Number Theory
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
Wednesday 8 April
We establish an asymptotic formula for the logarithmic mean value of a 1-bounded multiplicative function that is sharp in many cases of interest. We derive from it a variety of applications, making progress on several old problems. As a first application, we show that if $f$ is a completely multiplicative function taking values in $[-1,1]$ then there is a constant $c > 0$ such that for every $x \geq 3$, $$ L_f(x) := \sum_{n \leq x} \frac{f(n)}{n} > -\frac{c}{(\log x)^{1-2/\pi}}, $$ thus significantly improving on a 20-year-old result of Granville and Soundararajan. We also show that the exponent of $\log x$ in this result can be improved to $-1+o(1)$, as long as $f$ does not “behave like” the Liouville function $\lambda$ in a precise sense. As a second application, we show that for a Rademacher random completely multiplicative function $\mathbf{f}$, the probability that $L_{\mathbf{f}}(x)$ is negative is $O(\exp(-x^c))$ for some $c \in (0,1)$, thus establishing a previously conjectured bound. Finally, we obtain a converse theorem for small absolute values $|L_f(x)|$, and construct examples $f$ that show that it is (essentially) best possible.
Numerical Analysis · Cross-listed from physics.data-an
Fast and accurate noise removal by curve fitting using orthogonal polynomials
Local polynomial smoothing is a widespread technique in data analysis, and Savitzky-Golay (SG) filters are one of its most well-known realizations. In real settings, the effectiveness of SG filtering depends critically on proper tuning of its parameters, constrained in turn by repeated polynomial fitting over large data windows and for varying polynomial degrees. Standard implementations based on monomial bases and Vandermonde matrix formulations are known to suffer from ill-conditioning and unfavorable scaling as the problem size increases. In this work, we present a fast and numerically stable method for computing polynomial fitting and differentiation matrices by reformulating the problem in terms of discrete orthogonal (Chebyshev) polynomials. Exploiting their recursive structure and the intrinsic symmetry properties of the resulting matrices, we derive two algorithms designed to reduce computational overhead. Both methods significantly reduce memory usage and improve scalability with respect to the polynomial degree and window length. A discussion of the performance demonstrates that the proposed algorithms achieve orders-of-magnitude improvements in numerical accuracy compared to standard matrix multiplication, while also providing potential gains in execution time for large-scale problems. These features make the approach particularly well suited for applications requiring repeated local polynomial fits, such as the optimization of SG filters in high-resolution spectral analyses, including axion dark matter searches and the ALPHA haloscope.
Read the full abstract
Combinatorics
Extremal Mostar Index of Graphs with Given Number of Cut Edges
The Mostar index of a connected graph \(G\) is defined as \[ Mo(G)=\sum_{uv\in E(G)}\bigl|n_u(uv)-n_v(uv)\bigr|, \] where for an edge \(e=uv\), \(n_u(e)\) denotes the number of vertices of \(G\) that are closer to \(u\) than to \(v\). In this paper, we determine the maximum possible Mostar index among all connected graphs of order \(n\) with exactly \(k\) cut edges, where \(1\le k\le n-1\). We prove that the maximum value is given by \(k(n-2)+(n-k-1)k\), and the unique extremal graph is \(K_{n-k}^k\) (a complete graph on \(n-k\) vertices with \(k\) pendant edges attached to a single vertex). We also establish a sharp lower bound and characterise the extremal graphs for the minimum value. Furthermore, we extend the results to graphs with a given cyclomatic number and a given number of cut edges. Our findings complete the extremal characterisation of the Mostar index for this fundamental graph class.
Read the full abstract
Analysis of PDEs
Bourgain-Brezis-Mironescu formula for Riesz Potentials
We identify the Bourgain-Brezis-Mironescu pointwise limit of the nonlocal potential operator $(1-\alpha)\, I_\alpha(\mathcal D^\alpha f)$, $0<\alpha<1$, where $I_\alpha$ denotes the Riesz potential and $\mathcal D^\alpha$ a nonlinear fractional differential operator. Specifically, for every $f\in C_c^\infty(\mathbb R^n)$ and every $x\in \mathbb R^n$, we show that \begin{equation*} \lim_{\alpha\to 1^-} (1-\alpha)\, I_\alpha(\mathcal D^\alpha f)(x) = K_n\, I_1(|\nabla f|)(x), \end{equation*} where $K_n$ is the geometric constant appearing in the well-known Bourgain-Brezis-Mironescu formula [BBM02]. By a density argument, we further extend this result to every $f\in W^{1,1}(\mathbb R^n)$, obtaining almost everywhere convergence along subsequences.
