Peer Revenue Returns subject to peer review

Papers

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

All of arXiv

Tuesday 1 September

Signal Processing

Why Is Cubic-Phase Airy Beamforming Sufficient for Blockage Recovery?

Blockage is a critical challenge for near-field communications, where reliable transmission depends heavily on the line-of-sight (LoS) path and can suffer severe power degradation when that path is obstructed. Near-field Airy beams offer a promising solution for blockage mitigation by forming curved trajectories that guide energy around obstacles, and can be practically generated with phased arrays by imposing a cubic source phase. However, trajectory-based interpretations explain how Airy beams propagate, but not why cubic-phase Airy beamforming is sufficient for blockage recovery or how much received-power gain the cubic term itself contributes. To answer these questions, we identify the blockage-induced phase mismatch relative to conventional near-field focusing and quantify how successive phase orders compensate it. The resulting analysis reveals that the linear and quadratic degrees of freedom, originally used to compensate the free-space geometric phase, can be reoptimized under blockage to provide resteering and refocusing, respectively. The quadratic term can compensate the dominant quadratic component of the additional mismatch, while the Airy cubic provides the first independent correction to the remaining non-quadratic mismatch. Simulations show that linear and quadratic compensation recover most of the available gain. The Airy cubic adds only $0.1433$ dB on average, yet enables the cubic-phase family to attain $99.77\%$ of the phase-only upper bound. Residual-phase analysis further determines when the remaining higher-order components are negligible within a prescribed received-power tolerance. These results explain why cubic-phase Airy beamforming is sufficient: lower-order phase terms provide most of the recovery, while the cubic term closes nearly all of the remaining gap.
Read the full abstract

Other Condensed Matter

Inversion-symmetric topological insulators in cut-and-project binary chains

We investigate the electronic properties of binary tight-binding chains generated by the cut-and-project method for rational slopes $\alpha=p/q$, leading to periodic and inversion symmetric chains with $n=p+q$ sites. The binary structure is encoded in two hopping amplitudes $t_a$ and $t_b$. For fixed $t_a \neq t_b$, the support of the energy spectrum as a function of $p/n$ gives rise to a "Cut-and-Project butterfly". We concentrate on insulators with $M$ filled bands among a total of $n$ bands and vary $t_a/t_b$. Inversion symmetry constrains the electric polarization $P$ to $0$ or $P_q/2$ modulo a polarization quantum $P_q = \gcd(M,n)/n$. A topological transition, between two insulators that differ by their quantized polarization, occurs if and only if $n/\gcd(M,n)$ is odd. When $n/\gcd(M,n)$ is even, the two insulating regimes have a vanishing polarization and no topological transition occurs, despite the gap closing at $t_a=t_b$. When $n$ is even and $M$ odd, we find an adiabatic path between $t_a>t_b$ and $t_a<t_b$ that maintains inversion symmetry and a gap.
Read the full abstract

All authors Zhipeng Zeng, Yuge Chen, Jean-Noël Fuchs, Jianxin Zhong, Rémy Mosseri

Machine Learning

One-Layer Transformer Provably Learns Multiclass One-Nearest Neighbor in Context

We extend recent work establishing an equivalence between one-layer transformers and nearest-neighbor classifiers in the binary setting to the multiclass case. By leveraging the simplex encoding, we show that one-layer transformers with an argmax classification head behave identically to a one-nearest-neighbor classifier in the multiclass setting. This closes a gap left by prior work, whose multiclass result relied on a non-standard rounding-based approach rather than the typical argmax head used in practice.
Read the full abstract

Image and Video Processing

GazeRefine: Expert Gaze as a Test-Time Prompt for Training-Free Medical Image Segmentation

Medical image segmentation remains difficult to scale because high-performing methods typically rely on dense expert annotations and task-specific training. We introduce GazeRefine, a training-free framework that uses gaze as an inference-time prompt for zero-shot medical image segmentation. Sparse, duration-weighted fixations are converted into foreground and background priors that initialize semantic prototypes in frozen DINOv3 feature space. These prototypes are iteratively refined through foreground-background discrimination, feature-space affinity propagation, and anchoring to the initial gaze guidance, allowing segmentation to extend beyond directly fixated regions while limiting semantic drift. GazeRefine requires no segmentation masks, fine-tuning, adapters, prompt encoders, or gradient updates. We evaluate the method on gaze-annotated polyp segmentation and prostate MRI segmentation. The results show strong performance on colonoscopy images and competitive performance on prostate MRI, supporting gaze-guided prototype refinement as a promising approach for segmentation-label-efficient, human-in-the-loop medical image segmentation. Our tools and code can be found in the following repository: https://github.com/MohammedOussamaBEN/GazeRefine.git
Read the full abstract

