Peer Revenue Play credits. Real papers.

Papers

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

All of arXiv

Tuesday 8 September

Computer Vision

Vision-language models see more about agriculture than they report, and test-time scaling remains bounded by parametric knowledge

Vision-language models (VLMs) show promise for agricultural diagnosis but fail unpredictably in zero-shot settings. We test whether this stems from poor visual features or an inability to verbalize what the model already perceives, using a benchmark spanning classification and detection drawn from 118 AgML classification and 54 detection datasets. Linear probing shows agricultural features are linearly separable in frozen VLM vision towers, with Gemma-4-31B's encoder reaching F1-macro of 0.60–0.78 across tasks and approaching the vision-only DINOv3 backbone, indicating the bottleneck lies downstream of perception. We then test whether test-time scaling, an iterative in-context skill-memory loop, can close this gap without updating weights. Gains are modest and inconsistent. Skill-memory improves classification accuracy by up to 3.1 points on three of four tasks while regressing on species, with gains concentrated on classes seen during optimization and unseen classes regressing. Detection F1 is effectively unchanged (0.412 to 0.421), though an oracle filtering condition reaches 0.432 F1, revealing real headroom unconstrained prompting fails to surface. These results indicate VLMs see more agricultural features than they report, but test-time scaling alone fails to surface that hidden parametric knowledge, motivating agentic approaches grounded in external, retrievable knowledge.
Read the full abstract

All authors Earl Ranario, Jared Smith, Lars Lundqvist, Urmil Jatin Chandarana, Bharath Sridhar, J Mason Earles

Differential Geometry

Area-constrained Willmore surfaces in Lorentzian manifolds: curvature inequalities and the Hawking quasi-local energy

We develop an area-constrained Willmore theory for closed spacelike surfaces in four-dimensional Lorentzian manifolds. Since the Euler–Lagrange equation is normal-bundle-valued, we introduce directional criticality along a prescribed normal direction and identify a class of connection-compatible directions determined by the normal connection. Our main result is the sharp inequality $2\lambda|\Sigma|+\int_\Sigma |\vec H|^2,d\mu\leq 8\pi\chi(\Sigma)$ for criticality in a connection-compatible spacelike direction under an Einstein-tensor sign condition. Under the dominant energy condition and $\lambda\geq0$, this yields $\int_\Sigma |\vec H|^2,d\mu\leq16\pi$. Equality is rigid: $\Sigma$ is round and satisfies Minkowski causal-diamond rigidity. We also prove an analogous sharp inequality for null critical directions. These results give positivity, rigidity, stationarity, and infinitesimal monotonicity statements for the Hawking quasi-local energy. A hypersurface formulation recovers the previously studied Hawking-surface problem as criticality in the spatial normal direction. Explicit examples in spherical symmetry, Kerr, and FLRW spacetimes distinguish directional from full criticality and show the necessity of the sign assumptions in the rigidity theory.
Read the full abstract

Instrumentation and Detectors

Intrinsic time resolution of a hexagonal BC404 scintillator: A Monte Carlo study of interaction geometry and optical reflectivity

Time resolution is an important parameter in particle detectors, as it determines their ability to distinguish between two temporally separated events. It can be decomposed into electronic and intrinsic contributions, the latter being influenced by scintillation, optical transport, detector geometry, particle type and energy, and the sensitive area of the photosensor. In this work, a Monte Carlo simulation based on Geant4 is used to investigate the intrinsic time resolution of a hexagonal BC404 plastic scintillator. The analysis includes the evaluation of the optical-photon distribution over the hexagonal surfaces and the characterization of the intrinsic time resolution for different interaction configurations. The intrinsic time resolution is investigated using muons as cosmic-ray proxies with energies ranging from 4~GeV to 1~TeV. In addition, the effect of surface reflectivity on optical-photon transport and timing performance is studied for reflectivities ranging from 0% to 100%. The results show that the intrinsic time resolution exhibits little dependence on the incident muon energy, while a clear dependence on the interaction position is observed. Furthermore, although the number of detected photons varies with the interaction position and surface reflectivity, the intrinsic time resolution is not solely determined by photon statistics, indicating that optical transport and geometrical effects play an important role in the timing performance. Across the investigated configurations, intrinsic time resolutions ranging from approximately 15 to 100~ps are obtained. These findings highlight the importance of the interaction position, optical transport, photon statistics, and surface reflectivity in determining the timing performance of hexagonal plastic scintillators and provide guidance for the optimization of their optical properties and geometry for timing-sensitive applications.
Read the full abstract

Combinatorics

Existence of regular sets in circulant quartic graphs: a complete characterization

