cyclotomic fields produce ties; and second, via an explicit bijection arising from the $\mathrm{GL}_2(\mathbb{F}_q)$-action. Our examples also include characteristic 2 cases.
Read the full abstract

All authors Graeme Bates, Ryan Jesubalan, Seewoo Lee, Jane Lu, Hyewon Shim

Statistics Theory · Cross-listed from econ.EM

Power Bounds and Efficiency Loss for Asymptotically Optimal Tests in IV Regression

We characterize the maximal attainable power-size gap in overidentified instrumental variables models with heteroskedastic or autocorrelated (HAC) errors. Using total variation distance and Kraft's theorem, we define the decision theoretic frontier of the testing problem. We show that Lagrange multiplier and conditional quasi likelihood ratio tests can have power arbitrarily close to size even when the null and alternative are well separated, because they do not fully exploit the reduced-form likelihood. In contrast, the conditional likelihood ratio (CLR) test uses the full reduced-form likelihood. We prove that the power-size gap of CLR converges to one if and only if the testing problem becomes trivial in total variation distance, so that CLR attains the decision theoretic frontier whenever any test can. An empirical illustration based on Yogo (2004) shows that these failures arise in empirically relevant configurations.
Read the full abstract

Group Theory

$p$-Parts of Stabilizers in Primitive Permutation Groups

Let G be a primitive permutation group on a finite set Omega. Let p^2 divide |G|, for a prime p. We show that when G is solvable, there exists a subset of Omega whose stabilizer S has the property that 1<|S|_p<|G|_p. We offer a counting argument which should be helpful when G is not solvable.
Read the full abstract

Algebraic Geometry

Motivic Cohomology and K-groups of varieties over higher local fields

For quasi-projective varieties over a higher local field $k_N$, we prove that its $K$-groups, above a suitable degree, are divisible-by-finite. We also prove the finiteness of the prime-to-$p$ torsion subgroup of certain higher Chow groups for smooth projective varieties over such fields, where $p$ denotes the final residue characteristic of $k_N$. As an application, we show that the kernel of the tame reciprocity map is uniquely $p'$-divisible. A key ingredient in achieving these results is the finiteness of étale cohomology groups over such fields.
Read the full abstract

Probability

On Path-dependent Volterra Integral Equations: Strong Well-posedness and Stochastic Numerics

The aim of this paper is to provide a comprehensive analysis of the path-dependent Stochastic Volterra Integral Equations (SVIEs), in which both the drift and the diffusion coefficients are allowed to depend on the whole trajectory of the process up to the current time. We investigate the existence and uniqueness (aka the strong well-posedness) of solutions to such equations in the $L^p$ setting, $p>0$, locally in time and their properties specifically their path regularity and flows. Then, we introduce a numerical approximation method based on an interpolated $K-$integrated Euler-Maruyama scheme to simulate numerically the process, and we prove the convergence, with an explicit rate, of this scheme towards the strong solution in the $L^p$ norm.
Read the full abstract

Information Theory

On the Bit Error Rate Fluctuation Induced by Multipath Interference in the Coherent Regime for Intra Data Center Applications

We theoretically explain how multipath interference resizes PAM-4 constellation in the coherent regime and thus increases bit error rate fluctuation in intra data centers for the first time.

Combinatorics

Imprimitive association schemes and elimination theory

We prove that a commutative association scheme is imprimitive if and only if it admits a multivariate $P$- or $Q$-polynomial structure with respect to an elimination-type monomial order. This provides a direct bridge between the classical theory of block and quotient schemes for imprimitive association schemes and elimination theory in computational commutative algebra. For an imprimitive multivariate $P$- or $Q$-polynomial association scheme, we determine the induced multivariate polynomial structures on the quotient and block schemes and describe their associated polynomials via explicit specializations, variable deletions, and rescalings of the original associated polynomials. At the level of zero-dimensional ideals, we show that the ideal of the block scheme is exactly an elimination ideal, whereas the ideal of the quotient scheme is obtained by adjoining the valency relations for the eliminated variables and then eliminating. As applications, we study direct products and crested products from the viewpoint of multivariate polynomiality, and we characterize the schemes that are multivariate $P$- or $Q$-polynomial with respect to every monomial order as precisely the direct products of univariate $P$- or $Q$-polynomial schemes. We also discuss formal duality, composition series, and several related open problems.
Read the full abstract