All authors Mohammed Oussama Benyahia, Marouane Tliba, Mohamed Amine Kerkouri, Taifour Yousra, Bin Wang, Max Bengtsson, Gorkem Durak, Elif Keles, Zuheng Ming, Marek Penhaker, Azeddine Beghdadi, Ulas Bagci, Aladine Chetouani

Group Theory

Stabilizers in Coxeter groups with a transpositional action

We study simply-laced Coxeter groups whose Coxeter diagram is the line graph $\mathcal{L}\Gamma$ of some simple graph $\Gamma$; such groups act on the set of vertices of the graph by transpositions. Our goal is to describe the point stabilizer of this action in detail. For any tree $\Sigma$, we show that the stabilizer is a finite-index reflection subgroup, and identify its Coxeter diagram, thereby discovering finite-index embeddings of Coxeter groups $W(\mathcal{L}^{[2]}\Sigma) \hookrightarrow W(\mathcal{L}\Sigma)$ for any tree $\Sigma$, including infinitely many cases where the stabilizer itself is simply-laced.
Read the full abstract

Computer Vision

CMRVision: A Foundation Model for Cardiac MR Image Analysis

Cardiac magnetic resonance (CMR) imaging provides complementary information on cardiac anatomy, function, and tissue characterization across multiple sequences and views. In this work, we investigate foundation model pretraining for 2D CMR and introduce CMRVision, a CMR-specific foundation model trained using DINOv3-style self-supervised learning on a multi-center, multi-sequence cohort of 36 million CMR images. We systematically evaluate architectural and training design choices for domain-specific pretraining. CMRVision is evaluated on two downstream tasks: multi-task segmentation across cine, late gadolinium enhancement (LGE), and mapping sequences, and cine view classification. Our experiments show that CMR-specific pretraining, smaller patch sizes, and patch-level objectives consistently improve downstream performance. Across a multi-task segmentation benchmark, CMRVision achieved the strongest overall performance, outperforming prior natural-image (NI), medical-image, supervised, and CMR foundation model baselines. Improvements were modest but consistent across structures and sequences, with Dice scores ranging from 0.940-0.967 for LV and 0.855-0.905 for myocardium, and reaching 0.929 for RV, 0.920 for LA, and 0.931 for RA. The largest gains were observed for myocardium segmentation in LGE and mapping images. In a zero-shot segmentation task on unseen LGE long-axis views, the model achieved an average Dice score of 0.692, demonstrating cross-view generalization. For cine view classification, CMRVision achieved the highest average accuracy (0.906), compared to prior methods reported in the literature. These results highlight the potential of CMRVision to support robust and generalizable cardiac MRI analysis across multiple sequences and views.
Read the full abstract

Instrumentation and Methods for Astrophysics

HEL1OS on Aditya-L1 Mission: Operations, Data Processing and Monitoring of Sun in Hard X-rays

HEL1OS (High Energy L1 Orbiting X-ray Spectrometer) is the hard X-ray spectrometer on-board Aditya-L1 solar mission, monitoring the Sun continuously in the hard X-ray band of 8 keV to 150 keV from the vantage observation platform around the Sun-Earth Lagragian Point 1 (L1). HEL1OS was commissioned in end of October, 2023 and the final instrument configuration was set by end of June, 2024 after all the performance verifications. The instrument always operates in the event mode, and the down-linked data is processed in an automated pipeline to generate Science-ready products comprising mainly of solar spectra in 8 keV to 150 keV energy band. The data products of observations since December, 2023 are hosted on the internet with open access. This includes data from the performance verification phase of the payload operations. In this paper, the details of the commissioning and performance verification phases, all the stages of the fully automated data pipeline, and the data products of HEL1OS are discussed. Also a few scientific results are discussed to illustrate the timing and spectral capabilities of the instrument in monitoring Sun in the X-ray energy bands.
Read the full abstract

All authors Ravishankar B. T., Manju Sudhakar, Srikar Paavan Tadepalli, Kiran Kumar Kalisetti, Madhavan R. Menon, Anand Jain, Reenu Palawat, Brajpal Singh, Rethika T., Priyanka Upadhyay, Prapti Mittal, Amit Purohit, Srikanth M., Anuj Nandi