Let $\Gamma=(V,E)$ be a graph and let $a,b$ be nonnegative integers. A nonempty proper subset $C\subset V$ is an \emph{$(a,b)$-regular set} if each vertex of $C$ has exactly $a$ neighbors in $C$ and each vertex outside $C$ has exactly $b$ neighbors in $C$. Perfect codes and total perfect codes are the $(0,1)$- and $(1,1)$-regular sets. For every connected circulant quartic graph $\operatorname{Cay}(\mathbb{Z}_n,\{x^{\pm k},x^{\pm\ell}\})$ and every pair $(a,b)$ we decide, in terms of $n$, $k$ and $\ell$, whether an $(a,b)$-regular set exists. Existence is proved by exhibiting unions of residue classes; non-existence comes from a Fourier-analytic obstruction combined with the fact that certain sums of roots of unity cannot vanish.
Read the full abstract

All authors A. Abdollahi, J.Bagherian, F. Jafari, M. Khatami, Z. Shokoohi, R. Sobhani

Artificial Intelligence

Decision-Focused Active Learning for Scale-Aware Critical-Materials Recovery

Choosing a recovery process for scale-up requires connecting laboratory results with product requirements, process costs, and scale effects. We analyze records from Pacific Northwest National Laboratory's Computer Intelligence for Critical Element Recovery and Optimization (CICERO) workflow for autonomous selective precipitation. Active learning uses prior results to choose experiments. In a conditional retrospective benchmark with fitted models and recycled neodymium-iron-boron (NdFeB) magnet records, active learning finds the best recorded result with fewer experiments than nonadaptive space filling. Enrichment is the selected rare-earth-to-iron ratio relative to that in the feed. Adaptive policies reach the recorded enrichment maximum by 16 to 24 wells (individual experiments), versus 48. Our two-stage reconstruction ties two adaptive alternatives at 16 wells. Conditional analyses of recycled samarium-cobalt (SmCo) magnets show a Round 2 tradeoff between purity and nominal yield, the recovery fraction calculated from an assumed starting amount - NdFeB Round 1 routes differ in enrichment. Rankings for produced water from oil and gas extraction depend on phase and dilution assumptions requiring confirmation. We propose choosing batches by their expected reduction in downstream Bayes risk: the minimum expected loss among available process decisions under current beliefs. In exploratory simulations, a hybrid that filters candidates has lower estimated loss than the implemented joint search across routes and conditions. Differences involving the synthetic two-stage policy are small relative to estimation uncertainty. We outline a pre-registered prospective test under a shared loss and logging standard, requiring clarified measurements and records, a defined process decision and relevant outputs, credible economic inputs, and validation at the intended scale.
Read the full abstract

Nuclear Experiment

Detailed experimental study of excited states in $^{50}$Ti via the $(d,p)$ and $(d,p\gamma)$ reactions

Excited states of semi-magic $^{50}$Ti were studied up to the neutron-separation energy via the $(d,p)$ and $(d,p\gamma)$ reactions. In total, 82 excited states were identified based on the measurement of angular distributions with the Super-Enge Split-Pole Spectrograph (SE-SPS) at Florida State University. From the experimental data, sum rules related to vacancies were calculated for the $2p_{3/2}$, $2p_{1/2}$, $1f_{5/2}$, $1g_{9/2}$, and $2d_{5/2}$ neutron single-particle orbitals and compared to predictions obtained with the time-dependent continuum shell model (TDCSM), the quasiparticle-phonon model (QPM), and the relativistic equation of motion theory (REOM$^3$). A comparison for the $(d,p)$ data obtained for $^{51}$Ti and $^{50}$Ti is also presented, focusing on differences of the single-particle strength fragmentation in even-$A$ and odd-$A$ nuclei.
Read the full abstract

All authors B. Kelly, M. Spieker, L. T. Baby, S. Baker, A. L. Conley, D. Houlihan, K. W. Kemper, E. Litvinova, N. Tsoneva, A. Volya

Astrophysics of Galaxies

TDCOSMO XXX: Spatially resolved kinematics of the deflectors in time-delay lens systems B1608+656 and SDSSJ1206+4332 from JWST-NIRSpec observation

We present spatially resolved stellar kinematics of the deflector galaxies in two time-delay lens systems, B1608+656 and SDSS J1206+4332, measured from JWST NIRSpec integral field spectroscopy. B1608+656 is a quadruply imaged quasar lensed by a pair of tidally interacting galaxies, the primary deflector (G1) and a close companion (G2), for which we quantified G2's tidal influence on G1 and define a criterion identifying the region where this perturbation is negligible and standard Jeans-equation modeling remains valid. SDSS J1206+4332 is a doubly imaged quasar whose extended host galaxy is itself quadruply imaged by a foreground massive galaxy (G0), accompanied by a fainter perturber (G1). For both systems, we report the first spectroscopic redshift and velocity dispersion measurements of the perturber galaxies. For G2 in B1608+656, we measure {\sigma}_v = 188 +/- 7 (stat.) +/- 2 (sys.) km/s at z = 0.6307, identical to the redshift of the main deflector G1. For G1 in SDSS J1206+4332, we find {\sigma}_v = 100 +/- 21 (stat.) +/- 2 (sys.) km/s at z = 0.7403, slightly lower than the primary deflector G0. The resolved velocity maps of both primary deflectors show clear rotation signatures and we visually classify them as fast or regular rotators. The data products will be made publicly available and will be combined with time delays, lens models, and line-of-sight convergence to measure cosmological parameters in the upcoming TDCOSMO 2026 milestone publication.
Read the full abstract

