AI for Science

A network of oscillators that synchronizes perfectly computes nothing further, so an attention architecture built from synchronization must locate its computation in structured departures from agreement. We introduce the Frustrated Synchronization Network (FSN), whose token states are phases on a torus and whose entire value pathway is one learned complex coupling kernel over harmonics and a one-step delay. Each component of the kernel is a frustration in the sense of the synchronization literature. The complex phases are static Kuramoto-Sakaguchi frustration angles, the signed harmonics are repulsive Daido components, and the delay term, which couples each token to the successors of the tokens it attends to, is algebraically identical to Kuramoto-Sakaguchi coupling whose frustration angle is the data's own transition, so next-token prediction is implemented as synchronization frustrated by the data. At matched one-million-parameter and training budgets on character-level text and code, the FSN's validation loss is below a tuned RoPE-SwiGLU transformer's at every epoch measured, and the comparison survives training the baseline to convergence: every thirty-epoch enwik8 seed finishes below the transformer's converged fifty-epoch loss of 1.611, and the FSN's completed fifty-epoch runs converge to 1.5953 +/- 0.0014. A variant with every feed-forward block replaced by mean-field coupling to learned collective modes, leaving no multilayer perceptron in the stack, tracks the transformer. On natural text the unfrustrated base layer falls behind the converged transformer at every copy depth, worst on long-range copy events; the kernel reverses the deficit at every depth of four and beyond. Headline comparisons are at the one-million-parameter scale; a scale ladder is complete through four million parameters with the advantage persisting, and remaining arms are marked as in progress.

Hyperuniform systems strongly suppress long-wavelength density fluctuations, which is quantitatively characterized by the small-wavenumber scaling. In finite samples, however, accurately estimating the hyperuniformity exponent α can be challenging. The inferred value depends strongly on the range of length scales accessible in the measurement, finite-size effects, and the specific characterization method employed, whether based on Fourier-space structure factors, real-space density fluctuations, or dynamical probes such as diffusion spreadability. In particular, the structure-factor method provides the most direct estimate of α, but is sensitive to empirical low-k fitting cutoffs. The number-variance method offers a real-space Class-like diagnosis, but contributes a numerical exponent only when the finite-size data retain Class III-like scaling information. The spreadability method provides a smoother dynamic estimate and reduces configuration-level fluctuations, but requires a physically admissible long-time fitting window. Here, we develop a practical method-aware protocol for robust estimation of the effective scaling exponent α in finite-size hyperuniform point configurations, combining three complementary methods with distinct roles. Our protocol summarizes the method-specific estimates through a joint empirical estimator and reports the internal dispersion among the participating methods to determine the optimal estimate.

In this paper, we propose a general-purpose multi-dimensional symbol construction for computing an arbitrary symmetric function with digital over-the-air computation (OAC) and discuss the practical aspects of coherent aggregation. For our first contribution, we discuss the categorical representation of a symmetric function. By using this representation and leveraging the sufficiency of the histogram to evaluate a symmetric function, i.e., inspired by type-based multiple access (TBMA), we introduce a general approach to design a single set of OAC symbols to compute any digital function. For our second contribution, we use a comprehensive platform based on low-cost nodes that maintain synchronization in time, frequency, phase, and amplitude via a trigger mechanism, enabling coherent OAC experiments without Global Positioning System (GPS) or cable-based synchronization. Using measurements from the platform, we characterize the phase and amplitude statistics of the composite channel to derive a realistic impairment model for coherent OAC. Through a comprehensive analysis, we demonstrate the effectiveness of the proposed scheme under impairments captured by the proposed model

We use temporal unitary transforms to generate 16-QAM up to 220 GBd using only 50-GHz electrical bandwidth. The technique is theoretically lossless and can generate arbitrary optical waveforms beyond the bandwidth of the constituent modulators.

We classify the distributional self-embeddings of the centered first limit Bourgain-Rosenthal-Schechtman space $R_ω^{p,0}$, $1<p<\infty$. Using a Boolean rigidity principle for its canonical independent-sum realization, we show that every such embedding is induced by a finite packing of Bernoulli factors. As a consequence, we also prove that $R_ω^{p,0}$ admits no proper non-zero internal compressions. Moreover, for $p\notin2\mathbb N$, we obtain a complete description of the linear isometric embeddings of the non-centered space $R_ω^p$, and, for $p\neq2$, we determine its group of surjective linear isometries.