Graphics

MeshSuite: A Unified Benchmark for Triangle- and Mesh-Based Neural Rendering

Triangle- and mesh-based neural rendering seeks to bridge neural representations and graphics engines through triangle primitives compatible with standard rasterization. Yet existing methods were developed and evaluated under inconsistent settings, with few direct comparisons and little evaluation in graphics engines. We introduce MeshSuite, the first benchmark for systematic evaluation of engine deployment and Non-Engine Render in triangle- and mesh-based neural rendering. MeshSuite evaluates two deployment settings: Standard Deployment, using an opaque mesh pipeline with vertex colors and hardware Z-buffering; and Dedicated Deployment, using method-specific engine implementations to preserve appearance and compositing features. For mesh splatting, we also audit exported surfaces for topological and geometric integrity as downstream graphics assets. Our results show that engine deployment reduces image quality in both settings; among evaluated methods, MeshSplatting degrades least under Standard Deployment. Dedicated Deployment preserves most fidelity at roughly 5-33x the runtime of Standard Deployment across two deployment datasets. For mesh splatting, explicit connectivity and shared indexing remain limited: shared vertex indexing alone guarantees neither manifoldness nor global connectivity. These findings show that rasterizability is only part of graphics readiness and underscore the need to assess engine deployment. Project page: https://x-up-lab.github.io/research/meshsuite.
Read the full abstract

All authors Kaixuan Zhang, Minxian Li, Mingwu Ren, Xiatian Zhu

Information Theory

Lifted-Product QLDPC Codes in the Polynomial Domain

This paper presents a finite-length polynomial-domain formulation of lifted-product quantum low-density parity-check (QLDPC) codes. We formulate the code construction over the quotient ring F 2[D]/(D L + 1), where polynomial base matrices are lifted entrywise to binary circulant blocks. This representation gives a compact algebraic description of the lifted-product parity-check matrices and allows CSS orthogonality to be analyzed before binary expansion. We show that the standard circulant lifting map is compatible with polynomial conjugation, which implies that the resulting binary matrices satisfy the CSS commutation constraint. The construction is illustrated with a constraint length 7, rate 1/2 NASA convolutional code example, and numerical examples are provided from a 3 x 4 polynomial parity-check matrix. The finite-length performance of selected constructed codes is then evaluated over the depolarizing channel using various benchmark decoders. The resulting framework gives a structured method to construct finite-length lifted-product QLDPC codes from small polynomial base matrices.
Read the full abstract

All authors Vahid Nourozi, Mohsen Moradi, Roxana Smarandache, David G. M. Mitchell

Rings and Algebras

Deformation theory and the controlling $L_{\infty}$-structure of extended Rota-Baxter algebras

In this paper, we introduce the controlling $L_{\infty}[1]$-algebra of extended Rota-Baxter algebras. As applications, we obtain the cohomology of extended Rota-Baxter algebras and homotopy extended Rota-Baxter algebras. We also study abelian extensions and formal deformations by the lower cohomology group. Lastly, we describe the intrinsic relationship among extended Rota-Baxter algebras, extended Rota-Baxter Lie algebras, and the eigenvalues of the extended Rota-Baxter operator. As the generalization of Rota-Baxter algebras with weight and modified Rota-Baxter algebras, the corresponding results also hold for those algebras.
Read the full abstract

Quantum Physics

Completing or Refusing Low-Dimensional Records of Structured Quantum Circuits: Measurement Loss, Compression Loss, and Hardware Drift

Structured quantum circuits are often represented by low-dimensional records such as occupation numbers or constraint feasibility rather than by full outcome distributions. A record loses control-dependent information when it merges outcomes whose probabilities respond differently to circuit parameters. We assess this loss without assuming a parametric hardware-noise model. Quantum Fisher information bounds premeasurement sensitivity; Fisher metrics of measurements and recorded statistics describe device-attainable sensitivity. For finite control changes, a Hellinger residual measures the response removed by a record, while its local quadratic term is the conditional covariance of the full-outcome score. Simultaneous confidence bounds support approval, refusal, or deferral. We prove a finite-library completion theorem: executable augmentations terminate either with a record preserving every declared response or with proof that no library augmentation removes the loss. For an analytic control germ, integral closures characterize preservation along every analytic control arc, and finitely many Rees valuations detect failure. This connects the state–measurement–record chain to an executable completion-or-refusal procedure, an attainable-Fisher kernel criterion, and finite-sample decisions. A three-qubit calculation separates measurement loss from record loss. Prospectively fixed IBM Kingston and Marrakesh experiments test decisions in four-qubit fixed-particle-number and GHZ families, including negative controls. In two-epoch IQM Garnet data, median record-level drift is 0.0797 times full-distribution drift, but a significant residual remains in 35 of 36 settings. On these circuits, the rules detect loss and approve records only within stated tolerances; they do not establish universal compression performance or device quantum Fisher information.
Read the full abstract