All authors Pritom Mozumdar, Shawn Knabel, Anowar J. Shajib, Tommaso Treu, Michele Cappellari, Aymeric Galan

Computation and Language

Benchmarking Hybrid Deep Research Across Database Querying and Web Search

While autonomous agents have made significant strides in "deep research" by iteratively navigating the open web to synthesize information, real-world problem-solving is rarely confined to a single environment. Complex analytical tasks inherently require agents to weave together evidence from both ambiguous unstructured text (e.g., the open web) and highly precise structured data (e.g., relational databases). However, existing benchmarks evaluate these modalities in isolation, failing to capture the critical "handoff" - the ability to preserve constraints when moving evidence between systems. We introduce HybridDeepResearch, to our knowledge the first deep-research benchmark that requires both web search and SQL to form a complete, verifiable answer. The benchmark contains 380 tool-dependent tasks grounded in LiveSQLBench-Base-Lite databases and public web corpora, validated through automated checks and human review, and covering three reasoning patterns: SQL2S, S2SQL, and Parallel. Evaluations across proprietary and open-weight models under various agentic scaffolds reveal that even state-of-the-art models like GLM-5.2, Claude-Sonnet-4.6 and GPT-5 achieve only about 50-54% Pass@8 on the hard subset. Notably, results show that directional reasoning is substantially more difficult than parallel intersection, highlighting that bridging structured and unstructured information spaces without losing constraints remains a major open challenge for agentic systems. Code and datasets are publicly available at GitHub (https://github.com/Snowflake-AI-Research/HybridDeepResearch) and Hugging Face (https://huggingface.co/datasets/Snowflake/HybridDeepResearch).
Read the full abstract

All authors Ruofan Wu, Peiran Xu, Xiaolong Li, Fan Shu, Soyoung Yoon, Yite Wang, Xiaodong Yu, Boyi Liu, Feng Yan, Debiao Li, Yuxiong He, Zhewei Yao

Signal Processing

Reliable Near-Field Multi-User Positioning Informed by Two-Stage MUSIC

Near-field localization is a promising technique for high-resolution multi-user positioning in future wireless systems, but its performance is often degraded by scattering-induced coherent propagation. Existing near-field localization methods, which require separate parameter estimation and path/source association, suffer from high computation overhead and accumulated errors, and usually do not provide any guarantee on reliability. In this paper, we propose MUSIC-Net, an end-to-end near-field positioning deep learning (DL) framework informed by two-stage MUltiple SIgnal Classification (MUSIC) in mixed line-of-sight (LoS) and non-LoS (NLoS) multi-path scenarios, which embeds the two-stage MUSIC objects into training to isolate the LoS-related signal subspace and to identify a surrogate distance. The proposed framework directly recovers multi-user positions without the need for involved NLoS parameter estimation or path/source association. Furthermore, we introduce split conformal prediction (SCP) to move beyond point-estimation-based positioning towards statistically guaranteed (confidence) set estimation for all users. Numerical results show that the proposed MUSIC-Net achieves lower mean positioning error (MPER) than existing benchmarks and yields tighter SCP-calibrated prediction regions, demonstrating both accurate LoS localization and efficient uncertainty quantification (UQ) in coherent multi-path environments.
Read the full abstract

All authors Jiaying Li, Haifeng Wen, Changsheng You, Yuanwei Liu, Hong Xing

Plasma Physics

RF-Specific Tungsten Erosion and Global Transport in ITER under Neon Seeding

Ion cyclotron radio-frequency heating (ICRH) is a key auxiliary heating system in ITER, but high-power RF operation can enhance plasma-material interactions through rectified RF sheath potentials on antenna structures and nearby plasma-facing components. We present the first predictive application of the STRIPE (Simulated Transport of RF Impurity Production and Emission) framework to assess RF sheath-driven tungsten (W) erosion and global impurity transport from the ITER ICRH antenna under ITER-relevant neon-seeded conditions. STRIPE couples SOLPS-ITER plasma backgrounds, full-wave RF sheath calculations, geometry-specific ion energy-angle distributions, sputtering physics, and three-dimensional impurity transport. Simulations predict RF sheath potentials of 1 to 3 kV on antenna limiter sidewalls, increasing gross W erosion by about a factor of 64 relative to thermal sheath conditions and producing a gross source of 3.34e18 W atoms per second. Erosion is governed by RF-modified ion energy-angle distributions together with local plasma flux rather than sheath voltage alone. About 10 percent of sputtered W is locally redeposited, giving a net source of 3.01e18 W atoms per second. The RF-induced antenna source remains about three orders of magnitude smaller than the thermal divertor source and more than two orders of magnitude smaller than the integrated thermal main-chamber source. After 100 ms, about 22 percent of the mobile W inventory resides within the SOLPS-covered confined-plasma region, corresponding to an annular W concentration of 1.70e-6. These results indicate that the ITER ICRH antenna is unlikely to dominate the total W source budget under the conditions considered and demonstrate the need for coupled modeling of RF waves, sheaths, sputtering, redeposition, and global impurity transport.
Read the full abstract