Information Theory

Probability of super-regular matrices and MDS codes over finite fields

Let $C$ be an $[n,k]$ linear code chosen uniformly at random over a finite field $\mathbb{F}_q$ of size $q$. The following asymptotic probability of $C$ being maximum distance separable (MDS) as $q,n,k\to\infty$ is known: If $\frac{1}{q}\binom{n}{k} \to 0$, then $P(C\ \text{is MDS}) \to 1$. We demonstrate that this growth rate is in fact a threshold by proving: If $\frac{1}{q}\binom{n}{k} \to \infty$, then $P(C\ \text{is MDS}) \to 0$. A matrix is ($\textit{contiguous}$) $\textit{super-regular}$ if all of its (contiguous) square submatrices are nonsingular. The above results imply that for any $k \times k$ matrix $A$ chosen uniformly at random over $\mathbb{F}_q$, the following hold: If $\frac{4^k/\sqrt{k}}{q} \to 0$, then $P(A \text{ is super-regular}) \to 1$. If $\frac{4^k/\sqrt{k}}{q}\to \infty$, then $P(A \text{ is super-regular}) \to 0$. We also obtain the following asymptotic probabilities for two variations of the above questions: If $\frac{1}{q}\binom{n}{k} \to \lambda \in (0,\infty)$ and $k/n\to 0$, then $P(C\ \text{is MDS}) \to e^{-\lambda}$. If $\frac{k^3/3}{q} \to \lambda \in [0,\infty]$, then $P(A \text{ is contiguous super-regular}) \to e^{-\lambda}$. The number of super-regular $3\times 3$ matrices is known to be a polynomial in $q$. We show that the number of contiguous super-regular $3\times 3$ matrices is also a polynomial. Finally, for $4\times 4$ matrices, we show that the number of super-regular matrices is not a polynomial, nor even a quasi-polynomial of period less than $7$, whereas our experimental evidence suggests that the number of contiguous super-regular matrices is a polynomial.
Read the full abstract

All authors Rathinakumar Appuswamy, Marco Bazzani, Spencer Congero, Joseph Connelly, Matthew Ekaireb, Kenneth Zeger

Saturday 21 March

Algebraic Topology

Operational Calculus on Curved Differentials: Optimal N-Complex Bounds and Persistent Homology

We establish a canonical normal form for the iterates of a curved differential in curved differential algebras (CDA). This operator calculus clarifies the underlying algebraic structure of CDAs and bypasses the need for complex combinatorics. Using this framework, we provide sharp criteria for curvature constraints to induce N-complex structures. We demonstrate that, while the nilpotency of the curvature element to the n-th power is insufficient to bound the nilpotency of d to 2n, it fundamentally guarantees a strict (4n-2)-complex structure. On the applied side, we model curvature as a filtration controller on a genuine square zero chain complex. This places us under the standard persistence stability framework and yields a Lipschitz control of barcodes with respect to degreewise curvature variation. A reproducible toy example on a four vertex flag complex illustrates the mechanism
Read the full abstract

Differential Geometry

Clairaut Generic Riemannian Maps from Nearly Kahler Manifolds

In this paper, we study Clairaut generic Riemannian map from a nearly Kahler manifold to a Riemannian manifold. Further, we obtain a condition for a Clairaut generic Riemannian map to be a totally geodesic foliation on the total manifold. Lastly, we give non-trivial examples of such Riemannian maps.
Read the full abstract

Optimization and Control

Large-Scale Resilience Planning for Wildfire-Prone Electricity-System via Adaptive Robust Optimization

Wildfire risk poses a growing challenge for electric utilities, as powerline failures can ignite wildfires while large fires can disrupt grid operations. Utilities increasingly rely on operational interventions such as Public Safety Power Shutoffs (PSPS) and fast-trip protection to mitigate ignition risk, but these measures can cause widespread service disruptions if deployed indiscriminately. Infrastructure planning decisions–such as feeder sectionalization and protection configuration–play a key role in determining how effectively these interventions can be targeted. We develop a planning framework for wildfire resilience that jointly optimizes long-term infrastructure configuration and short-term operational response under uncertain ignition risk. The problem is formulated as a tri-level optimization model capturing the interaction between infrastructure planning, wildfire risk realization, and adaptive operational decisions. To represent system-wide ignition uncertainty, we construct a data-driven uncertainty set that combines segment-level prediction intervals with group-level uncertainty budgets. Leveraging the model structure, we reformulate the problem as a tractable robust optimization model and develop a scalable column-and-constraint generation algorithm. Synthetic experiments and a large-scale case study on an investor-owned utility distribution system show that coordinated planning of sectionalization and operational mitigation strategies can substantially reduce wildfire risk while maintaining service reliability.
Read the full abstract