Michael Farber introduced a generalization of the Lusternik--Schnirelmann category for closed 1-forms. In this paper, we introduce and study a corresponding version of topological complexity for closed 1-forms. We establish analogues of the basic properties of ordinary topological complexity, including inequalities relating this invariant to the Lusternik--Schnirelmann category for closed 1-forms. We also prove a closed 1-form analogue of the navigation-function method: upper bounds for our invariant can be obtained from the dynamical properties of gradient-like flows of closed 1-forms.

This paper establishes for the first time the iterative and rigid theory of weak$^{*}$ slices within a non-metric framework, demonstrating that dual convex sets under the pure weak$^{*}$ topology can achieve localization, diameter control, and fine structural analysis. It fundamentally transforms the traditional understanding of the geometric properties of weak$^{*}$ topology and thereby pioneers a new direction in non-metric weak$^{*}$ slice geometry. By developing a new technique involving intricate manipulations of weak$^{*}$ slices and a carefully designed iterative selection process, we prove that if $M$ is a closed subspace of a G$\mathrm{\hat{a}}$teaux differentiability space $X$, then $M$ is a G$\mathrm{\hat{a}}$teaux differentiability space. As a Corollary, we get that if $X$ is a weak Asplund space and $M$ is a closed subspace of $X$, then $X$ is a G$\mathrm{\hat{a}}$teaux differentiability space. Thus, we definitively solve an open problem raised 46 years ago by D.G. Larman and R.R. Phelps (J. London Math. Soc., 20(1979), 115--127).

Evaluating visualization systems in niche domains such as genomics is challenging due to scarcity of domain experts and difficulty recruiting a representative user base. While LLM-based synthetic personas are increasingly used to ease evaluation bottlenecks, they face well-founded skepticism. Rather than weighing synthetic personas as substitutes for real users, we ask a fundamental open question: when synthetic personas evaluate a real visualization system, what do they actually produce, and how does that output change when grounded in documented human contexts? We present Sycamore, an exploratory three-condition probe design using Geranium, a search engine for multimodal genomics visualization, as a case study. Sycamore evaluates Geranium using: (1) ungrounded synthetic personas from generic LLM priors; (2) grounded synthetic personas constrained by voice-of-customer artifacts from a prior interview study; and (3) a published baseline study of real domain experts. We observe that grounding shifts synthetic feedback toward the language and concerns of documented users, while ungrounded evaluators drift toward operational specifics that real participants did not raise; both synthetic conditions, however, converge on a find-and-adapt frame and miss the image-modality preference observed in the expert study. We discuss what these observations imply for where synthetic personas might fit alongside expert studies in domain-specific visualization evaluation. All supplemental materials are available at https://osf.io/kdfr3/.

We revisit the optimization problem solved in Løkka & Zervos (2008), i.e., the maximization of dividends, in a Brownian risk model, with the possibility (not the obligation) of making capital injections. Following the approach introduced in Alvarez & Shepp (1998), Renaud & Simard (2021), Renaud et al. (2023), we consider instead absolutely continuous (AC) dividend strategies with an affine bound on the payment rates, while singular capital injections are still allowed. In addition, we incorporate a parameter for the cost of ruin or, said differently, a penalty at ruin in the performance function. We show that the solution is a so-called Løkka-Zervos dichotomy: the surplus is never ruined by making bail-out payments, or no capital is injected and bankruptcy can occur; in either case, dividends are paid at full rate when the surplus is above a threshold. Our framework allows us to provide explicit conditions to express the dichotomy, either using the cost of capital injections or the cost of ruin as a criterion, which also exposes the underlying structure of the solution. In particular, for some values of the parameters, we show that it is optimal to liquidate. Moreover, we perform a numerical analysis highlighting the range of values generated under this AC affine-bound structure.

Energy costs of information processing are growing exponentially. Bit erasure is a key problem in this energy-information nexus, and a number of seminal relationships have been deduced regarding the relationship between thermodynamic costs and memory storage. To continue making progress in the modern era, however, requires confronting thermodynamic costs in realistic physical systems which operate away from equilibrium. Here, we explore the thermodynamic costs of bit erasure in a complementary metal oxide semiconductor model of two types of random access memory. We find dynamic random access memory dissipates the least amount of energy when operated in the quasistatic limit, where errors are also minimized. By contrast, static random access memory is most efficiently operated in finite time due to the energy required to maintain the state of the bit. We demonstrate a numerically robust optimization scheme using mean field theory and automatic differentiation, finding optimal protocols compatible with electrical engineering insights. These results provide a framework for operating realistic circuits in thermodynamically advantageous ways.

