Peer Revenue Past citations are no guarantee of future returns

Papers

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

Mathematics

math Watch Mathematics (sign in to use)

Saturday 16 May

Quantum Algebra

Shifted quantum toroidal algebra of type $\mathfrak{gl}_{1|1}$ and the Pieri rule of the super Macdonald polynomials

The super Macdonald polynomials indexed by the super partitions form a basis of the level zero super Fock module (combinatorial representation) of the quantum toroidal algebra $\mathcal{U}_{q,t}(\widehat{\widehat{\mathfrak{gl}}}_{1|1})$. The action of the super charges of $\mathcal{U}_{q,t}(\widehat{\widehat{\mathfrak{gl}}}_{1|1})$ implies the Pieri rule of the super Macdonald polynomials. We can express the Pieri rule in terms of differential operators in the power sums $p_k$ and the fermionic power sums $\pi_k$, which leads to the operators on the Fock space of a free boson and a free fermion. From the Pieri rule we compute the supersymmetric Hamiltonians given by the anti-commutator of the super charges and recover the results previously obtained in the literature. It is remarkable that we have to deal with a shifted quantum toroidal algebra.
Read the full abstract

Dynamical Systems

A Mathematical Characterization of Neural Activation Induced by Temporal Interference Stimulation

Temporal Interference Stimulation (TIS) is a non-invasive neuromodulation technique in which two high-frequency sinusoidal currents with slightly different frequencies generate a low-frequency envelope that can activate deep neural structures. This study investigates the conditions under which TIS elicits action potentials in a single neuron modeled by the FitzHugh-Nagumo system. This research integrates phase-plane analysis and geometric singular perturbation to develop a mathematical framework for analyzing TIS. By combining a mathematical analysis of differential equations with computer simulations, the study elucidates how the amplitudes and beat frequency jointly determine whether the neuron remains quiescent, exhibits only transient responses, or undergoes persistent (tonic) firing.
Read the full abstract

Dynamical Systems · Cross-listed from q-bio.QM

Viability Space Decomposition: A geometric partition of survival outcomes in single- and multi-agent systems

What determines whether an organism or collective will survive under particular conditions? This question is asked across the life sciences when determining adaptive fit, developing efficacious treatments for diseases, and assessing the risks posed by ecological shifts. To aid their investigations, researchers employ models of agents which must respect particular constraints to remain alive. By constraining the dynamics of these agents to bounded viability regions, these models form a class of extended dynamical systems where transient dynamics can lead to death, making traditional attractors and separatrices insufficient for characterizing the global space of possible behaviors. To remedy this, we develop viability space decomposition, an analysis framework for ordinary differential equation models of agents with viability constraints. We first introduce the general theory, revealing how several new classes of manifolds (mortality, ordering, and collapse) permit a complete decomposition of state space into regions of qualitatively similar survival outcomes: a viability portrait. We then demonstrate the method by completely analyzing the global behavior of three models: a subcellular network, a behaving cell with the same physiology, and two coupled cell networks. Finally, we finish by discussing how the framework scales and future directions for its development and application.
Read the full abstract

Optimization and Control

Data-Driven Linear Quadratic Control Using Output-Feedback via Non-Minimal Realization

In this paper, we investigate a continuous-time linear quadratic control problem for systems with unknown matrices, where only input-output data are available. We propose an output-feedback learning framework based on a canonical nonminimal realization constructed through Kreisselmeier's adaptive filter. The filter admits an observer interpretation, which leads to an augmented system that preserves the input-output response of the realization and provides accessible state trajectories. We show that the optimal gain of this augmented system explicitly recovers the optimal gain associated with the canonical non-minimal realization, and hence achieves the optimal state-feedback solution of the original plant. Exploiting this relation and the known structure of the augmented input matrix, we develop a data-driven value iteration algorithm within the adaptive dynamic programming framework. The resulting controller is implementable from input-output data, and its performance is validated via simulations.
Read the full abstract