Read the full abstract
Probability
On spectrum of sample correlation matrices from large fold tensor vectors
In this paper, we investigate the limiting spectral distribution of the sample correlation matrix, whose sample vectors are $k$-fold tensor products of $n$-dimensional vectors with i.i.d. entries. We focus on the limiting regime $n,k \to \infty$ with $k = o(n)$, and we show that the limiting spectral distribution is the Marčenko-Pastur law. As a consequence, we show that the limiting spectral distribution of the Whishart matrix from the $k$-fold tensor product of independent uniformly distributed unit vectors in $\mathbb C^n$ is the Marčenko-Pastur law.
Read the full abstract
Information Theory
Non-RS type cyclic MDS codes over finite fields via cyclotomic field reduction
Cyclic maximum distance separable (MDS for short) codes are a special subclass of linear codes and have received a lot of attention, as these codes have very important applications in many areas including quantum codes, designs and finite geometry. However, the existing construction methods for cyclic MDS codes mainly rely on strict restrictions on certain parameters or involve relatively complex procedures. In this paper, we investigate this problem further via norm reduction in cyclotomic fields. By converting the verification of the MDS property over a finite field into checking non-zero minors in characteristic zero, we propose a construction method of cyclic MDS codes over finite fields via cyclotomic field reduction. Based on this method, we obtain several cyclic MDS codes over finite fields and many non-RS type cyclic MDS codes are produced. Compared with the existing construction method in [Non-Reed-Solomon Type Cyclic MDS Codes, IEEE Trans. Inf. Theory, 71(5): 3489–3496, 2025], we give a new and relatively direct characterization method for constructing cyclic MDS codes over finite fields and the obtained non-RS cyclic MDS codes have flexible parameters in terms of the code length $n$, the defining set $J$, and the underlying finite field $\mathbb{F}_q$.
Read the full abstract
Combinatorics
Bounds for (strong) Roman $k$-dominations
Motivated by resource defense models in networks, such as protecting territories with varying legion strengths, let $k \geq 2$ be an integer. Roman $k$-domination and strong Roman $k$-domination generalize Roman, double Roman, Italian, and double Italian domination to arbitrary number of legions. The main goal of this note is establishing sharp upper bounds for the Roman and strong Roman $k$-domination numbers of connected graphs. These bounds unify and extend prior results for $k=2$ and $k=3$. We also precisely characterize the graphs achieving these bounds.
Read the full abstract
Numerical Analysis
A time-nonlocal multiphysics finite element method with Crank-Nicolson scheme for poroelasticity model with secondary consolidation
The paper studies a time-nonlocal multiphysics finite element method with Crank-Nicolson scheme for poroelasticity model with secondary consolidation. For the case where the physical parameters $\lambda,\lambda^*$ and $c_0$ are all finite positive constants, by introducing two auxiliary variables-the fluid content $\eta$ and the generalized pressure $\xi$ – the original strongly coupled poroelasticity model is reformulated into a generalized Stokes equation with time integral terms and a diffusion equation. The reformulated model not only reveals the underlying multiphysics processes in the original model, but also exhibits time-nonlocal characteristics. A time-nonlocal multiphysics finite element method is designed for the reformulated model: the spatial discretization employs high order Taylor-Hood mixed finite element method, and the temporal discretization adopts the Crank-Nicolson scheme. The time integral terms are approximated using the composite trapezoidal rule, and the integral terms $J_{\xi}^n$ and $J_{\eta}^n$ are introduced for real-time updates, which not only avoids repeated calculations and improves efficiency, but also maintains second-order temporal accuracy. The existence and uniqueness of weak solutions for the reformulated model are proved via energy estimate methods, the stability of the fully discrete time-nonlocal multiphysics finite element method is established, and optimal-order error estimates are derived using projection operator techniques. Finally, numerical example verified the theoretical results and compared the long-time convergence of the Crank-Nicolson scheme and the backward Euler scheme.