All authors Atul Kumar, Dhyanjyoti Nath, Wouter Tierens, Jeremy D. Lore, Andrei Pshenov, Tom Wauters, Andrea Galvan, Davide Curreli, Syun'ichi Shiraiwa, Nicola Bertelli, Guillaume Urbanczyk, Sebastijan Brezinsek, Onkar Sahni, Mark Shephard, Walid Helou, Xavier Bonnin, Klaus Schmid, Volodymyr Bobkov

Quantum Physics

Balancing Expressivity and Overfitting in Quantum Gaussian Process Regression

Active learning is a paradigm of machine learning that can be utilized to model expensive black-box functions by training a surrogate model from actively queried training points. The performance of this framework depends heavily on the choice of the surrogate model. When the surrogate is Gaussian Process Regression (GPR), its performance is largely determined by the expressivity of the underlying kernel. In this work, we investigate the peculiarities of using a quantum kernel to change the computational dynamics of active learning with GPR, focusing on the sensitivity to regularization by hyperparameter tuning. While generic, unstructured kernels suffer from exponential concentration at large scale, we empirically demonstrate that even at small scale overfitting can collapse GPR performance. Kernel regularization can be used to counteract this effect, but due to the smoothness of the quantum fidelity kernel, regularization must be carefully chosen to balance expressivity and overfitting. Our qualitative results transfer to the practical application of restricted quantum kernels designed to avoid exponential concentration and also present the kinds of noise that may be valuable to kernel training in near-term quantum devices.
Read the full abstract

Dynamical Systems

Finite Cores Exhaustion and the Failure of Strict Sequential-Pressure Approximation in Coded Shifts

Let $X=X(\mathcal{G})$ be a coded shift whose generating set uniquely represents its concatenation set. We prove that, for every continuous potential $\varphi$, the pressures of the finite-generator subshifts (the finite cores) converge to the sequential pressure $P_{\mathrm{seq}}(\varphi,\mathcal{G})$, the supremum of free energy over invariant measures giving full mass to the concatenation set. Thus full sequential pressure is the exact condition for the finite cores to recover global pressure. The proof is a refinement of the inducing argument of Burr, Das, Wolf, and Yang. Given an arbitrary enumeration of the generators and a length-ordered enumeration of the language, we obtain an algorithm which computes the global pressure of every computable potential with full sequential pressure. The set of equilibrium states is recursively compact, and a unique equilibrium state is computable from the same data. For the zero potential, $h_{\mathrm{con}}\geq h_{\mathrm{res}}$ therefore implies computability of topological entropy and of any unique measure of maximal entropy. We also construct a uniquely represented presentation of the full binary shift for which the zero potential has full sequential pressure but is not a uniform limit of potentials satisfying $P_{\mathrm{seq}}>P_{\mathrm{res}}$, disproving a conjecture of Burr, Das, Wolf, and Yang.
Read the full abstract

Statistical Finance

The Log S-fBM model: Statistical analysis

The Log S-fBM model, introduced by Wu et al., is a stochastic volatility model whose log volatility is a stationary fractional Brownian motion (S-fBM): a stationary Gaussian process with power-decaying autocovariance driven by the Hurst exponent $H$, and variance scaled by an intermittency coefficient. A key property is that it reconciles rough volatility, where $H$ is typically near $0.1$ (see Gatheral et al.), with multifractal volatility, where $H$ is close to $0$ as in Bacry, Muzy et al.: the model's volatility measure converges to a multifractal random measure as $H\to0$. Numerical findings in Wu et al. show intermittency of order $0.02$ across financial assets, motivating a small intermittency approximation of log volatility moments for calibration via the general method of moments (GMM). In this work, we conduct a statistical analysis of the Log S-fBM model. We derive scaling properties of the S-fBM process and the Log S-fBM integrated volatility measure, present deviation inequalities with tail distributions sensitive to $H$ and intermittency, and develop a hypothesis test for the null Hurst exponent, i.e.\ rough versus multifractal dynamics. Finally, we revisit scale invariance of the log volatility increment process via explicit small-intermittency formulas, reproducing analogous properties in both regimes.
Read the full abstract

Cryptography and Security

An Experimental Evaluation of Multimodal Prompt Injection Attacks on Agentic AI Frameworks

Agentic AI frameworks let a language model plan, keep memory, and call tools that reach real files, mail, and services. Most of these agents also read images, which gives an attacker a way to put text into the agent's context without going through the user. We present MMPIBench, a reproducible benchmark that measures what happens next. It delivers a fixed set of attacks through six visual carriers (OCR text, overlays, EXIF metadata, QR codes, fake interfaces, and hybrids) and records how far each injected instruction travels through the agent, from perception through planning to the tool call. Across 720 runs covering six frameworks, five foundation models, six carriers, and four attacker objectives, attacks complete in approximately 1% of runs but are attempted in 12.8%, and the gap is closed almost entirely at the planning step, where the model reads the injected instruction and declines to act on it. The model matters far more than the framework for whether an instruction is acted on. One model never attempts an attack and recognizes the injection in 59.7% of runs, while two others attempt in 23.6%. We then extend the benchmark to audio, the only other raw perceptual channel current frontier models accept. Only two of the five models ingest audio and only three of the six frameworks deliver it, but where the signal arrives the attack completes in 49% of cells, and in 75% for one model. Reporting completion alone therefore understates exposure, and perceptual channels beyond vision are narrower but much less defended.
Read the full abstract