All authors Weijian Li, Bowen Yi, Panos J. Antsaklis, Hai Lin

Analysis of PDEs · Cross-listed from cs.LG

Propagation of Chaos in Contextual Flow Maps

We develop a quantitative statistical theory of transformers in the large-context regime by adopting the abstraction of contextual flow maps (CFMs): dynamical systems that evolve a distinguished token in the presence of a contextual measure across a stack of attention blocks. Within this framework, the finite-context model approximates an idealized infinite-context system in which the contextual measure is replaced by its underlying population, so that the context length $n$ becomes a statistical resource. Exploiting the McKean–Vlasov structure of the dynamics and the classical machinery of propagation of chaos, we establish a forward bound controlling the deviation between the finite- and infinite-context CFMs uniformly along depth, and a backward bound controlling the deviation between the corresponding training trajectories uniformly across iterations of online gradient descent. Both bounds achieve the optimal Wasserstein rate $n^{-1/d}$ for general CFMs and parametric rate $n^{-1/2}$ for a restricted class of CFMs that includes transformers as a special case. The analysis rests on a new Eulerian adjoint formulation of the loss gradient and stability estimates for the resulting forward–adjoint system, both of which may be of independent interest.
Read the full abstract

All authors Shi Chen, Zhengjiang Lin, Kaizhao Liu, Philippe Rigollet

Combinatorics

Hypermaps with hyperedges of length at most $3$

We study the computation of our recently introduced Whitney polynomial and the enumeration of the spanning hypertrees for hypermaps whose hyperedges have length at most $3$. This is a class of hypermaps where the computation of the above invariants depends only on the underlying (multi)hypergraph structure. We develop deletion-contraction formulas involving six types of generalized loops and bridges, and we prove results on special substitutions into our Whitney polynomial. We generalize the reliability polynomial and the random cluster model to hypermaps in general in such a way that they can be computed using our Whitney polynomial. Finally we explicitly count the spanning hypertrees in reciprocals of plane graphs in which every vertex has degree at most $3$.
Read the full abstract

Probability

Concentration Inequalities for Sample Cross-Covariances

This paper establishes sharp dimension-free concentration and expectation bounds for the deviation of a sample cross-covariance matrix from its mean. For sub-Gaussian random vectors, we prove a high-probability operator-norm bound governed by the effective ranks of the two marginal covariance matrices. In the Gaussian case, we prove a matching expectation lower bound, allowing arbitrary correlation between the two random vectors.
Read the full abstract

Optimization and Control

A Proximal Gradient Framework for Composite Multiobjective Optimization on Riemannian Manifolds

This paper proposes a Riemannian Multiobjective Proximal Gradient Method (RMPGM) for composite optimization problems on manifolds. Unlike scalarization-based approaches, the proposed framework directly handles vector-valued objectives and establishes global convergence to Pareto stationary points, together with an $\mathcal{O}(1/k)$ convergence rate. We further develop two variants to enhance practicality and performance: an inexact RMPGM that allows controlled inexactness in solving subproblems, and a trust-region RMPGM that adaptively adjusts the penalty parameter and achieves an $\mathcal{O}(\epsilon^{-2}) $iteration complexity. Numerical experiments demonstrate that the proposed methods are consistently outperform subgradient-based baselines.
Read the full abstract

Metric Geometry

An immersed flat polyhedral Klein bottle

We present a polyhedral surface in Euclidean 3-space with the topology of a Klein bottle such that every vertex has zero angle defect and the star of every vertex is embedded. From the perspective of metric geometry, the polyhedron can be viewed as the image of a piecewise smooth isometric immersion of a flat Klein bottle. It is apparently the first such explicit example.
Read the full abstract

Number Theory

Factorization of Additive Polynomials and van der Geer–van der Vlugt curves in characteristic 2

In our previous work, we gave a formula for the Frobenius eigenvalues of van der Geer–van der Vlugt curves in characteristic 2 by considering suitable quotients of the curve. Although the formula is explicit, it depends on many choices, which makes the formula complicated. In this article, we take a different approach using a factorization of additive polynomials, and prove a new formula. The resulting formula is simpler and is useful for explicit computations. As applications, we provide a method for constructing maximal and minimal van der Geer–van der Vlugt curves, and show that every such curve arises from this construction. We also compute various examples of van der Geer–van der Vlugt curves and study their periods.
Read the full abstract