Read the full abstract
Functional Analysis
Composition operators on de Branges spaces of entire functions
This paper aims to study the boundedness and compactness of composition operators from model spaces to the Hardy Hilbert spaces in the upper half-plane. Consequently, we investigate the boundedness and compactness of composition operators on de Branges spaces of entire functions. Moreover, we observe that the boundedness of a composition operator on a regular de Branges space forces the inducing symbol to be affine; conversely, affine symbols under appropriate conditions yield bounded composition operators. Furthermore, we show that the behaviour of boundedness and compactness of composition operators on general de Branges spaces is different from that on the Paley-Wiener spaces.
Read the full abstract
Algebraic Topology
A distance between maps via interleavings of relative Sullivan algebras
In this article, we consider extended tame persistence commutative differential graded algebras (CDGAs) associated with relative Sullivan algebras. In particular, if the relative Sullivan algebra is a model for a map between spaces, then the persistence CDGA is isomorphic to the persistence object obtained by a Postnikov tower for the map with the polynomial de Rham functor in the homotopy category of extended tame persistence CDGAs. Moreover, the interleaving distance in the homotopy category (IHC) in the sense of Lanari and Scoccola enables us to introduce a pseudodistance on the homotopy set of maps via the persistence CDGA models for maps. In contrast to persistence cochain complexes, the IHC of persistence CDGAs does not coincide with the cohomology interleaving distance in general. Due to the reason, we also discuss formalities of a persistence CDGA with interleavings. Computational examples of the pseudodistances between maps are showcased.
Read the full abstract
Analysis of PDEs
Calderon-Zygmund estimates for generalized double phase equations with matrix weights
We prove Calderon-Zygmund estimates for generalized double phase equations with Orlicz growth and variable matrix weights. The operator combines a non-uniformly elliptic double phase structure with a degenerate or singular matrix weight satisfying a small log-BMO condition. Under appropriate structural assumptions, we show that higher integrability of the weighted datum yields higher integrability of the weighted gradient of weak solutions. Our results extend the existing Calderon-Zygmund theory for double phase problems and weighted elliptic equations to a unified framework capturing the interaction between Orlicz growth and matrix-weighted structures, thereby building upon and unifying the results in [BBO20] and [BCR26].
Read the full abstract
Number Theory
Boundary Cohomology of Sp6(Z): Trivial Representation
In this article, we compute the boundary cohomology of the arithmetic group Sp6(Z) with coefficients in the trivial representation. Our computation utilizes the Borel-Serre compactification and the associated spectral sequence.
Functional Analysis · Cross-listed from cs.LG
Sparse-Aware Neural Networks for Nonlinear Functionals: Mitigating the Exponential Dependence on Dimension
Deep neural networks have emerged as powerful tools for learning operators defined over infinite-dimensional function spaces. However, existing theories frequently encounter difficulties related to dimensionality and limited interpretability. This work investigates how sparsity can help address these challenges in functional learning, a central ingredient in operator learning. We propose a framework that employs convolutional architectures to extract sparse features from a finite number of samples, together with deep fully connected networks to effectively approximate nonlinear functionals. Using universal discretization methods, we show that sparse approximators enable stable recovery from discrete samples. In addition, both the deterministic and the random sampling schemes are sufficient for our analysis. These findings lead to improved approximation rates and reduced sample sizes in various function spaces, including those with fast frequency decay and mixed smoothness. They also provide new theoretical insights into how sparsity can alleviate the curse of dimensionality in functional learning.
Read the full abstract
Combinatorics
Simultaneous avoidance of length-4 patterns in ascent sequences
Ascent sequences form a central class of combinatorial objects, as they are in bijection with several important families such as (2+2)-free posets, Stoimenow matchings, and other Fishburn objects, and are enumerated by the Fishburn numbers. We study pattern avoidance in ascent sequences for the five patterns of length 4: $0101$, $0102$, $0112$, $0120$, and $0121$. These patterns arise naturally from recent work on pattern avoidance in related families of Fishburn objects, including Stoimenow matchings and (2+2)-free posets. We enumerate ascent sequences avoiding any subset of these patterns, with the exception of the sets $\{0120\}$, $\{0121\}$, and $\{0120,0121\}$, for which the enumeration remains open. Our results reveal that the corresponding avoidance classes fall into $16$ Wilf equivalence classes and exhibit a wide range of enumerative behaviour, including connections to classical sequences such as the Catalan and Fibonacci numbers, as well as polynomial formulas and rational generating functions; several of the sequences we obtain appear to be new. Our methods combine structural decompositions with generating-tree techniques and, in several cases, rely on reductions to shorter patterns via restricted growth functions. This work contributes to the broader study of pattern avoidance across Fishburn families and highlights further connections between ascent sequences and other combinatorial structures.