Robotics

A Decade of Bayesian Optimization for Controller Tuning and Robot Learning: Tutorial, Review, and Future Prospects

In the past decade, Bayesian optimization (BO) has emerged as a powerful and adaptable framework for automatic controller tuning and robot learning. This article offers a comprehensive overview of the state-of-the-art in BO, designed to support both researchers and practitioners in understanding recent advancements, practical applications, and future research directions. We begin by adopting a practitioner's perspective, illustrating how to effectively set up BO through a representative controller tuning example. We position BO within the broader context of learning paradigms, ranging from deep reinforcement learning to data-driven control, and highlight scenarios where BO is most advantageous. Next, we discuss the diverse range of BO methods that have been developed to tackle complex problems and specific applications. This article provides a unified perspective on the current landscape of BO, emphasizing its relevance to control systems and robotics, and it highlights future prospects by identifying key research challenges and promising avenues for advancing BO in the field. This includes addressing a significant gap in the BO landscape: the lack of standardized benchmark problems specifically for control-related applications. To foster future research and ensure rigorous evaluation, we start an effort towards a lightweight benchmark suite for control engineering and robotics. We also present metrics and best practices to facilitate direct comparisons between new BO algorithms and established state-of-the-art methods.
Read the full abstract

All authors David Stenger, Paul Brunzema, Johanna Menn, Alexander von Rohr, Angela P. Schoellig, Sebastian Trimpe

Cryptography and Security

X-amine509: Predicting the Practical Risk Level of Enterprise X.509 Certificates

Enterprises managing large X.509 certificate inventories face a prioritization problem: deterministic analysis tools that precisely identify standards violations are indispensable for remediation, but applying them exhaustively across millions of certificates is operationally impractical. We present X-amine509, a two-stage triage system that uses machine learning to rapidly rank certificates by predicted risk and route only the highest-risk items to full deterministic analysis. Certificate risk is quantified as a composite score derived from 177 defect checks grounded in CA/Browser Forum Baseline Requirements, NIST IR 8547/SP 800-57, and cryptographic strength criteria, weighted by security severity across four tiers ranging from cryptographic breaks to minor compliance deviations. We collected 1,027,714 X.509 certificates from Fortune 500, .gov, and .edu domains and scored each using this rubric. On a held-out test set of 201,976 certificates, our best model (Extra Trees) achieves $R^2$ of 0.993 with MAE of 2.26, while Decision Tree scores $R^2$ of 0.986 at 3.7 million certificates per second on a single machine. Ranking quality confirms the triage value: aggregate NDCG exceeds 0.997, and severity-tier classification reports 99.76% accuracy with 98.90% recall on critical-tier defects. Thirteen months later, we retrieved another 571,374 certificates to test our models' durability over time, and the Extra Trees and Decision Tree models maintain MAE below 6.8, $R^2$ of at least 0.915, aggregate NDCG above 0.988, severity-tier accuracy of at least 99.52%, and critical-tier recall of at least 97.03%. Feature importance analysis identifies validity period, Extended Key Usage configuration, negative serial number encoding, and self-signed status as the strongest risk predictors, providing coarse interpretability at the triage stage.
Read the full abstract

All authors Cameron Keith, Shubh Patel, JD Kilgallin, Caleb Shorter

Statistical Methodology

Semiparametric Receiver Operating Characteristic Analysis in the Presence of an Imperfect Reference Standard via a Box-Cox Density Ratio Model

Receiver operating characteristic (ROC) analysis is commonly used to evaluate the diagnostic accuracy of continuous biomarkers. In practice, the true disease status may be unavailable and only a nominal disease status provided by an imperfect reference standard is observed. Existing nonparametric methods have been developed for ROC analysis in this setting, but may suffer from reduced estimation efficiency, numerical instability, or sensitivity to the choice of biomarker scale. We propose a semiparametric method based on a Box-Cox density ratio model, which links the biomarker distributions of the truly healthy and diseased populations while leaving the baseline distribution unspecified. A key feature of the proposed method is that the transformation parameter is estimated from the data rather than specified in advance, allowing the density-ratio structure to adapt to different transformation scales. We develop an empirical likelihood approach for estimation and an expectation-maximization algorithm for computation. We establish the asymptotic distributions of estimators of the ROC curve, area under the curve, Youden's index, and the sensitivity and specificity at the Youden-optimal cutoff, and develop bootstrap confidence intervals and a goodness-of-fit test. Simulation studies demonstrate that the proposed method provides accurate and numerically stable estimation and inference across a range of distributional settings without requiring the transformation scale to be specified in advance. The proposed method is illustrated using data from a malaria study.
Read the full abstract