Materials Science

Thickness-Dependent Orbital-to-Spin Torque Signatures in Cr/Gd/Co Thin Films

We studied orbital-torque generation in Cr(10nm)/Gd($t_{\mathrm{Gd}}$)/Co(3nm)/TaO$_x$ and the inverted stack Co(3nm)/Gd($t_{\mathrm{Gd}}$)/Cr(10nm) with $t_{\mathrm{Gd}}$ from 0 to $5\,\mathrm{nm}$ by combining electrical harmonic Hall measurements with magnetometry. A detailed understanding of the magnetometric data is obtained by cross-sectional chemical composition mapping. The data show temperature-dependent magnetic compensation points, while elemental analysis provides evidence of pronounced intermixing, in particular of Gd and Co layers. From the harmonic Hall dataset we extract the damping-like (DL) and field-like (FL) torque efficiencies normalized to the applied electric field, $\xi_{\mathrm{DL}}^{E}$ and $\xi_{\mathrm{FL}}^{E}$, and interpret their dependence on the Gd interlayer thickness using two different descriptions: (i) a naive-layer model and (ii) an alloy model that accounts for interfacial mixing. Notably, upon reversing the stack order, the FL contribution changes sign, whereas the DL contribution does not change sign within the harmonic Hall measurements.
Read the full abstract

All authors Tiago de Oliveira Schneider, Michel Heidkamp, Luana Caron, Inga Ennen, Matthias Opel, Alexey Arzumanov, Lambert Alff, Markus Meinert

Strongly Correlated Electrons

Self-Healing Diffusion Monte Carlo applied to a simple fermionic model: A critical assessment of the method

We investigate the Self-Healing Diffusion Monte Carlo (SHDMC) method using a one-dimensional model with periodic boundary conditions. An inversion symmetry is introduced to mimic the antisymmetry property of fermionic wave functions, with the bosonic and fermionic sectors being modeled by the even and odd eigenstates, respectively. As in realistic fermionic systems, the nodal structure is only partially constrained by symmetry, making this model a non-trivial testbed for nodal optimization algorithms such as SHDMC. We show that the nodal evolution under SHDMC iterations can be cast into a dynamical system exhibiting both attractive and repulsive fixed points. When applying SHDMC to this model in the absence of statistical noise, the fixed-node energy is found to increase during iterations and the node converges to an incorrect value. This occurs for any finite basis, even when the exact node is representable in it; the attractive fixed point approaches the exact node only as the basis becomes complete, while the convergence toward it becomes increasingly slow. We further show that this problem can be largely cured by modifying the nodal update criterion to give more importance to the nodal region. Our results suggest that, in the general case, an efficient nodal optimization may require both an improved nodal update criterion and a localized basis set allowing local distortions of the trial wave function, as is the case for this model.
Read the full abstract

Algebraic Geometry

Virasoro Constraints for Orbifold Curves

We prove the Virasoro constraints for the relative Gromov–Witten theory of all smooth projective effective orbifold curves, allowing relative conditions at ordinary points. The absolute Jiang–Tseng Virasoro conjecture for smooth projective effective orbifold curves follows as a corollary.
Read the full abstract

All authors Jianxun Hu, Pengrong Jin, Hua-Zhong Ke, Jingwei Lu, Hsian-Hua Tseng, Longting Wu

Quantum Physics

Emerging Personas and Tools for Hybrid Quantum-HPC Systems

Quantum computing is entering an era where quantum processing units (QPUs) with tens to hundreds of qubits are increasingly integrated with high-performance computing (HPC) systems to support hybrid quantum-classical workloads. This paradigm, often referred to as Quantum-Centric SuperComputing (QCSC), introduces new operational, software, and infrastructure challenges. In this paper, we identify and characterize six distinct personas involved in QCSC environments. Particular attention is given to emerging roles that arise at the interface between quantum and HPC systems. We analyze the responsibilities and tools associated with each persona and examine their differing requirements for operational data. Consistent data representations and analytics frameworks will be required to support diverse users across the quantum-HPC ecosystem. By defining personas and their data needs, we provide a foundation for future discussions in this critical area.
Read the full abstract

All authors Eun-Kyung Lee, Jessie Yu, Claudio Carvalho, Yoonho Park, Marcelo Amaral, Tim Osborne, Woong Shin

Algebraic Geometry

Relative Fourier–Mukai theory for algebraic spaces

We study the variation of Fourier–Mukai equivalences in proper flat families of Noetherian algebraic spaces. For bounded pseudocoherent kernels, we prove that the equivalence locus is open and commutes with arbitrary Noetherian base change. To do so, we develop a theory of relative perfectness and relative quasi-perfectness for algebraic spaces. This includes a fiberwise criteria for integral transforms with nonperfect kernels. As an application, we construct an explicit autoequivalence for elliptic fibrations.
Read the full abstract

Quantum Physics

A Scalable Multi-Protocol Platform for Quantum Key Distribution Simulation with Rigorous Statistical Evaluation