Read the full abstract
Optimization and Control
Linear Programming Problem Solved By a Special Substitution Method
In this paper we develop a very special substitution method for solving a general linear programming problem (LPP). Of course the substitution is a kind of elimination of variable but this method must not be confused with the so-called Fourier-Motzkin elimination. The susbtitution developed in this paper only differs by the set of criteria that a variable must verify to be substitued. Most of the criteria are associated with the cost function of the LPP. We prove that the research of the criteria is strongly polynomial. Thus, the special substitution inehrits of the strong polynomiality which characterizes the classical substitution for linear systems. Moreover, as for the classical substitution the backward substitution for finding a vertex associated with the optimum is still valid and does not require to inverse a matrix.
Read the full abstract
Analysis of PDEs
Optimal decay of heteroclinic solutions of the fractional Allen-Cahn equation with a degenerate potential
We refine the asymptotic estimates for minimizers of a class of nonlocal energy functionals of the form \[ \frac{1}{4} \iint_{\R^{2n} \setminus (\R^n \setminus \Omega)^2} \snr{u(x) - u(y)}^2 K(x - y) \,dx\,dy + \int_\Omega W(u(x)) \,dx, \] as originally studied in~\cite{DPDV}, and we prove the optimality of our improved bounds. Here, $W$ denotes a possibly degenerate oscillatory double-well potential, satisfying a polynomial control on its second derivative near the wells. The kernel~$K$ belongs to a broad class of measurable functions and is modeled on the one of the fractional Laplacian.
Read the full abstract
Analysis of PDEs
Reconstructing double-well potentials from transition layers in long-range phase coexistence models
In models of phase coexistence, the precise form of the double-well potential is of central importance, yet it cannot be derived from first principles. In this paper, we investigate an inverse problem: starting from a prescribed transition layer with power-type decay at infinity, we reconstruct the structural properties of the associated double-well potential. We focus on the case of long-range interactions, where the dependence of the potential on the layer and its derivatives is particularly delicate. Our analysis establishes a correspondence between the decay rate of the transition layer and the regularity of the potential, revealing the existence of specific patterns and the possible emergence of degeneracies.
Read the full abstract
Optimization and Control
Uncertainty Propagation in Stochastic Hybrid Systems with Dimension-Varying Resets
This paper studies probability density evolution for stochastic hybrid systems with reset maps that change the dimension of the continuous state across modes. Existing Frobenius–Perron formulations typically represent reset-induced probability transfer through boundary conditions, which is insufficient when resets map guard sets into the interior or onto lower-dimensional subsets of another mode. We develop a weak-form formulation in which reset-induced transfer is represented by the pushforward of probability flux across the guard, yielding a unified description for such systems. The proposed framework naturally captures both cases: when the reset decreases dimension, the transferred probability appears as an interior source density, whereas when the reset increases dimension, it generally appears as a singular source supported on a lower-dimensional subset. The approach is illustrated using a stochastic hybrid model in which two particles merge into one and later split back into two, demonstrating how dimension-changing resets lead to source terms beyond classical boundary-condition-based formulations.
Read the full abstract
Algebraic Topology
On the bialgebra structure of the free loop homology
We introduce a commutative product of degree $-n$ on the homology $H_\ast(X)$ of an $n$-dimensional special cubical set $X$ and lift it on the free loop homology $H_\ast(\Lambda M)$ for $M=|X|$ to be the geometric realization. These products agree with the intersection and string topology products respectively when $M$ is an oriented closed manifold, and we establish the compatibility relation between the string topology product and the standard coproduct on $H_\ast(\Lambda M).$ Motivated by the above relationship we introduce the notion of loop bialgebra for differential graded coalgebras $C$ by means of the coHochschild complex $\Lambda C.$ We calculate the loop bialgebra structure for some spaces.