All authors Yi Chang, Siyan Liu, Qinglong Tian, Pengfei Li

Networking

QPS-ToR: A Parallel Iterative Switching Algorithm for Reconfigurable Optical Datacenter Switching

Reconfigurable optical data center networks (RODCNs) have emerged as a promising solution for scaling DCN capacity, yet their scheduling mechanisms remain a performance bottleneck: traffic-oblivious schemes inherently limit throughput, while the state-of-the-art traffic-aware scheme, NegotiaToR, uses single-iteration iSLIP as its scheduling engine, which limits throughput to around 60% and treats all source-destination pairs with equal priority regardless of queue length. We propose QPS-ToR, which replaces NegotiaToR's scheduling logic with SW-QPS, a sliding-window algorithm originally proposed for crossbar scheduling that achieves around 90% throughput with a single low-complexity iteration. QPS-ToR operates within NegotiaToR's existing workflow, requiring only a revision of the scheduling cycle from three-step Request-Grant-Accept (RGA) to two-step Request-Grant (RG) with a sliding window mechanism. In flow-level simulations with 128 ToRs under realistic datacenter workloads, QPS-ToR achieves up to 36% higher throughput and 82% lower flow completion time (FCT) compared to NegotiaToR, and consistently outperforms RotorNet, a representative traffic-oblivious scheme, on the parallel network topology.
Read the full abstract

Systems and Control

A general representation form of system data with feedback control and fundamental lemma as examples

This note introduces a general data representation of dynamic systems.

Soft Condensed Matter

A force model for dense granular flows and particle segregation

Dense granular flows are a common occurrence throughout nature and industry, but still exhibit phenomena that are not fully understood. One of these phenomena is the segregation of particles that constitute a dense granular flow, where particles separate due to differences in properties such as size and density. This work approaches segregation from a new perspective by developing a force model for a single particle based on the observation that there are two opposing drag forces acting on the particle. Validation of the model in mono-disperse flows reveals a family of velocity profiles, including a Bagnold profile. Further validation is done by comparing with known particle scaling laws in literature. The force model shows that the mechanism of segregation is always present in dense granular flows and that mono-disperse flows are a unique case where it is not distinctly visible. The insights in this paper prove a fundamental new way of thinking about dense granular flows and its individual particles, opening the field for new research directions.
Read the full abstract

Information Theory

Degree Sequence Reconstruction from Subgraph Traces

The goal of degree sequence reconstruction is to recover the ordered vector of degrees of an unknown graph from vertex deleted traces, where each vertex is deleted independently with probability $p$. We provide two algorithms for reconstruction; the first uses rejection sampling to reduce the problem to an estimation problem for a mixture distribution. Combined with prior trace reconstruction results, this gives a reconstruction algorithm using $\Exp{\tilde O (n^{1/3})}$ traces, although no sub-exponential time decoder is known. Our other approach involves recovering certain graph invariants, degree moments, that can identify a graphs degree sequence. Extremal polynomial bounds show that $\tilde \Theta(n^{1/2})$ degree moments are necessary and sufficient to reconstruct the degree sequence, which leads to an algorithm with $\Exp{\tilde O(n^{1/2})}$ trace complexity. The same polynomial machinery yields a sub-exponential time decoder for the degree sequence from the moments. Additionally, we give an $O(n^{3})$ upper bound and a $\Omega(n^2)$ lower bound for the trace complexity of recovering the number of edges.
Read the full abstract

Computer Vision

VANTAGE-Bench: Evaluating the Infrastructure AI Gap in Vision-Language Models

As Vision-Language Models (VLMs) advance toward physical deployment, the focus has remained on action-oriented Embodied AI evaluated on subject-centric consumer video. This overlooks a pervasive class of Physical AI: Infrastructure AI, which relies on fixed cameras for open-loop insights like safety monitoring and operational logging. We introduce VANTAGE-Bench, a benchmark measuring this "Infrastructure AI Gap." It spans three operational domains (Logistics, Transportation, and Smart Spaces), unifies image and video evaluation across semantic, spatial, temporal, and spatio-temporal capabilities, and moves beyond multiple-choice to eight task formulations including dense captioning and spatio-temporal grounding. It adds a single-pass trajectory protocol for Single Object Tracking and, to our knowledge, the first such evaluation on fixed-camera infrastructure video, scored against specialist trackers. Annotation spans three regimes over 3,346 media assets: 3,342 video-task annotations, 4,281 image-grounding annotations, and 27,404 detection boxes. Evaluating 17 models zero-shot, we find the shortfall relative to consumer-centric benchmarks is concentrated, not general. Event verification, referring expressions, and temporal localization fall roughly 9 to 24 points at every model scale, while video question answering stays within 5.3 points of VideoMME and 2D spatial pointing shows no shortfall against BLINK. The temporal pillar is weakest in absolute terms: no system exceeds 55.7 mIoU on temporal localization or 37.3 SODA_c on dense video captioning. On tracking, frontier models come within roughly 5 points of specialist trackers over short horizons but separate as the horizon extends. Open-weight models lead 2D object localization outright, so neither scale nor proprietary access explains the pattern. Data, evaluation harness, and leaderboard: https://vantage-bench.org/
Read the full abstract