Commercial fusion energy requires frameworks to assess both the scientific and economic viability of a wide variety of fusion concepts. Inspired by the Lawson criterion's ability to universally describe fusion energy gain, a generalized framework is developed to determine the economic gain of fusion power plants. The model exploits temporal equilibrium, and engineering and cost parameters normalized to the energy capture surface. The derived criteria for economic gain are therefore independent of the power plant's absolute power, impartial to the particulars of its fusion technology, and can be applied to any fusion confinement concept. The derivation of the economic gain factor, $Q_{econ}$, results in nonlinear equations with ten controlling normalized design parameters ranging from fusion power density and surface component lifetime to energy fluence, price of energy, and component efficiency and cost. These ten controlling parameters are varied over a wide range to provide high-level insights in design, finance and operational tradeoffs that improve the prospects for economically viable fusion energy.

We present a framework to compute non-Gaussian likelihoods for two-point correlation functions. The non-Gaussianity is most pronounced on large scales that will be well-measured by stage-IV weak-lensing surveys. We show how such a multivariate likelihood can be constructed and efficiently evaluated using a copula approach by incorporating exact one-dimensional marginals and a dependence structure derived from the exact multivariate likelihood. The copula likelihood is found to be in better agreement with simulated sampling distributions of correlation functions than Gaussian likelihoods, particularly on large scales. We furthermore investigate the effect of the non-Gaussian copula likelihood on posterior inference, including sampling the full parameter space of contemporary weak-lensing analyses. We find potential parameter shifts in $S_8$ on the order of one standard deviation for $1 \ 000 \ \mathrm{deg}^2$ surveys but negligible shifts for areas of $10 \ 000 \ \mathrm{deg}^2$, suggesting Gaussian likelihoods are sufficient for stage-IV surveys, though results depend on the detailed mask geometry and data-vector structure.

Duality is a foundational tool in robust and distributionally robust optimization (RO/DRO), underpinning both analytical insights and tractable reformulations. While most RO/DRO duality results are derived through saddle-point, Lagrangian, or conic arguments, this paper leverages perturbation duality. We show that this perspective provides a natural and unifying framework for deriving RO/DRO dual formulations, proving the associated duality results, and diagnosing the regularity assumptions on which they depend. First, guided by perturbation duality, we establish new duality theorems for a recent DRO framework that unifies several canonical models, including $ϕ$-divergence and Wasserstein models, through optimal transport subject to conditional moment constraints. Our results resolve an open conjecture on this DRO duality by clarifying the role of compactness: compactness itself is not necessary, but can be replaced by perturbation-based regularity conditions. Second, we revisit \emph{robust duality}, commonly described as \emph{primal-worst equals dual-best.} Using bifunctions, we unify dual-best formulations appearing in the literature and derive concise perturbation-based proofs that streamline recent results. Overall, the paper positions perturbation duality as a versatile and underutilized tool for RO and DRO, offering both conceptual unification and technical generality across a broad class of models.

We prove that for any real number $ξ\neq 0$ and any coprime integers $p>q\ge1$ such that $ξ$ is irrational or $q>1$, the image in $\mathbb{R}/\mathbb{Z}$ of the sequence $(ξ(-p/q)^n)_{n\ge 0}$ is not contained in any interval of length less than $(1+q/p-q^2/p^2)/p$.

Semidefinite programming (SDP) provides a fundamental framework for studying properties of sum-of-squares (sos) representations of nonnegative polynomials. In this paper we study the quartic forms GF = (|x|^4 + F(x))/2 associated with isoparametric polynomials F of OT-FKM type with g = 4. We characterize the sos property of GF in terms of the feasibility of an explicit SDP determined by the underlying Clifford system, and in the sos cases we obtain quantitative rank bounds for sos representations, with rigidity when m >= 3.

We study a class of maps having the Collatz function (famously related to the Collatz Conjecture) as an example, under topological and ergodic perspectives, including an approach with thermodynamic formalism. By introducing a key topology and its Borel sigma-algebra we show that recurrence implies periodicity. Moreover, we establish that if every continuous potential with finite pressure possesses some equilibrium state then we have either finiteness of cycles or infinitely many cycles sharing the same period. The existence of some continuous potential with no equilibrium state is equivalent to the unboundedness of periods of cycles. The uniqueness of periodic orbits is equivalent to the uniqueness of equilibrium state for every bounded and continuous potential. We also prove that we have either infinitely many cycles or no divergent orbits. Finally, we apply our technique to the Baker and Syracuse maps, obtaining a similar result for this general class of important maps.