Read the full abstract
Optimization and Control · Cross-listed from eess.SY
A Markov Decision Process Framework for Enhancing Power System Resilience during Wildfires under Decision-Dependent Uncertainty
Wildfires pose an increasing threat to the safety and reliability of power systems, particularly in distribution networks located in fire-prone regions. To mitigate ignition risk from electrical infrastructure, utilities often employ safety power shutoffs, which proactively de-energize high-risk lines during hazardous weather and restore them once conditions improve. While this strategy can result in temporary load loss, it helps prevent equipment damage and wildfire ignition development in the system. In this paper, we develop a state-based decision-making framework to optimize such switching actions over time, with the goal of minimizing total operational costs throughout a wildfire event. The model represents network topologies as Markov states, with transitions influenced by both exogenous weather conditions and endogenous power flow dynamics. To address the computational challenges posed by the large state and action spaces, we propose an approximate dynamic programming algorithm based on post-decision states. The effectiveness and scalability of the proposed approach are demonstrated through case studies on 54-bus and 138-bus distribution systems, showcasing its potential for enhancing wildfire resilience across different grid configurations.
Read the full abstract
Geometric Topology
Legendrian position of veering triangulations
We make a first step towards connecting the theory of veering triangulations and bicontact structures as tools for studying (pseudo-)Anosov flows: We show that given a veering triangulation corresponding to an Anosov flow with orientable stable and unstable foliations, the edges of the triangulation can be realized as Legendrian arcs with respect to a strongly adapted bicontact structure that supports the Anosov flow. Along the way, we show that every veering triangulation can be placed in `steady position', where each pair of edge projections that intersect in the orbit space only intersect once transversely. By a previous result of the author, this implies that horizontal surgery of veering triangulations correspond to horizontal Goodman surgery of pseudo-Anosov flows.
Read the full abstract
Group Theory
Relative numbers of ends and quasi-median graphs
Given a finitely generated $G$ and a subgraph $H \leq G$, the relative number of ends $e(G,H)$ is the number of ends of a Schreier graph $\mathrm{Sch}(G,H)$ and the number of coends $\tilde{e}(G,H)$ is the maximal number of $H$-infinite components of the complement of a neighbourhood of $H$ in $G$. Generalising Sageev's characterisation of codimension-one subgroups in terms of actions on CAT(0) cube complexes, we characterise the number of relative ends and the number of coends of a pair $(G,H)$ in terms of actions on quasi-median graphs.
Read the full abstract
Information Theory
Tag-based Physical-Layer Authentication Against Message Interference
Tag-based Physical-Layer Authentication (PLA) has attracted significant attention in recent years due to its low complexity, high security, and low latency. Traditional tag-based PLA schemes typically estimate tags by decoding the message and then subtracting the estimation of the message from the received signal. However, these approaches suffer from two main limitations. First, decoding errors introduce message interference that degrades authentication performance. Second, the analytical complexity of decoding errors leads to sub-optimal threshold settings, thereby limiting detection probability. To address these limitations, this paper proposes a Tag-Based Challenge-Response (TBCR) scheme and a Series Cancellation Authentication (SCA) scheme. Specifically, in the TBCR scheme, the tags are superimposed on a forwarded challenge signal, enabling the receiver to estimate tags by removing the known challenge signal rather than relying on decoding. However, the challenge-response mechanism introduces extra noise. Here, we propose the SCA scheme without the noise interference, where both the series signal generation and cancellation modules are well-designed to generate authentication signals and estimate tags, respectively. Furthermore, we derive the closed-form expressions to evaluate the robustness and security of both proposed schemes. Notably, on one hand, the optimal threshold and detection probability are derived, which theoretically reveal that the SCA scheme always achieves the ideal detection performance, while the TBCR scheme does so in the absence of noise at Alice. On the other hand, the TBCR scheme provides enhanced security at high Signal-to-Noise Ratio (SNR) regions with fewer keys. Theoretical analysis and simulation demonstrate that both proposed schemes significantly outperform the benchmarks in detection probability with reduced time complexity.