Combinatorics

Hadamard Hypercubes

Although Hadamard matrices have been investigated since the nineteenth century, relatively little is known about their higher-dimensional analogues. In this paper, we introduce two constructions of Hadamard hypercubes. The first construction is derived from conference matrices, while the second is recursive, combining Hadamard matrices (and hypercubes) of smaller order with Latin hypercubes. The former approach draws on the theory of association schemes on triples, whereas the latter yields applications to the construction of higher-dimensional symmetric designs.
Read the full abstract

Dynamical Systems

Regularity of Lyapunov exponents at one-point Lyapunov spectra: the semisimple case

We study the regularity of Lyapunov exponents as functions on the space of compactly supported probability measures on $\mathrm{GL}(d,\mathbb{R})$. We prove that the Lyapunov exponents are pointwise log-Hölder continuous with respect to the Wasserstein distance, at semisimple probability measures with one-point Lyapunov spectrum. The proof relies on a decomposition of the action into virtually conformal subspaces and a Berry-Esseen type estimate for the random walk towards these subspaces.
Read the full abstract

Friday 15 May

Numerical Analysis · Cross-listed from cs.DC

The 2M Multiplication Algorithm for Complex Matrices

Complex matrix multiplication is typically computed using 4 real matrix multiplications (GEMMs) of the same size. The well-known 3M multiplication algorithm reduces this cost to 3 real GEMMs, together with quadratic time pre- and post-processing steps. In this paper, we reduce 3M to 2M for matrices with integer real and imaginary parts, performing complex GEMM with only 2 real GEMMs of the same size, along with quadratic time pre- and post-processing. For floating-point matrices, 2M multiplication combines naturally with the Ozaki-II scheme, yielding a practical, high-performance algorithm for computing a complex floating-point GEMM in roughly twice the time of a real GEMM of the same size. As corollaries, we derive new algorithms for symmetric rank-$k$ updates (SYRK/HERK) that internally use full rectangular GEMMs.
Read the full abstract

Algebraic Geometry

Autoequivalences of Derived Categories of Moduli Spaces of Vector Bundles

Let \(C\) be a smooth projective curve over an algebraically closed field of characteristic zero. For the moduli space \(N(r,L)\) of stable vector bundles on \(C\) of rank \(r\) with fixed determinant \(L\), we study the group of exact autoequivalences of its bounded derived category. Combining the Bondal–Orlov reconstruction theorem with the descriptions of \(\operatorname{Aut}(N(r,L))\) due to Kouvidakis–Pantev and Newstead, we obtain an explicit description of \(\operatorname{Aut}\mathrm{D^b}(N(r,L))\). We also construct extension correspondences between moduli spaces of vector bundles of different ranks and use them to define natural Fourier–Mukai type exact functors between their bounded derived categories.
Read the full abstract

Combinatorics · Cross-listed from cs.GT

The Geometry of Cooperative Game Solutions: Stratified Egalitarian Shapley Values

The space L of linear value maps on a finite-player cooperative game G^N is finite-dimensional, and admits a canonical inner product induced by the Harsanyi-dividend decomposition of G^N. We show that this inner product is intrinsic: the same value arises from any orthonormal basis of G^N with respect to the Harsanyi inner product. Within this geometry, the subspace L^{ESL} of efficient, symmetric, linear value maps admits a clean structure theorem. The induced orthogonal stratification of L by coalition size yields a canonical linear isomorphism L^{ESL} = R^{n-1}, under which every efficient symmetric linear value map decomposes uniquely into n-1 stratified epsilons, one per coalition size. The classical egalitarian Shapley family of Joosten (1996) is precisely the diagonal slice of this R^{n-1}. The orthogonal projection of any Psi in L^{ESL} onto this diagonal yields an optimal parameter eps*(Psi) equal to the weighted mean of the stratified epsilons under an explicit probability distribution {w_a} over coalition sizes, and the goodness-of-fit R^2(Psi) equals one minus the relative weighted variance of those epsilons. The framework is a literal regression-statistics analogue of the coefficient of determination. At n=4 it produces a clean three-way classification of the standard alternatives to the Shapley value: the Banzhaf value is nearly orthogonal to the egalitarian Shapley axis (R^2 ~ 1%); the equal-surplus-division value is moderately aligned (R^2 ~ 38%); the solidarity value is almost entirely aligned (R^2 ~ 99.6%). Asymptotically R^2(ESD) -> 1, R^2(So) -> 1, and R^2(Bz) -> 1/2, the last reflecting a structural identity between the efficiency defect and the egalitarian-Shapley deviation of the Banzhaf value at every coalition size.
Read the full abstract