Information Theory · Cross-listed from quant-ph

Geometric Classification of Biased Quantum Capacity via Harmonic Translation

We establish an exact noise-model-derived characterization of quantum error correction under diagonal local phase noise. Under uniform locality, the maximal logical dimension under t-local phase errors equals Aq(n,2t+1), the classical q-ary packing function. Because no affine or stabilizer structure is imposed, nonlinear spectral supports achieve this bound and strictly exceed all affine constructions whenever Aq(n,2t+1)>Bq(n,2t+1). This follows from a harmonic translation principle: diagonal phase operators act as rigid translations in the Fourier domain, reducing the Knill-Laflamme conditions exactly to an additive non-collision constraint (S-S) cap Et={0}. For structured phase noise, exact correction is equivalent to independence in an additive Cayley graph, connecting biased quantum capacity to classical zero-error theory and the Lovasz theta function. Under mixed Pauli noise, simultaneous protection in conjugate domains incurs an intrinsic rate penalty R <= 1-(gamma_X+gamma_Z)/2, exposing a discrete harmonic uncertainty principle. In contrast with stabilizer- or graph-based frameworks, this classical correspondence is derived directly from the phase-noise model itself rather than from an auxiliary algebraic construction.
Read the full abstract

All authors Eliseo Sarmiento Rosales, Egor Maximenko, Dionisio Manuel Tun Molina, Juan Carlos Jimenez Cervantes, Jose Alberto Guzman Vega, Rodrigo Leon Morales

Dynamical Systems

Metric-Independent Expansiveness

In this article we introduce and study a natural form of expansivity, that we call metric-independent expansiveness, for group actions on metrizable spaces. This notion means expansive with respect to every compatible metric. For actions on locally compact $\sigma$-compact metric spaces, we show that this property admits a purely topological characterization: it is equivalent to what we call cocompact expansivity and to the existence of an expansive extension to the one-point compactification. We apply this characterization to ordinal spaces and to totally bounded spaces, obtaining criteria and examples that distinguish expansive actions from genuinely metric-independent ones. A central theme in these applications is that metric-independent expansiveness can be recovered from expansive compact dynamics when the boundary of the compactification is dynamically isolated. Finally, we introduce the notion of Cauchy expansiveness and prove that every uniformly continuous Cauchy expansive action extends uniquely to an expansive action on the completion.
Read the full abstract

Combinatorics

Local $\epsilon$-uniform mixing in continuous quantum walks

Let $X$ be a weighted graph and $M$ be its adjacency, Laplacian or signless Laplacian matrix. In a continuous quantum walk on $X$, local $\epsilon$-uniform mixing occurs at vertex $u$ if the $u$th column of the matrix $U(t)=e^{itM}$ can be made arbitrarily close to a vector whose all entries have equal magnitude. Using the spectral and combinatorial properties of $X$, we derive necessary conditions for local $\epsilon$-uniform mixing to occur in $X$. This includes an inequality involving all entries of each eigenvector of $M$, as well as an upper bound on the degree of vertex $u$ when $M$ is the Laplacian or signless Laplacian matrix. We use these necessary conditions to rule out local $\epsilon$-uniform mixing in numerous classes of graphs, most of which are non-regular. We also show that almost all planar graphs (resp., trees) contain a vertex that does not admit local $\epsilon$-uniform mixing for any assignment of edge weights. Furthermore, we prove if $X$ has $n$ vertices and admits local $\epsilon$-uniform mixing at a vertex contained in a subgraph with a twin, then the number of vertices of this twin subgraph must be at least $\sqrt{n}$. In particular, we establish that a graph on $n\geq 5$ vertices does not admit local $\epsilon$-uniform mixing at a vertex with a twin.
Read the full abstract