Read the full abstract
Analysis of PDEs
On the existence of vector solutions to nonlinear Schrödinger equations with weak three-wave interaction
We study a nonlinear Schrödinger system with three-wave interaction: \begin{equation*} \left\{\begin{aligned} & - \Delta u_1 = f_1(u_1) + \alpha u_2u_3 \quad \text{ in } \R^N, & - \Delta u_2 = f_2(u_2) + \alpha u_3u_1 \quad \text{ in } \R^N, & - \Delta u_3 = f_3(u_3) + \alpha u_1u_2 \quad \text{ in } \R^N, & \quad \vec{u}=(u_1,u_2,u_3)\in (H_rad^1(\R^N))^3, \end{aligned}\right. \end{equation*} where $3\leq N\leq 5$, $\alpha\in \R$ and each nonlinearity $f_i(\xi)$ satisfies the Berestycki-Lions conditions. Let $S_i$ denote the set of all least energy solutions of the scalar equation $-\Delta u = f_i(u)$ in $H_{\rm rad}^1(\R^N)$. A solution of the systems is called vector if all its components are nontrivial. We establish the existence of two distinct families of vector solutions $\{\vec{u}_\alpha\}$ with different asymptotic behaviors as $\alpha \to 0$. One family satisfies ${\rm dist}(\vec{u}_{\alpha},S_1\times S_2\times S_3) \to 0$, while another satisfies ${\rm dist}(\vec{u}_{\alpha},S_1\times S_2\times \{0\}) \to 0$. By contrast, we prove that no family of vector solutions satisfies ${\rm dist}(\vec{u}_{\alpha},S_1\times \{0\}\times \{0\}) \to 0$. Together, these results give a complete description of the asymptotic structure of vector solutions when the three-wave interaction is weak.
Read the full abstract
Optimization and Control
A Sample-Wise Adjoint Regression Framework for Mean-Field Control with Connections to Adjoint Matching
This work proposes a novel numerical approach for solving mean-field control (MFC) problems using an adjoint-based optimization framework motivated by the stochastic maximum principle (SMP). Rather than solving the adjoint processes $(Y_t,Z_t)$ explicitly, we construct sample-wise unbiased estimators of their discretized counterparts and use them to approximate the Hamiltonian control gradient through recursive regression. The control is then updated by a gradient-descent scheme. This approach differs from conventional deep-learning methods, which typically follow a “discretize-then-optimize” paradigm and directly optimize a globally parameterized control through the discretized objective. Numerical experiments demonstrate competitive, and in many cases improved performance compared with direct deep-learning approaches. The sample-wise and regression-based structure also supports scalable implementation in high-dimensional settings and makes the method attractive for generative modeling problems involving particle-based representations and distributional objectives. On the theoretical side, we establish a connection between a particular class of MFC problems and the adjoint matching framework. Using the SMP under the mean field control setting, we further show that the self-consistent critical point of the resulting mean-field adjoint matching loss coincides with the optimal control.
Read the full abstract
Mathematical Physics · Cross-listed from cond-mat.soft
Cholesteric Fingers from a Magnetic Perspective: Topology, Energetics, and Interactions
Chiral liquid crystals and chiral magnets host a wide variety of topological solitons described by closely related continuum theories, namely the Frank-Oseen and Dzyaloshinskii models. Exploiting this correspondence, we develop a unified description of cholesteric fingers in confined liquid crystals and their magnetic counterparts. Within a continuum framework including bulk and surface anisotropies, we analyze the topology, structure, interactions, and collective states of the two main finger types, CF-1 and CF-2. We show that cholesteric fingers are composite chiral solitons built from merons. CF-2 corresponds to a bimeron with unit topological charge, while CF-1 is a topologically trivial composite of two merons with identical vorticities. From a homotopic viewpoint these textures correspond to skyrmions and droplets. Strong homeotropic anchoring induces confinement effects that reshape the meron structure and redistribute topological charge across the film thickness. Isolated fingers in the homogeneous state interact repulsively and behave as particle-like objects. Periodic phases emerge when the energy of an isolated finger becomes negative, leading to nucleation-type transitions with a diverging lattice period. Degenerate finger types allow mixed periodic sequences, analogous to stacking polytypes. In a conical background, interactions become attractive due to overlap of distortion regions. Film thickness controls stability and structure: at small thickness solitons collapse, while at large thickness bimerons exhibit bistability between surface-stabilized and bulk-like states.