Combinatorics · Cross-listed from cs.GT

An Axiomatic Theory of Tie-Breaking: Impossibility, Characterization, and Decomposition

We develop an abstract axiomatic theory of tie-breaking. A tie-breaking input consists of a finite set N of players, a weak order on N representing the standings to be refined, and an auxiliary information item drawn from a set on which the symmetric group Sym(N) acts. Within this minimal framework we prove three theorems. First, no tie-breaking rule producing a strict linear order can be anonymous, provided the input space contains even one intrinsically symmetric situation, a condition met in essentially every realistic application. Second, when we allow the rule to output a partition of N (rather than a strict ranking), there is a unique rule satisfying two natural axioms: it is the partition of N into orbits of the joint stabilizer of the input. Third, every reasonable strict tie-breaking rule decomposes uniquely as the canonical orbit partition followed by an arbitrary completion. The decomposition makes precise the informal observation that real tie-breaking systems are honest until forced to be arbitrary. The framework is broad enough to capture chess tournament tie-breakers, sports league regulations, voting tie-breakers, tie-breaking among symmetric players in cooperative games, and ranking by network centrality measures, all within a single uniform formalism.
Read the full abstract

Information Theory · Cross-listed from cs.SI

Conditional Entropy of Heat Diffusion on Temporal Networks

Diffusion-based information-theoretic approaches provide new theoretical and practical tools to study complex networks. So far, they have not been generalized to temporal networks. In this work, we show that common entropic measures based on modeling diffusion on graphs, such as the entropy rate and the spectral entropy, do not generalize straightforwardly to temporal networks due to the process's non-stationarity and temporal asymmetries. Instead, we propose the conditional entropy of heat diffusion as an entropic measure for continuous-time temporal networks and study its properties. We show that this quantity is monotone in time, yielding an information-theoretic analog of the second law of thermodynamics for inhomogeneous diffusion on temporal networks. We provide an upper bound and suggest a lower bound on its evolution and explain how discrepancies from it arise due to asymmetric temporal paths. We then introduce a local version of conditional entropy, designed to probe diffusion over finite temporal windows, and show that it provides an informative signal for change-point detection in continuous-time temporal networks. We evaluate the proposed methodology on synthetic benchmarks, including comparative experiments with existing nonparametric baselines in the snapshot setting, and then apply it to a real-world temporal contact network. Finally, we show how to use detected change points to guide community detection on targeted sub-intervals, improving the quality and interpretability of the clustering results.
Read the full abstract

Number Theory

On the Natural Density of Monic Integer Polynomials with Roots in a Fixed Number Field

In this article, we investigate the statistical distribution and asymptotic behavior of the family of monic integer polynomials of degree $n$ having at least one root in a fixed number field $K$. Although the framework of thin sets implies that the natural density of this family in the parameter space of bounded height is zero, explicitly quantifying this vanishing rate is a central challenge in arithmetic statistics. Employing a hybrid approach that integrates the Mahler measure, Dirichlet's unit theorem, and residue analysis of the Dedekind zeta function, we demonstrate that the rate of convergence of this density to zero is strictly dependent on the degree $n$. Specifically, we prove that the degrees of the factors induce a phase transition in the asymptotic behavior; for polynomials of degree $n = 2$, the decay rate is bounded by $O(H^{-1} \log H)$, whereas for higher degrees, the asymptotic behavior is dominated by the contribution of rational roots, yielding a bound of $O(H^{-1})$. Beyond deriving these asymptotic estimates, we apply principles from the geometry of numbers to establish explicit combinatorial bounds for counting both the reducible and irreducible components of these polynomials. These explicit bounds provide practical tools for computational evaluations within this domain.
Read the full abstract