Statistics Theory

Finite Sample Smeariness on Spheres and Modulation-aware Tests

Directional and shape data often live on manifolds, where standard central limit theorems (CLT) and the associated Wald-type \(\chi^2\)-tests require curvature-dependent recalibration: the limiting normal covariance is modified by the Hessian of the Fréchet function. On spheres, under the positive-Hessian assumptions considered here, this curvature correction yields Type~I finite sample smeariness (FSS), meaning that the asymptotic modulation exceeds \(1\). This paper studies FSS on \(\mathbb{S}^m\) and develops modulation-aware tests. We first show that, for absolutely continuous distributions on \(\mathbb{S}^m\) with positive definite Hessians, Type~I FSS is unavoidable under the stated classical CLT and moment assumptions. The geometric mechanism is that, for the absolutely continuous spherical distributions considered here, positive curvature of the sphere gives the strict spectral bound \(H \prec 2I\), so that the limiting modulation is strictly larger than one. We then prove a curse-of-dimensionality result for dimensionally comparable rotationally symmetric families: their asymptotic modulation is monotone increasing with dimension and can become arbitrarily large as the support approaches a hemisphere. Motivated by these results, we derive explicit consistent plug-in estimators of the Fréchet-function Hessian and use them to construct one-sample and two-sample Hessian-corrected Hotelling-type statistics with \(\chi^2\) limits. Numerical experiments on low- and high-dimensional spheres show that the proposed modulation-aware tests have rejection behavior comparable to bootstrap-based FSS corrections, while being substantially faster due to the explicit Hessian formulas.
Read the full abstract

Information Theory

Composition Theorems for Multiple Differential Privacy Constraints

The exact composition of mechanisms for which two differential privacy (DP) constraints hold simultaneously is studied. The resulting privacy region admits an exact representation as a mixture over compositions of mechanisms of heterogeneous DP guarantees, yielding a framework that naturally generalizes to the composition of mechanisms for which any number of DP constraints hold. This result is shown through a structural lemma for mixtures of binary hypothesis tests. Lastly, the developed methodology is applied to approximate $f$-DP composition.
Read the full abstract

Statistics Theory · Cross-listed from stat.ML

Hard labels sampled from sparse targets mislead rotation invariant algorithms

One of the most common machine learning setups is logistic regression. In many classification models, including neural networks, the final prediction is obtained by applying a logistic link function to a linear score. In binary logistic regression, the feedback can be either soft labels, corresponding to the true conditional probability of the data (as in distillation), or sampled hard labels (taking values $\pm 1$). We point out a fundamental problem that arises even in a particularly favorable setting, where the goal is to learn a noise-free soft target of the form $\sigma(\mathbf{x}^{\top}\mathbf{w}^{\star})$. In the over-constrained case (i.e. the number of samples $n$ exceeds the input dimension $d$) with examples $(\mathbf{x}_i,\sigma(\mathbf{x}_i^{\top}\mathbf{w}^{\star}))$, it is sufficient to recover $\mathbf{w}^{\star}$ and hence achieve the Bayes risk. However, we prove that when the examples are labeled by hard labels $y_i$ sampled from the same conditional distribution $\sigma(\mathbf{x}_i^{\top}\mathbf{w}^{\star})$ and $\mathbf{w}^{\star}$ is $s$-sparse, then rotation-invariant algorithms are provably suboptimal: they incur an excess risk $\Omega\!\left(\frac{d-1}{n}\right)$, while there are simple non-rotation invariant algorithms with excess risk $O(\frac{s\log d}{n})$. The simplest rotation invariant algorithm is gradient descent on the logistic loss (with early stopping). A simple non-rotation-invariant algorithm for sparse targets that achieves the above upper bounds uses gradient descent on the weights $u_i,v_i$, where now the linear weight $w_i$ is reparameterized as $u_iv_i$.
Read the full abstract

All authors Avrajit Ghosh, Bin Yu, Manfred Warmuth, Peter Bartlett

Number Theory

Graham conjecture on small sets in abelian groups

A famous conjecture of Graham asserts that every set $A \subseteq \mathbb{Z}_p \setminus \{0\}$ can be ordered so that all partial sums are distinct. Although this conjecture was recently proved for sufficiently large primes by Pham and Sauermann in~\cite{PM} (combined with earlier results of \cite{BBKMM}), it remains open for general abelian groups, even in the cyclic case $\mathbb{Z}_k$. In this paper, using a recursive approach, we investigate the sequenceability of subsets $A$ in generic abelian groups for small values of $|A|$. We prove that any subset $A \subseteq G\setminus\{0\}$ with $|A| \leq 20$ is sequenceable where previously it was known only for $|A|\leq 9$. This bound is improved to $|A| \leq 22$ for zero-sum subsets. Finally, regarding the related CMPP conjecture, we show that zero-sum subsets without inverse pairs are sequenceable for $|A| \leq 23$.
Read the full abstract