All authors Zaid Pervaiz Bhat, Nimra Nayyar, Arihant Jain, Lap Fung Chan, John Suchanek, Yu Wang, Varun Praveen, Tomasz Kornuta, Vidya Nariyambut Murali

Artificial Intelligence

The Menu Is an Execution Prior: State-Path Tool Menus for Online Agents

Language models act through tools, yet practical agents face libraries containing thousands of interfaces. We introduce the tool menu as the short, ordered subset of available tools shown to an agent before execution. The agent can call only tools in this menu. Multi-step tasks require the final action and the prerequisite tools that create its inputs in a usable order. Current constructors rank tools by request relevance, which can surface the final action while omitting or delaying less obvious producers. We introduce the state path, a pre-execution route from the observable request state to the desired outcome, and propose State-Path Tool Menu to learn it. Our framework treats the menu as an execution prior over these routes. Its encoder represents which tools can run from the current state, how their outputs satisfy later inputs, and which orders recur in training paths. A retriever covers an executable entry, the missing-input producers, and the final action. A reranker then places producers before consumers. On ToolBench, our menu raises online success from 0.737 to 0.898 and outperforms retrieval, reranking, generation, and routing baselines without changing the agent. The State-Path menu also covers more complete chains with 32 tools than the official list covers with 128, and its success gain persists across executor families with different model capacities. Our code is at https://github.com/Met2348/State-Path.
Read the full abstract

Computer Vision

OmniPoint: Universal Monocular Metric Pointcloud from Any Camera

Recovering metric 3D geometry from monocular images is a fundamental computer vision task, yet current methods remain heavily fragmented by fixed camera model assumptions and inflexible input schemes. We present OmniPoint, a unified framework designed to generalize metric reconstruction across diverse imaging sensors, including pinhole, fisheye, and equirectangular projections, while accommodating varying geometric priors. To overcome projection rigidity, OmniPoint abandons conventional planar depth regression. It instead adopts a decoupled ray and distance representation alongside a decoupled training objective, explicitly separating the camera projection model from the scene structure. To address the severe scarcity of training data for alternative cameras, we introduce a bidirectional augmentation strategy that explicitly bridges labeled perspective data and unlabeled omnidirectional domains in 3D space. Furthermore, to seamlessly integrate optional inputs like camera intrinsics or sparse depth without destabilizing the network through feature distribution shifts, we propose a robust information injection mechanism. This mechanism utilizes learnable input state embeddings to resolve architectural ambiguity and applies vectorized Gaussian smoothing to densify irregular measurements. Extensive experiments demonstrate that OmniPoint achieves state-of-the-art zero-shot performance across multiple benchmarks, establishing a robust new standard for unified monocular 3D reconstruction.
Read the full abstract

All authors Botao Ye, Marc Pollefeys, Ming-Hsuan Yang, Abhijit Kundu

Signal Processing

Frame-Induced Doppler Self-Ambiguity in SiriusXM Satellite Signals for Passive Radar

Satellite Digital Audio Radio Service (SDARS) signals are attractive illuminators of opportunity for passive radar due to their continuous geostationary coverage and relatively high effective radiated power. However, the deterministic time-division multiplexed (TDM) framing employed by these waveforms introduces inherent self-ambiguity that can constrain coherent integration and detection performance. This paper analyzes the self-ambiguity function (SAF) of the XM/SiriusXM (SXM) waveform and explicitly links the observed Doppler periodicity to the repetition of the fast synchronization preamble (FSP) embedded in the TDM master frame. A frame-accurate XM TDM signal simulation is used to establish the expected ambiguity behavior, which is then validated experimentally using real direct-path data collected from the SXM-8 satellite. Results demonstrate pronounced and persistent Doppler replicas at integer multiples of the XM FSP repetition frequency, forming a deterministic Doppler “comb” (grating-lobe structure) that spans Doppler extents relevant to GEO-based passive radar operation with multi-second coherent processing intervals.
Read the full abstract

All authors Khalid El-Darymli, Christoph H. Gierull, Ryan A. English, Tony Laneve

Rings and Algebras

Classification of Nijenhuis operators on three-dimensional multiplicative simple Bihom-Lie algebras

We provide a complete classification of Nijenhuis operators on three-dimensional multiplicative simple BiHom-Lie algebras over the complex field C. Based on the classification of such algebras by Saadaoui into three families L1, L2, and L3, we derive the complete system of equations arising from the equivariance conditions and the Nijenhuis identity. We provide explicit matrix representations of all solutions, including all degenerate parameter cases. We construct the BiHom-Frölicher-Nijenhuis bracket on the cochain complex of a BiHom-Lie algebra, showing that Maurer-Cartan elements of this graded Lie algebra are precisely Nijenhuis operators. This enables us to define the cohomology of Nijenhuis operators and study formal deformations. We introduce BiHom-NS-Lie algebras as the algebraic structure underlying Nijenhuis operators. We provide explicit deformed BiHom-Lie algebra structures and prove the non-isomorphism of different families. We also study the moduli space of Nijenhuis operators and provide computational verification of all results.
Read the full abstract