Optimization and Control · Cross-listed from physics.bio-ph

Universal Extremum Seeking Mechanism for Lift Variation in Soaring Birds Flight: A New Paradigm in Computational Physics and Biology

In this letter, we reveal a universal, very simple extremum seeking natural feedback law and mechanism that governs, adapts, and generates in real-time, optimized lift variations for successful energy gain flight in presence of wind shear. The introduced law/mechanism, which is computationally minimal and needs only sensory information of the wind or local energy rate (i.e., model-free and data-driven) is able to characterize and replicate dynamic soaring optimized flight physics of windward climb in real-time for a variety of soaring birds species, namely wandering albatross, black-browed albatross and grey-headed albatross. We confirm the effectiveness of this new simple, real-time law by successful comparisons with sophisticated non-real-time optimal control solver and reported biological data. Our results establish the proposed mechanism as a new paradigm in soaring flight physics. That is, our results substantially advance the computational physics/biology aspects of the problem while providing a biologically plausible theory for avian soaring behavior.
Read the full abstract

General Mathematics

A Maxwell Quadratic-Form Representation of the Parallel-Plate Casimir Trace from Codimension-Three Riesz Reduction

We formulate a Maxwell version of the codimension-three Riesz/Gaussian quadratic-form representation for perfectly conducting parallel plates. This paper is the Maxwell follow-up to the scalar codimension-three Riesz/Gaussian representation theorem presented earlier in arXiv:2605.06693(2026): the same transverse Riesz reduction and prescribed-covariance quadratic-form mechanism are carried over here to the physical parallel-plate Maxwell operator. The construction is carried out in finite lateral volume $\Omega_{L,a}=T_L^2\times[0,a]$, using the physical electric-field Hilbert space of divergence-free fields satisfying the perfect-conductor tangential condition $n\times E=0$, with the static normal zero mode removed. The Maxwell curl-curl operator is defined by its closed quadratic form, and an explicit Fourier-domain analysis proves the finite-volume spectral gap, compact resolvent, and heat-trace admissibility needed for the stochastic construction. For this reduced Maxwell operator $\mathcal L_{\mathrm{Mx}}$, the codimension-three Riesz integral gives the transversely reduced Riesz mediator $g\mathcal L_{\mathrm{Mx}}^{-1}$. A prescribed heat-regularized Gaussian source with covariance $(\hbar c/g)\mathcal L_{\mathrm{Mx}}^{3/2}e^{-\tau\mathcal L_{\mathrm{Mx}}}$ then has expected quadratic Green energy equal to the heat-regularized physical Maxwell trace. The finite-volume trace is shown to be spectrally equivalent to a scalar Dirichlet channel plus a scalar Neumann channel with its constant zero mode removed. Under the standard parallel-plate interaction finite-part prescription, the large-area energy density is $$ -\frac{\pi^2\hbar c}{720a^3}. $$ The result is a representation theorem for the Maxwell parallel-plate trace under a prescribed covariance in the flat-plate geometry considered here.
Read the full abstract

General Mathematics

The Arithmetic of Chess Piece Strength on the n x n Board

On the n x n chessboard, the move totals of distinct pieces satisfy a small number of striking arithmetic identities. The total diagonal mobility of the bishop and the total 8-neighbor mobility of the king are exactly proportional, with constant n/12, valid for every n. Among nontrivial boards, the strengths of two distinct pieces drawn from a natural thirteen-piece alphabet coincide only for n in {6, 8, 12}. We define the strength of a piece P on the n x n board as the probability that a uniformly random ordered pair of distinct squares forms a legal P-move on the empty board, and prove four main results. (1) An asymptotic dichotomy classifies pieces into riders (Theta(1/n) strength) and leapers (Theta(1/n^2) strength), with explicit rational leading constants. (2) A stable-ordering theorem identifies the threshold n* = 24 beyond which the strength order becomes fixed, with a complete tabulation of every transition for 4 <= n <= 24. (3) A complete classification of strength coincidences shows they occur only at the three magic boards n in {6, 8, 12}, accompanied by the closed-form identity str(K) - str(N) = 12/(n^2(n+1)), the unique near-coincidence between bishop and knight at n = 10 (gap 0.0606%), and the bishop-king proportionality str(B)/str(K) = n/12. (4) A Strength Algebra Theorem expresses the strength of any compound army as a linear functional of a four-dimensional atomic vector, and confines strength coincidences between distinct single pieces to the three magic boards. As immediate consequences we obtain explicit strength-preserving single-piece substitution rules on each magic board, and a characterization of the 8 x 8 board as the unique nontrivial board on which the rook attains a strength matched by another piece in the alphabet (the archbishop).
Read the full abstract