All authors Simone Costa, Stefano Della Fiore, Mattia Fontana, Lluís Vena

Probability · Cross-listed from stat.CO

Surrogate-Guided Adaptive Importance Sampling for Failure Probability Estimation

We consider the sample efficient estimation of failure probabilities from expensive oracle evaluations of a limit state function via importance sampling (IS). In contrast to conventional “two stage” approaches, which first train a surrogate model for the limit state and then construct an IS proposal to estimate failure probability using separate oracle evaluations, we propose a single stage approach where a Gaussian process surrogate and a surrogate for the optimal (zero-variance) IS density are trained from shared evaluations of the oracle, making better use of a limited budget. With such an approach, small failure probabilities can be learned with relatively few oracle evaluations. We propose kernel density estimation adaptive importance sampling (KDE-AIS), which combines Gaussian process surrogates with kernel density estimation to adaptively construct the IS proposal density, leading to sample efficient estimation of failure probabilities. We show that KDE-AIS density asymptotically converges to the optimal zero-variance IS density in total variation. Empirically, KDE-AIS enables accurate and sample efficient estimation of failure probabilities compared to the state of the art, including previous work on Gaussian process based adaptive importance sampling.
Read the full abstract

Logic

Ultrapowers of determinacy models as iteration trees on HOD

In the 1990s, Steel and Woodin showed that under large cardinal hypotheses, the HOD of $L(\mathbb R)$ admits a fine-structural analysis. Although this theorem sheds light on various problems in descriptive set theory, the fine-structural representations of many fundamental objects of determinacy theory are still unknown. For example, Woodin asked whether the ultrapower of HOD by the closed unbounded filter on $\omega_1$ is given by an iteration tree on HOD according to its fine-structural extender sequence and canonical iteration strategy. In this paper, we give a positive answer to Woodin's question, not only for the closed unbounded filter but for any ultrafilter on an ordinal. The key tool that enables the solution of Woodin's problem is a recent advance in inner model theory: the Steel–Schlutzenberg theory of normalizing iteration trees, which allows us to represent HOD and its ultrapowers as normal iterates of a single countable mouse. Despite our results, the precise structure of the iteration trees that lead from HOD into its ultrapowers remains a mystery.
Read the full abstract

Mathematical Physics · Cross-listed from cond-mat.mes-hall

Super-Klein tunneling in 2D Lorentzian-type barriers in graphene

We introduce a two-dimensional model of spin-1/2 Dirac fermions in graphene subjected to a highly tunable electric field, which exhibits super-Klein tunneling. The electric field can be continuously interpolated between two limiting configurations: a uniform electrostatic Lorentzian barrier with translational invariance and a chain of well-separated electrostatic scatterers. We demonstrate that super-Klein tunneling arises naturally as a direct consequence of the intrinsic connection of the model to free-particle dynamics, a relation that is established through methods of supersymmetric quantum mechanics, which provide an elegant and analytically tractable framework. Besides the mentioned super-Klein tunneling, scale invariance of the model and invisibility of the potential for particles of specific energy are revealed, and possible routes toward experimental realization are discussed.
Read the full abstract

All authors Alonso Contreras-Astorga, Francisco Correa, Luis Inzunza, Vit Jakubsky, Raul Valencia-Torres

Combinatorics

Block Structure and Spectrum of Zero-Divisor Graphs of Lipschitz Quaternion Rings Modulo $n$

We study the adjacency spectra of zero-divisor graphs associated with Lipschitz quaternion rings modulo $n$. For an odd prime $p$, the identification $\mathbb{L}_p\cong M_2(\mathbb{F}_p)$ gives a kernel–image partition of the nonzero singular matrices and an explicit block representation of the adjacency matrix. This structure leads to a spectral decomposition, closed expressions for the spectral radius, and information on rank, nullity, eigenvalue multiplicities, and graph energy. For powers of $2$, we determine the graph at $n=2$ and use square-zero ideals in $\mathbb{L}_{2^t}$ to construct large complete subgraphs and obtain corresponding spectral and energy bounds. The resulting framework provides a unified structural approach to the odd-prime and two-adic cases and reduces large adjacency-matrix computations to substantially smaller algebraic models.
Read the full abstract

Numerical Analysis

Adjoint DSMC Method for Spatially Inhomogeneous Boltzmann Equation with General Boundary Conditions