Quantum Key Distribution (QKD) offers information- theoretically secure key establishment grounded in the laws of quantum physics, yet its practical reach is limited by the prohibitive cost of photonic hardware and the fragmented nature of existing simulation tools. Most simulators support only a single protocol and report results from individual stochastic runs, making systematic protocol comparison and reproducible statistical inference difficult. This paper presents a unified QKD simulation platform that implements four foundational protocols BB84, B92, E91, and BBM92 within a single Python/Qiskit engine. A shared impairment model covers fiber attenuation, source and detector losses, po- larization drift, and configurable intercept-resend eavesdropping. The platform is accessible through two independent interfaces that share the same backend: a desktop application (Tkinter, Matplotlib) for local experimentation and a browser-based web client (React, Node.js/Express) for zero-install remote access. All reported results are drawn from repeated-run studies (20 independent runs, 10000 qubits each), with mean, standard deviation, and 95% confidence intervals stated throughout. At a 25 km fiber link, BB84 achieves the highest mean key-rate of 160,045 Hz, followed by BBM92 (80023 Hz), E91 (52815 Hz), and B92 (40011 Hz) ordering that tracks simulation-derived sifting efficiencies precisely. Under the E91 protocol, the CHSH S-statistic averages 2.12 at baseline and falls to 1.58 when an eavesdropper is activated, demonstrating Bell-inequality-based intrusion detection independent of QBER
Read the full abstract

Classical Analysis and ODEs

Asymptotic safety regions for Gabor frames generated by Hermite functions

The aim of this paper is to establish new regions in the frame sets of Hermite functions $h_n$. A classical result of Gröchenig and Lyubarskii shows that the Gabor system $\mathcal{G}(h_n,a\mathbb{Z}\times b\mathbb{Z})$ forms a frame for $L^2(\mathbb{R})$ whenever the lattice density exceeds $n+1$. We show that, for every $\eta>0$ and all sufficiently large $n$, the same Gabor system forms a frame whenever $ab\leq n^{-\frac{2}{3}-\eta}$. Moreover, we obtain an asymptotically sharp result near the coordinate axes, i.e., when one of the parameters $a$ or $b$ is small. Namely, for every $\delta>0$ and $\rho\in(0,\frac{1}{2})$ and all sufficiently large $n$ we prove that if $\min\{a,b\}\leq n^{-\frac{1}{2}-\delta}$ and $ab\leq \frac{1}{2}-\rho$ then $\mathcal{G}(h_n,a\mathbb{Z}\times b\mathbb{Z})$ forms a frame.
Read the full abstract

Probability

The circular law for non-Hermitian random band matrices: optimal bandwidth, periodic profile and discrete law

We consider non-Hermitian random band matrices with growing bandwidth and study convergence of their empirical spectral distributions to the circular law. Let $N$ denote the matrix size and $W_N$ the bandwidth. From a universality perspective, it is conjectured that the circular law holds whenever $W_N\to\infty$. Previous results have mainly required $W_N\gg N^{1/2}$, with the principal exception of \cite{Han2511}, which proves the $W_N\to\infty$ circular law for an open-boundary block-tridiagonal model. Here we prove the circular law at this optimal threshold for several periodic models and under the near-optimal condition $W_N\gg\log N$ for a genuinely discrete model. For the periodic hard-indicator profile and its uniform and polynomially tapered generalizations, we prove the circular law under bounded-density and finite-third-moment assumptions whenever $W_N\to\infty$. At the same threshold, we prove the circular law for continuous, integrable profiles locally bounded below on every finite interval, including exponentially and Gaussian decaying profiles, with circular complex Gaussian entries. For the periodic full-block model, we prove the circular law under bounded-density and finite-third-moment assumptions whenever $W_N\to\infty$. The finite-third-moment assumption in the bounded-density results can be weakened to a finite $(2+\alpha)$-moment assumption. For the same full-block model without a density assumption, we prove the circular law for centered variance-one real subgaussian atoms when $W_N\gg\log N$. The proof uses compositions of random transfer operators. An established high-band circular-law input on a short auxiliary ring calibrates the full exterior coefficient norm, and a boundary-uniform local comparison lifts this calibration to target rings even when $N/W_N$ is arbitrarily large.
Read the full abstract