General Mathematics

A Limit-Free Algebraic-Geometric Construction of Derivatives for Elementary Functions

This paper continues the author's previous work on a limit-free algebraic-geometric construction of the derivative in the class of polynomial functions and extends the proposed framework to elementary functions. Derivatives of rational power, exponential, logarithmic, trigonometric, and inverse trigonometric functions are constructed through the geometric interpretation of the tangent line, inverse symmetry, and local linear structure, without treating the limit as the initial defining mechanism. Within the proposed approach, the derivative is introduced from the outset as a functional correspondence assigning to each point the slope coefficient of the tangent line. The paper demonstrates that the classical differentiation formulas arise naturally from interconnected geometric and algebraic structures and are subsequently consistent with standard limit-based analysis. From a methodological perspective, the study proposes the logical sequence: Tangent, Local Linear Structure, Limit formalisation. Thus, the paper presents a conceptual bridge between geometric intuition, algebraic construction, and classical differential calculus.
Read the full abstract

Combinatorics · Cross-listed from cs.FL

On the transversals of Latin squares generated by nonlinear bipermutive cellular automata

In this short paper, we begin to investigate the conditions under which a generic Bipermutive Cellular Automaton (BCA) with no-boundary conditions of diameter $d$ generates a Latin square of order $N=2^{d-1}$ admitting an orthogonal mate, without relying on the linearity of the local rule. Since an orthogonal mate exists if and only if the Latin square can be partitioned into $N$ disjoint transversals, we start by characterizing the subclass of BCA whose Latin squares have a transversal on the main diagonal. In particular, we prove that the main diagonal forms a transversal if and only if the generating function of the bipermutive local rule induces an invertible CA with periodic boundary conditions on a configuration of size $d-1$. We then perform exhaustive search experiments, showing that $d=6$ is the smallest diameter for which there exist nonlinear bipermutive CA that generate Latin squares with a transversal on the main diagonal.
Read the full abstract

Analysis of PDEs · Cross-listed from cs.LG

Multi-Headed Transformer Architectures as Time-dependent Wasserstein Gradient Flows

In recent years, transformer architectures have revolutionized the field of language processing, opening the door to previously unforeseen possibilities. However, from a theoretical point of view, the mathematical models proposed in the literature often lack direct contact with the actual architectures and depend on strong simplifying assumptions. In this paper, we reduce this gap by modelling the data flow in multi-headed transformer architectures as time-dependent gradient flows for a suitable interaction energy capturing the design of the attention mechanism. The explicit dependence on time allows us to consider different weights for each head and for each layer, without imposing constraints on the initialization method. Moreover, we prove that, under a suitable integrability assumption on the evolution of the weights, each element of the $\omega$-limit set of the gradient flows is a stationary point of the interaction energy at a limiting weight distribution. Finally, we analyse the stability of the gradient flows considering perturbations of both the initial data and the weights. Specifically, on the one hand, we study the robustness of the proposed models with respect to noisy inputs, establishing a continuous dependence of the gradient flows on the initial data and uniqueness of the flows. On the other hand, we prove the $\Gamma$-convergence of the perturbed interaction energy to the unperturbed one, leading to the convergence of the corresponding gradient flows. We complement these theoretical results with numerical experiments that confirm the predicted energy-dissipation identity and clarify the asymptotic behavior of the dynamics in both the autonomous-like (Ornstein–Uhlenbeck) and the genuinely non-autonomous (oscillating-weights) regimes.
Read the full abstract

All authors Alex Massucco, Leonardo Del Grande, Marcello Carioni, Christoph Brune, Carola-Bibiane Schönlieb