Group Theory

On the Hamiltonicity of generating graphs of almost simple groups

The generating graph $\Gamma(G)$ of a finite group $G$ has vertex set $G\setminus\{1\}$, and two distinct vertices are adjacent if and only if they generate $G$. Breuer, Guralnick, Lucchini, Maroti and Nagy [Bull. Lond. Math. Soc. 42 (2010), 621–633] conjectured that, for every finite group $G$ with at least four elements, $\Gamma(G)$ contains a Hamiltonian cycle if and only if every proper quotient of $G$ is cyclic. They proved their conjecture for sufficiently large almost simple groups with alternating socle and for all almost simple groups with sporadic socle. In this paper, we complete the asymptotic picture for almost simple groups by proving the conjecture for sufficiently large almost simple groups with socle of Lie type.
Read the full abstract

Signal Processing

Forward-Scatter Bistatic RCS of a PEC Sphere: A Mie-Theory Analysis

Forward-scatter radar (FSR) represents an extreme bistatic configuration in which the transmitter–target–receiver geometry approaches collinearity, yielding a bistatic angle near $180^\circ$. In this regime, diffraction-dominated scattering can produce bistatic radar cross sections (RCS) that substantially exceed their monostatic counterparts, a property that underpins FS and passive bistatic radar concepts. This paper presents a rigorous electromagnetic study of the bistatic RCS of a perfectly electrically conducting (PEC) sphere using exact Lorenz–Mie theory. The analysis addresses two practical objectives: (i) quantifying the sensitivity of FS enhancement to deviations from the ideal $180^\circ$ bistatic angle as a function of electrical size, and (ii) decomposing the FS response into the canonical Lorenz–Mie polarization channels $S_1$ and $S_2$, thereby establishing a polarization-resolved benchmark. The results, however, are governed by the dimensionless size parameter $ka$ and therefore generalize directly to other frequencies and bistatic configurations. By explicitly relating full-wave simulation results to Babinet-based FS scaling and optical-theorem interpretations in the high-frequency limit, the paper provides a physically transparent benchmark for FS and passive bistatic radar analysis.
Read the full abstract

Statistical Methodology

A Counterfactual Framework for Estimating Infectious Disease Prevalence under Repeated Testing with Symptomatic and Contact-Tracing Components

This paper addresses the problem of estimating infectious disease prevalence under longitudinal testing programs that include scheduled, symptomatic, and contact-tracing testing. Our study is motivated by data from The Ohio State University, where a mandatory once-per-week COVID-19 testing and isolation program was implemented during the Fall 2020 semester, supplemented by additional testing for symptomatic individuals and identified contacts. In this setting, the probability of being tested depends on symptoms or contact-tracing status, creating a complex observation process. We develop a counterfactual framework that links the observation process to a hypothetical process in which infection is prevented. This formulation enables unbiased estimation of disease prevalence by modeling the testing process, possibly nonparametrically, without requiring explicit modeling of transmission dynamics, even though the testing and infection processes are jointly dependent.
Read the full abstract

All authors Jeongjin Lee, Junke Yang, Grzegorz A. Rempala, Patrick M. Schnell

Machine Learning

XAI-Refine: An Automated Explanation-Knowledge Loop for Brain-Age Prediction

Brain-age prediction models are commonly evaluated by predictive accuracy, yet accurate predictions alone do not establish that a model relies on reproducible or neurobiologically supported mechanisms. Post-hoc explanation methods can expose these mechanisms, but existing workflows typically stop at diagnosis or require correction targets to be specified before model analysis. We propose XAI-Refine, an automated explanation-knowledge loop for brain-age prediction from resting-state functional connectivity. At each iteration, XAI-Refine consolidates complementary post-hoc analyses across repeated training runs into reliable, structured model explanations. It converts each reliable explanation into a neutral neurobiological question, retrieves and verifies relevant literature, and compiles the verified evidence into an admissible set in the same typed explanation space. The target for refinement is defined as the minimal projection of the current model explanation onto the admissible set induced by applicable verified knowledge. This revised explanation is then translated into a differentiable constraint while preserving the originating model variable, measurement operator, and applicable scope. Candidate updates are promoted only when multi-seed validation confirms target-directed explanatory movement, predictive performance remains within a prespecified guardrail, and non-target explanatory drift remains bounded. Experiments on functional-connectivity-based brain-age prediction evaluate predictive performance, explanation reliability, literature alignment, and target-specific model revision, illustrating a structured route from post-hoc analysis to evidence-guided model refinement.
Read the full abstract

All authors Yang Qiao, Junjie Wu, Deqiang Qiu, James J. Lah, Liang Zhao