We develop adjoint Direct Simulation Monte Carlo (DSMC) formulations for the spatially inhomogeneous Boltzmann equation with periodic, specular reflecting, diffuse thermal, and prescribed inflow boundary conditions. Periodic and specular boundaries are treated using a pathwise particle adjoint conditional on the realized event history. For diffuse thermal boundaries, we introduce a randomized-time regularization of wall-crossing events and use score-function terms to differentiate the resulting boundary probabilities. Reparameterization of the outgoing half-Maxwellian samples provides sensitivities with respect to wall temperatures and tangential wall velocities. Prescribed inflow requires a different construction because perturbations of incoming particles affect subsequent cell populations, local collision frequencies, and collision schedules. We therefore derive an ensemble adjoint based on the locally linearized Boltzmann collision operator and evaluate boundary sensitivities using local inflow-source scores, with injection counts held fixed. For a scalar objective, the dominant adjoint cost is largely independent of the number of parameters. Numerical experiments for Maxwell molecules validate the formulations against centered finite differences for thermal, mixed thermal-specular, two-sided inflow, and high-Mach Couette-flow configurations. The results demonstrate consistent gradient estimates, Monte Carlo convergence, stability with respect to the thermal regularization parameter, and accurate sensitivity calculation in a regime with limited relative statistical noise.
Read the full abstract

Probability

On the Effect of Bottlenecks in Block Spin Models

We study a bottleneck spin model with $N$ spins, split into two Curie-Weiss models at low temperature with a bottleneck between them. We propose multiple ways of how to realize such a bottleneck and study its influence on the phase transition in the thermodynamic limit $N \to \infty$. In all versions of this model we prove the existence of a threshold that determines whether or not the presence of the bottleneck is felt in the phase transition. This threshold depends on the size of the bottleneck and the interaction strength through it.
Read the full abstract

Optimization and Control

Periodic solutions of nonlinear control systems with switching: a Lie-algebraic and contraction approach

This paper is devoted to the analysis of periodic solutions of nonlinear control-affine systems with bang-bang controls. Such problems naturally arise in periodic optimal control with constrained inputs, which have, in particular, important applications in the performance optimization of chemical reactions. We reduce the problem of constructing a periodic solution to that of finding a fixed point of a composition of exponential maps. The latter problem is then addressed using the Baker-Campbell-Hausdorff-Dynkin (BCHD) formula. We establish the equivalence between periodic solutions of the original control system and those of an associated autonomous system involving iterated Lie brackets. Applying incremental stability arguments allows us to further simplify the problem to finding the equilibria of this autonomous system. The developed theory is then applied to nonlinear chemical reaction models with constrained controls.
Read the full abstract

Optimization and Control

Data-Driven Modeling of a Controlled Orthotropic Plate Using Machine Learning

We study the problem of learning the input-output map of a controlled vibrating plate with a composite structure from experimental measurements. Analytical modeling of this control system faces challenges due to the essential orthotropy and unknown damping characteristics of the material. Surrogate models based on linear regression, multilayer perceptrons, and gated recurrent units are constructed from the available sampled data. Through comparative analysis, we show that the multilayer perceptron model provides an acceptable approximation of this dynamical system, capturing the potentially nonlinear phenomena in its input-output behavior.
Read the full abstract

All authors Yongho Kim, Alexander Zuyev, Francesco Pellicano, Antonio Zippo

Information Theory · Cross-listed from cs.LG

Causally-Guided Diffusion for Stable Feature Selection

Feature selection is fundamental to robust data-centric AI, but most existing methods optimize predictive performance under a single data distribution. This often selects spurious features that fail under distribution shifts. Motivated by principles from causal invariance, we study feature selection from a stability perspective and introduce Causally-Guided Diffusion for Stable Feature Selection (CGDFS). In CGDFS, we formalized feature selection as approximate posterior inference over feature subsets, whose posterior mass favors low prediction error and low cross-environment variance. Our framework combines three key insights: First, we formulate feature selection as stability-aware posterior sampling. Here, causal invariance serves as a soft inductive bias rather than explicit causal discovery. Second, we train a diffusion model as a learned prior over plausible continuous selection masks, combined with a stability-aware likelihood that rewards invariance across environments. This diffusion prior captures structural dependencies among features and enables scalable exploration of the combinatorially large selection space. Third, we perform guided annealed Langevin sampling that combines the diffusion prior with the stability objective, which yields a tractable, uncertainty-aware posterior inference that avoids discrete optimization and produces robust feature selections. We evaluate CGDFS on open-source real-world datasets exhibiting distribution shifts. Across both classification and regression tasks, CGDFS consistently selects more stable and transferable feature subsets, which leads to improved out-of-distribution performance and greater selection robustness compared to sparsity-based, tree-based, and stability-selection baselines.
Read the full abstract

All authors Arun Vignesh Malarkkan, Xinyuan Wang, Kunpeng Liu, Denghui Zhang, Yanjie Fu