Computation and Language

Explore Before Committing: Hypothesis-Guided Search for Deep Research Agents

Deep-research agents answer complex questions by interacting with search and browsing tools, yet they often search along a single evolving trajectory. Our trajectory-level analysis reveals a common failure mode in which the agent may encounter an early search state with several plausible directions, but follow one direction before collecting enough comparative evidence. Once this happens, subsequent tool calls tend to reinforce the same path, increasing the chance of failure when the initial direction is misleading. We further find that successful trajectories reduce this risk through two behaviors: grounding vague exploration in concrete candidates and shifting directions when the current path is weak or incomplete. Based on these findings, we propose HypoSearch, which generates lightweight hypotheses as soft search hints, explores them through bounded independent branches, and compares branch-level evidence before commitment. Across four deep-research benchmarks and three backbone models, HypoSearch consistently outperforms single-trajectory search and standard parallel baselines, improving Qwen3.5-122B from 46.7 to 60.0 on BC-small while using fewer tool calls than five independent trajectories. A pilot supervised fine-tuning study further shows that these behavioral signals can curate compact training trajectories and reduce degradation from unfiltered data.
Read the full abstract

All authors Ruochen Zhou, Zhengyu Chen, Luan Zhang, Siyang Gao, Yee Whye Teh, Shiqi Chen

Networking

TriSLA: A Preventive and Closed-Loop SLA-Aware Architecture for Multidomain Decision-Making with Explainable Artificial Intelligence in 5G Networks

Network slicing in multidomain 5G environments introduces critical challenges in guaranteeing Service Level Agreements (SLAs) under dynamic resource variability and heterogeneous service requirements. This article presents TriSLA, a closed-loop, preventive, SLA-aware architecture designed to evaluate feasibility at request time and continuously ensure SLA compliance during operation. The architecture combines ontology-driven semantic intent interpretation, multidomain machine learning feasibility risk inference, Explainable Artificial Intelligence (XAI) feature attribution, and closed-loop runtime SLA assurance into a unified operational pipeline. A fully operational prototype was evaluated in a multi-node cloud-native environment integrating Radio Access Network (RAN), Transport Network (TN), and 5G Core (5GC) domains with real-time telemetry collection. Experimental evaluation demonstrates that TriSLA guarantees a 100% SLA satisfaction rate for admitted slices, completely eliminating post-deployment violations compared to reactive (51.2%) and static threshold (80.4%) admission baselines. The predictive feasibility assessment achieved a classification accuracy of up to 99.51% (98.68% for the default explainable Random Forest classifier), enabling preventive admission decisions before infrastructure commitment. Furthermore, the cognitive admission pipeline introduces minimal processing overhead, requiring 25.37 ms for ontology-driven semantic parsing and 231.66 ms for XAI-assisted feasibility inference. Concurrently, the closed-loop assurance engine resolves 100% of runtime telemetry anomalies within a 4.22 s recovery cycle. These results demonstrate that TriSLA provides reliable, explainable, transparent, and preventive SLA management through integrated predictive admission and closed-loop runtime assurance for next-generation 5G networks.
Read the full abstract

Programming Languages

Relational Task Generation Language: A Declarative Specification Framework for Relational Deep Learning

Relational Deep Learning (RDL) has become a powerful paradigm for learning from multi-tabular data. However, manually defining RDL prediction tasks is a laborious process that frequently results in data leakage. To address this issue, we introduce Relational Task Generation Language (RTGL) - an open-source declarative language that streamlines RDL task formulation by abstracting away low-level SQL details. We showcase RTGL by reconstructing existing RDL benchmark tasks and uncovering their inconsistencies stemming from manually crafted SQL definitions of RDL prediction targets, thereby underscoring the value of a dedicated declarative language. In addition, we demonstrate the practical utility of RTGL by designing various new tasks with diverse forms and target types. Our experiments confirm the robustness and usability of RTGL, as well as its seamless integration with the existing RDL frameworks, making it widely accessible to the community.
Read the full abstract

Optimization and Control

Optimal Inflow Control for Transport Equations with Uncertain Velocities and Demand

We study optimal inflow control for a linear transport equation subject to uncertainty in both time-dependent downstream demand and transport velocity. The random velocity induces a random travel time and thereby changes the relation between an inflow decision and the demand observed at its arrival time. For a fixed downstream observation window, we derive explicit optimal controls in the continuous-time setting and for piecewise constant controls. We decompose the irreducible stochastic error into contributions from demand and velocity uncertainty and show that the latter admits a bound that is linear in the velocity variance. As a computationally attractive alternative, we analyze a deterministic mean-velocity proxy whose additional performance loss admits a higher-order bound. We also establish Lipschitz stability of the optimal control with respect to perturbations of the velocity law in the Wasserstein distance. Numerical experiments illustrate the analytical results and investigate the proxy strategy for an uncertain nonlocal transport model beyond the linear theory.
Read the full abstract

General Relativity and Quantum Cosmology

Deviations from Kerr: Polar critical curves and photon rings in a class of separable spacetimes