Combinatorics

Brick Wall Excursions: Combinatorial Interpretation of Random Flight Moments

We study the expected distance of short uniform random walks in arbitrary dimensions with unit steps in random directions. It is known that for dimensions $d=2$ and $d=4$, all the moments of an $m$-step walk are integer. While for $d=2$, the $n$th moment can be interpreted as the number of abelian squares of length $2n$ over an alphabet with $m$ letters, for $d=4$ no interpretation was known. The goal of this paper is to provide such an interpretation, both for $d=2$ and $d=4$, in terms of $2n$-step lattice paths in dimension $m-1$. Our construction relies on a bijection between Dyck paths with a prescribed number of peaks and words of a certain type. In addition, this bijection allows us to derive closed formulas for the number of lattice paths provided with certain statistics.
Read the full abstract

Information Theory

Covert Multi-bit LLM Watermarking: An Information Theory and Coding Approach

We study the problem of multi-bit watermarking for non-autoregressive large language models (LLMs). We introduce an information-theoretic model inspired by diffusion language models (DLM), in which the encoder has limited non-causal access to token distributions within each token block. This formulation enables an information-theoretic characterization of the non-causal watermarking capacity, in which knowledge of LLM cover statistics is leveraged to enable a multi-bit covert embedding. We study the information-theoretic limits of the model by combining Gelfand–Pinsker and channel synthesis coding techniques and obtain an exact characterization of the capacity. The embedding strategy is further optimized across blocks using a constrained Markov decision process (CMDP) and we develop an explicit algorithm based on polar codes following the information-theoretic principles. We simulate the error performance on LLaDA, and provide empirical total variation (TV) analysis as a function of key randomness.
Read the full abstract

All authors Sidong Guo, Tyler Kann, Teodora Baluta, Matthieu R. Bloch

Combinatorics

On the geodesic midrange crossing constant

We introduce an analytic framework for studying the number of crossings in planar straight-line and spherical geodesic drawings; we begin by analyzing drawings in a continuous setting and then pass to finite graphs. We use this approach to derive multiple results. First, we show that the midrange crossing constant is strictly smaller than $8/(9\pi^2)$, improving the best known upper bound of Pach and Tóth from nearly thirty years ago. Second, we prove sharp lower bounds for some restricted drawing classes. For spherical geodesic drawings with uniformly distributed vertices, we determine the minimum number of crossings at every edge density. We also prove lower bounds for planar straight line drawings in compact convex domains with uniform vertex sets. Consequently, we determine the restricted midrange crossing constant to be $8/(9\pi^2)$ in both of these settings. Finally, for two-layer bipartite drawings, with straight edges between two parallel lines, we determine the midrange crossing constant to be $2/3$.
Read the full abstract

Probability · Cross-listed from q-bio.PE

Branching under First-Passage Resetting

Many biological processes, from cell division to viral lysis, are triggered when an internal stochastic variable reaches a threshold. Here we introduce Branching under First-Passage Resetting, a general framework in which replication events arise endogenously from first-passage dynamics rather than from externally imposed lifetime clocks. We show that the resulting population dynamics obey an exact renewal equation linking single-trajectory first-passage statistics to the population growth rate. This mapping shows that, for fixed offspring number and fixed mean replication time, stochastic timing fluctuations necessarily enhance growth relative to a deterministic clock. When offspring yield depends on the first-passage time, however, fluctuations have non-trivial effects and expose a fundamental yield-delay trade-off: waiting longer can increase the number of descendants, but delays all future lineages. Our framework allows us to address this optimization problem analytically, and upon application to bacteriophage lysis, gives an optimal lysis time and growth rate consistent with empirical data.
Read the full abstract

Analysis of PDEs

A Unified Integral Equation Approach to Conservation Laws for Nonlinear Schrödinger Equations

We present a unified framework for the rigorous derivation of conservation laws and related identities for nonlinear Schrödinger equations with power-type nonlinearities. This approach treats the equation in its Duhamel form and uses the space-time integrability provided by Strichartz estimates, without relying on smooth approximations or regularization procedures. It was first introduced by the third author in [20] and subsequently developed in [7, 13]. In this paper, we establish a single integral identity from which all of the laws and identities considered here follow systematically. These include the conservation of charge (mass), energy, and momentum, the pseudo-conformal conservation law, and virial-type identities.
Read the full abstract