Simulating a Gaussian process requires sampling from a high-dimensional Gaussian distribution, which scales cubically with the number of sample locations. Spectral methods address this challenge by exploiting the Fourier representation and treating the spectral density as a probability distribution suitable for Monte Carlo approximation. Although this probabilistic interpretation is valid for stationary processes, it is overly restrictive for the nonstationary case, where spectral densities are generally not probability measures. We propose regular Fourier features for harmonizable processes to avoid this limitation. Our method discretizes the spectral representation directly, preserving the correlation structure among spectral weights without requiring probability assumptions. Under a finite-spectral-support assumption, this yields an efficient low-rank approximation that is consistent and positive semi-definite by construction. When the spectral density is unknown, the framework extends naturally to kernel learning from data. We demonstrate the method on locally stationary and harmonizable mixture kernels, the latter with a complex-valued spectral density, and apply the kernel-learning extension to real and synthetic data.

Let $E$ be an Anderson $A$-module over $\mathbb{C}_{\infty}$. The period lattice of $E$ is related to its module of special functions by means of a non-canonical isomorphism introduced by the authors in [GM21]. In this paper, we explain how a modification of the inverse map is canonical by interpreting it as a residue morphism along the characteristic graph. This phenomenon has already been observed in various situations. The main innovation of this text is that of costability (costable admissible opens, costable site, etc.) which provides a convenient framework to develop the notion of sheaves of $E(\mathbb{C}_{\infty})$-valued meromorphic functions on the rigid analytic plane.

We establish the existence of a non-trivial, branched immersion of a closed Riemann surface $Σ$ with constant mean curvature (CMC) $H$ into any closed, orientable 3-manifold $\mathcal{M}$, for almost every prescribed value of $H$. The genus of the surface $Σ$ is bounded from above by the Heegaard genus $h$ of $\mathcal{M}$. Starting from a family of sweep-outs of $\mathcal{M}$ by surfaces of genus $h$, we apply a min-max construction for a family $\{E_{H,σ}\}_σ$ of perturbations of the energy involving the second fundamental form of the immersions to produce almost-critical points $u_k$ of $E_{H,σ}$. We then show, following ideas introduced by Rivière and developed by Pigati and Rivière, that the maps $u_k$ converge to a "CMC-parametrized varifold". This limiting object is then shown to be a smooth, branched immersion with the prescribed mean curvature $H$.

Motivated by recent developments in the theory of gravitation, we revisit the idea of topological variations, originally introduced by Wheeler and Hawking, from a rigorous perspective. Starting from a localized version of the Einstein-Hilbert variational principle, we encode the key aspects of the variational procedure in the form of a topology on a suitable space of Sobolev variational configurations, which is the final topology generated by the admissible variational maps. This framework naturally lends itself to generalization, and we rigorously introduce two distinct types of topological variations, corresponding to the infinitesimal addition of disconnected components and to infinitesimal surgeries, both motivated by related physical concepts. Using tools from the theory of Sobolev spaces and precise asymptotics, we establish dimensional obstructions for the continuity and differentiability of the Einstein-Hilbert action with respect to these variations, and show that in the extended variational framework the action does not admit critical points in dimension $n=4$, while higher dimensions are free of this problem. We also discuss the deeper geometric issue of scalar curvature blow-up of degenerating metrics within the context of our framework, and finally demonstrate the non-trivial effect of added higher order curvature terms on the critical dimension.

As a natural sequel to the study of A-motivic cohomology initiated in "On the integral part of A-motivic cohomology", we develop a notion of regulator for rigid analytically trivial Anderson A-motives. In accordance with the conjectural picture over number fields, we define it as the morphism at the level of extension modules induced by the exactness of the Hodge-Pink realization functor. The purpose of this article is twofold: first, we prove a finiteness result for A-motivic cohomology; second, under a weight assumption, we show that the source and the target of the regulator have the same dimension. It came as a surprise to the author that the image of this regulator may fail to have full rank, thereby preventing an analogue of Beilinson's celebrated conjecture from holding in our setting.

We give an example of a unital C*-algebra $\mathbf{A}$ with a computable presentation and for which neither $K_0(\mathbf{A})$ nor $K_1(\mathbf{A})$ has a computable presentation.

Ultrafast online learning is essential for high-frequency systems, such as controls for quantum computing and nuclear fusion, where adaptation must occur on sub-microsecond timescales. Meeting these requirements demands low-latency, fixed-precision computation under strict memory constraints, a regime in which conventional Multi-Layer Perceptrons (MLPs) are both inefficient and numerically unstable. We identify key properties of Kolmogorov-Arnold Networks (KANs) that align with these constraints. Specifically, we show that: (i) KAN updates exploiting B-spline locality are sparse, enabling superior on-chip resource scaling, and (ii) KANs are inherently robust to fixed-point quantization. By implementing fixed-point online training on Field-Programmable Gate Arrays (FPGAs), a representative platform for on-chip computation, we demonstrate that KAN-based online learners are significantly more efficient and expressive than MLPs across a range of low-latency and resource-constrained tasks. To our knowledge, this work is the first to demonstrate model-free online learning at sub-microsecond latencies.