Black-hole critical curves and photon rings provide complementary probes of strong-field geometry, motivating a unified analytic description of these observables beyond Kerr and a direct comparison with the Kerr predictions. We develop such a description for polar observers in a radial class of stationary, axisymmetric spacetimes with separable null geodesics. The spacetime dependence is organized through natural optical combinations selected by the null dynamics and a Kerr-relative representation suited to weak deviations from Kerr. We derive compact exact expressions for the polar critical radius, Lyapunov exponent, time delay, and rotation parameter, together with general first-order formulas that can be applied once the deformation functions of a specific metric are supplied. The weak-deviation analysis shows that the combined polar observables probe only a finite set of local deformation values and radial derivatives at the critical orbit, rather than the full radial profiles.
Read the full abstract

Computer Vision

Agentic Multimodal Models for Environmental Hyperspectral Unmixing

Hyperspectral unmixing is a key task in remote sensing that aims to decompose mixed pixels in hyperspectral images into their constituent material signatures, or endmembers, and their fractional abundances. Conventional modular approaches estimate the scene composition through successive model-order estimation, endmember extraction, and abundance estimation stages, whose errors can lead to redundant or ambiguous candidate components and ultimately affect the recovered decomposition. We introduce an algorithm-agnostic, large vision-language model (LVLM)-driven agentic framework that refines the outputs of such pipelines rather than replacing their underlying numerical algorithms. Starting from an initial decomposition, the agent iteratively gathers complementary spectral and spatial evidence through dedicated tools, including spectral-library retrieval and abundance-map visualization, and modifies the active endmember set through merge and discard operations followed by abundance re-estimation. We apply the same refinement procedure to several modular pipelines combining different model-order, extraction, and abundance-estimation methods, and evaluate it on HYDICE Urban, Jasper Ridge, and Stonewall Playa. Experiments show that the proposed agent consistently improves endmember cardinality and generally improves the recovered spectral signatures and abundance maps across heterogeneous modular pipelines, while remaining competitive with integrated end-to-end unmixing methods, including CNN-AE, uDAS, and R-CoNMF. These results highlight the potential of tool-using LVLM agents to combine spectral and spatial evidence for algorithm-agnostic refinement of physically grounded hyperspectral unmixing decompositions. Code is publicly available at https://anonymous.4open.science/r/agentic-hu.
Read the full abstract

All authors Michał Cholewa, Luca Ciampi, Nicola Messina, Przemysław Głomb, Giuseppe Amato

Numerical Analysis

Accelerating the Improved Arrow–Hurwicz Iteration via the Anderson Algorithm for Steady-State Navier–Stokes Equations

We apply Anderson acceleration to the improved Arrow–Hurwicz (IAH) method for the finite element solution of the steady-state incompressible Navier–Stokes equations. The IAH scheme avoids saddle-point solves by decoupling the velocity and pressure updates, but can require prohibitively many iterations, particularly at high Reynolds numbers. To place the acceleration on a rigorous footing, we reformulate the IAH iteration as a nonlinear fixed-point operator $G$ for the grad-div augmented discrete formulation induced by the scheme and establish its well-definedness, Lipschitz continuity, and Fréchet differentiability, thereby verifying the required smoothness conditions locally near the fixed point. Numerical experiments on problems with known analytical solutions, lid-driven cavity flow up to $Re = 15{,}000$, and channel flow over a full step demonstrate that the resulting Anderson-accelerated improved Arrow–Hurwicz algorithm substantially reduces iteration counts and CPU time while retaining the reported manufactured-solution convergence rates and centerline-velocity agreement.
Read the full abstract

Sound

Soft Posterior Speaker Injection for Multi-Talker Speech Recognition

Multi-talker automatic speech recognition (MT-ASR) remains challenging in the presence of overlapping speech. Hard segmentation introduces irreversible errors, whereas serialized output training (SOT) avoids explicit segmentation but does not condition a pretrained encoder on speaker activity. We propose Soft Posterior Speaker Injection (SPSI). A Soft Posterior Head predicts per-frame speaker posteriors $\hat{\mathbf{P}}$ and injects them into Whisper through Multi-layer Feature-wise Linear Modulation (MFLM) and Speaker Memory Prompts (SMP). The benefit of SPSI is largest where overlap is heaviest and under domain transfer. On controlled two-speaker LibriSpeech overlap, SPSI reduces concatenated minimum-permutation word error rate (cpWER) from $61.5\%$ to $60.0\%$ in the high-overlap bin, and from $51.9\%$ to $51.0\%$ on the full set, relative to SOT. By contrast, Speaker CE, SD-CTC, SA-DiCoW, and Pipeline (oracle/est.\ VAD) do not outperform SOT. Freeze-posterior overlap-heavy adaptation reduces held-out LibriCSS cpWER from $42.3\%$ to $36.8\%$ on sessions $8$–$9$, a $5.5$-point gain over SOT. The source code is available at https://github.com/HackerHyper/SPSI.
Read the full abstract