Read the full abstract
Information Theory
Network-Wide PAoI Guarantee in CF-mMIMO Networks with S&C Coexistence: A Unified Framework for Spatial Partitioning Toward xURLLC
As a key capability of 6G, sensing-communication (S&C) coexistence over distributed infrastructure is expected to support next-generation ultra-reliable and low-latency communication (xURLLC) applications, which demand both robust connectivity and real-time environmental awareness. This paper investigates network-wide information freshness in large-scale cell-free massive multiple-input multiple-output (CF-mMIMO) with S&C coexistence. A challenge arises from the spatial partitioning of access points (APs) into S&C roles: allocating more APs to sensing improves update generation, whereas allocating more APs to communication enhances reliable short-packet delivery. To address this, we develop a unified analytical framework by combining stochastic geometry and stochastic network calculus (SNC) to characterize the peak age of information (PAoI) violation probability (PAVP). Specifically, we derive the moment generating functions (MGFs) of sensory packet inter-arrival and service times, accounting for the joint stochastic spatial distribution of APs and users, imperfect channel state information (CSI), and finite blocklength coding (FBC). This facilitates the derivation of a tractable upper bound on the PAVP, which is minimized to determine the optimal AP partitioning. The derived bound accurately captures the performance trend and yields a minimizing partition factor that closely matches simulations. Therefore, the framework provides an efficient and low-complexity tool for network-wide PAoI guarantee and coexistence-oriented design in CF-mMIMO networks toward xURLLC.
Read the full abstract
Optimization and Control · Cross-listed from cs.LG
FlowAdam: Implicit Regularization via Geometry-Aware Soft Momentum Injection
Adaptive moment methods such as Adam use a diagonal, coordinate-wise preconditioner based on exponential moving averages of squared gradients. This diagonal scaling is coordinate-system dependent and can struggle with dense or rotated parameter couplings, including those in matrix factorization, tensor decomposition, and graph neural networks, because it treats each parameter independently. We introduce FlowAdam, a hybrid optimizer that augments Adam with continuous gradient-flow integration via an ordinary differential equation (ODE). When EMA-based statistics detect landscape difficulty, FlowAdam switches to clipped ODE integration. Our central contribution is Soft Momentum Injection, which blends ODE velocity with Adam's momentum during mode transitions. This prevents the training collapse observed with naive hybrid approaches. Across coupled optimization benchmarks, the ODE integration provides implicit regularization, reducing held-out error by 10-22% on low-rank matrix/tensor recovery and 6% on Jester (real-world collaborative filtering), also surpassing tuned Lion and AdaBelief, while matching Adam on well-conditioned workloads (CIFAR-10). MovieLens-100K confirms benefits arise specifically from coupled parameter interactions rather than bias estimation. Ablation studies show that soft injection is essential, as hard replacement reduces accuracy from 100% to 82.5%.
Read the full abstract
Optimization and Control
Nesterov Flow May Travel Infinitely Long to Converge to a Minimizer
Recent work has established that the trajectory of the Nesterov ODE, a the continuous-time model of Nesterov's accelerated gradient method, exhibits point convergence towards a minimizer of a convex potential. A natural next question is whether this point convergence can be upgraded to rectifiability, namely whether the convergent orbit has finite path length. This work provides the answer in the negative by constructing a differentiable convex potential in $\mathbb{R}^2$ for which the flow converges to its minimizer but still accumulates infinite path length. All proofs of this work are due entirely to an internal model at OpenAI.
Read the full abstract
Probability
Global in time solutions to stochastic reaction-diffusion systems with superlinear reactions satisfying a triangular control of mass
We study systems of reaction-diffusion equations perturbed by multiplicative noise, where the reaction terms satisfy quasipositivity, a triangular mass-control structure, and polynomial growth. Our results apply to a broad class of reaction-diffusion systems arising, most notably, in chemistry and biology. In the deterministic setting these assumptions are known to guarantee the global existence of solutions. In the stochastic setting, however, reaction-diffusion systems have typically been analyzed under different assumptions on the reactions that preclude many natural models, such as reversible chemical reaction networks modeled by the mass-action law, and the question of global existence and uniqueness under a mass-control structure has remained open. In this work, we show that stochastically perturbing reaction-diffusion systems with triangular control of mass by suitable multiplicative noise leads to solutions that exist for all time.