Engineered systems typically separate mechanical function from information processing, whereas biological systems can exploit physical structure as a medium for information processing and computation. Motivated by this contrast, recent work in mechanics has explored embedding information-processing capabilities directly into mechanical structures. However, quantitative frameworks for evaluating such capabilities remain limited. Here we address a foundational question: how does information propagate through a solid body? Using elastic bodies as a model system, we apply information-theoretic tools to treat an elastic domain as an information encoder and quantify how information transmits from applied loads to discrete sensor locations. We further connect these measures to familiar mechanical phenomena, including Saint-Venant's effect and principal stress lines. Moving toward design, we show how geometry and architected materials can tune transmission, enabling elastic domains to either transmit or block information. Overall, this work advances quantifiable metrics and benchmark tasks for mechanical intelligence, supporting comparable designs of mechanically embodied information processing.

In Evans and Hendel's recent proof of an outstanding conjecture on the resistance distances of a family of linear 3-trees, a key technique in the proof was calculating the recursion satisfied by a family of determinants. The underlying algorithm employed to prove the conjecture converged (i.e., terminated) in the particular case studied, and the paper presented an open question on when such a procedure converges in general. This paper proves the convergence of a Laplace expansion procedure for an arbitrary family of determinants of banded, square, Toeplitz matrices. A comparison of the procedure presented in this paper, the paper by Evans and Hendel, and a paper by Jia, Yang, and Li is presented.

We study monitorable sets from a topological standpoint. In particular, we use descriptive set theory to describe the complexity of the family of monitorable sets in a countable space $X$. When $X$ is second countable, we observe that the family of monitorable sets is $Π^0_3$ and determine the exact complexities it can have. In contrast, we show that if $X$ is not second countable then the family of monitorable sets can be much more complex, giving an example where it is $ Π^1_1$-complete.

We modify the proof of the disc embedding theorem for $4$-manifolds, which appeared as Theorem 5.1A in the book "Topology of 4-manifolds" by Freedman and Quinn, in order to construct geometrically dual spheres. These were claimed in the statement but not constructed in the proof. We also prove Proposition 1.6 from the Freedman-Quinn book regarding generic homotopies of discs or spheres in a 4-manifolds, which was not proven there.

We find a layered permutation $w\in S_n$ whose Schubert polynomial $\mathfrak S_w(x_1, \dots, x_n)$ has support of size asymptotically at least $n!/4^n$. This gives precise asymptotics for the growth rate of $β(n):= \max_{w\in S_n}|\mathrm{supp}(\mathfrak S_w)|$. We find a different layered permutation $w\in S_n$ whose Grothendieck polynomial has support of size asymptotically at least $n!/e^{\sqrt{2n} \cdot \ln(n)}$ and obtain more precise asymptotics for the growth rate of $β^{\mathfrak G}(n):=\max_{w\in S_n}|\mathrm{supp}(\mathfrak G_w)|$.

We give a short and relatively elementary proof of the Hilton-Milner Theorem.

This article focuses on establishing a positive weak solution to a class of semipositone problems over the Heisenberg group $\mathbb{H}^N$. In particular, we are interested in the positive weak solution to the following problem: \begin{equation}\label{p1} -Δ_{\mathbb{H}}u= g(ξ)f_a(u) \text{ in } \mathbb{H}^N \tag{$P_a$}, \end{equation} where $a>0$ is a real parameter and $g$ is a positive function. The function $f_a: \mathbb{R} \rightarrow \mathbb{R}$ is continuous and of semipositone type which means it becomes negative on some parts of the domain. Due to this sign-changing nonlinearity, we can not directly apply the maximum principle to obtain the positivity of the solution to \eqref{p1}. For that purpose, we need some regularity results for our solutions. In this direction, we first prove the existence of weak solutions to \eqref{p1} via the mountain pass technique. Further, we establish some regularity properties of our solutions and using that we prove the $L^\infty$-norm convergence of the sequence of solutions $\{u_a\}$ to a positive function $u$ as $a \rightarrow 0$, which yields $u_a \geq 0$ for $a$ sufficiently small. Finally, we use the Riesz-representation formula to obtain the positivity of solutions under some extra hypothesis on $f_0$ and $g$. To the best of our knowledge, there is no article dealing with semipositone problems in Heisenberg group set up.