All authors Jian Zhu, Jun Sun, Jiang Yang, Ying Zhou, Cheng Luo, Yang Ai, Hong-Hao Sun, Junhui Shi, Li-Rong Dai

Artificial Intelligence

Analog-DB: An Agent-First Analog Integrated Circuit Database, From Blocks to Systems

Sharing analog integrated circuit designs remains difficult: foundry non-disclosure agreements restrict the process details a design depends on, and the testbenches behind published results are rarely released. We present analog-db, an open-source, versioned database built on a shareable design representation. A domain-specific language captures each design as a process-neutral topology, reusable testbenches, and a machine-readable datasheet under one schema, so a design is shared in full and re-simulates on the process kits it is bound to. A parameterization scheme exposes functional sub-blocks and device sizes as named parameters that carry their matching constraints, making circuits composable and retargetable; a schema-governed contract and queryable catalog let AI design agents discover and reuse them directly. Across the regulator corpus, all 23 circuit-kit bindings on three open kits meet their own recorded specification bands (typical corner, matched devices, no layout) and 10 of 23 meet a common class band. Seventeen of the 23 imported sizings failed their testbenches and closed under a gm/ID sizing loop driven by the annotated sub-block roles, typically within one to three iterations. In a supervised case study, a coding agent working from the released artifacts sized the op-amp cores of a chopper instrumentation amplifier on an open 130nm kit, locating four hand-entry defects and a missing common-mode feedback loop that the sizing-only baseline did not repair. The database holds 68 circuits across sixteen classes, verifiable at schematic level under a tiered harness and tracked on a power/performance scoreboard, released at https://github.com/MacAnalog/spicexplorer-release.
Read the full abstract

High Energy Physics (Phenomenology)

Addendum: Production of $\pi^{+} \pi^{-}$ pairs in diffractive photon-proton and in proton-proton collisions revisited, in particular concerning the Drell-Söding contribution

We update our analyses concerning the non-resonant (Drell-Söding) contribution to the reaction $\gamma^{(*)} p \to \pi^{+} \pi^{-} p$, now including general vertex functions for the coupling of pomeron, $f_{2 \mathbb{R}}$, and $\rho_{\mathbb{R}}$, to pions. This takes into account the off-shellness of the pions. We present a comparison of our tensor-pomeron model, including our new Drell-Söding and $\rho$, $\omega$ vector-meson contributions, with data for the $\gamma p \to \pi^{+} \pi^{-} p$ reaction measured by the ZEUS and H1 Collaborations. We find very good agreement of our theory with the experimental data for the $\pi^{+} \pi^{-}$ invariant mass ($M_{\pi \pi}$) and for angular distributions. Here we restrict our analysis to the region $M_{\pi \pi} \lesssim 1.2$ GeV. For higher values of $M_{\pi \pi}$ we will have to consider production of higher-mass mesons. The purpose of the present article is to show that our tensor-pomeron model gives a very good fit to the so-called skewing of the $\rho$-resonance contribution in photoproduction. In fact, it is an old problem/puzzle to understood the difference of the shapes of the $M_{\pi \pi}$ distributions in the region of the $\rho(770)$ resonance in $e^{+} e^{-}$ annihilation and photoproduction. We think that with our results we have obtained for this problem a solution which is very satisfactory from the Quantum-Field-Theory point of view. Our results are also relevant for central-exclusive $\pi^{+} \pi^{-}$ production in ultraperipheral $pp$, A$p$, and AA collisions at the LHC and for photo- and electroproduction of $\pi^{+} \pi^{-}$ at future electron-proton/ion colliders, EIC and LHeC.
Read the full abstract

Metric Geometry

An exact hierarchy for Lebesgue's universal covering constant and a certified 0.834 lower bound

Posed by Lebesgue in 1914, the universal covering problem asks for the smallest-area planar convex set containing a congruent copy of every set of diameter at most one. We introduce an exact Reuleaux-type variational hierarchy for this constant: its monotone finite-arc values $\Lambda_M$ satisfy $a_{\mathrm{Leb}}=\lim_{M\to\infty}\Lambda_M$, and each level is a continuous finite-dimensional problem. We prove $0\le a_{\mathrm{Leb}}-\Lambda_M\le C M^{-2}$, giving a controlled finite-arc route to the constant itself. As a certified low-order realization, an outward-rounded interval certificate for a regular finite Reuleaux subtest proves $a_{\mathrm{Leb}}\ge0.834$, improving the lower-bound benchmark established by Brass and Sharifi in 2005.
Read the full abstract