Peer Revenue Returns subject to peer review

Papers

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

All of arXiv

Monday 31 August

Accelerator Physics

A particle-in-cell model of beam dynamics in a dielectric wall accelerator

Dielectric wall accelerator (DWA) technology has been proposed as a compact, cost-effective alternative to rf accelerators for proton therapy, but its beam dynamics and practical feasibility remain relatively unexplored. In this work, we derive a three-dimensional, time-dependent axisymmetric electromagnetic field model from a prescribed on-wall excitation and implement it as stacked external field elements in the particle-in-cell code Warp. In the absence of experimental DWA beam-transport data, the implementation is cross-checked against a previously developed linear optics model in TRANSOPTR using deliberately idealized beam conditions. Strong agreement is observed between the two models in this regime. The models are then compared for larger transverse and longitudinal emittances, bunch charges up to 1x10^8e, and beam parameters representative of a low-energy proton source. The PIC simulations remain consistent with the linear predictions over much of the investigated range, while also identifying wall interactions, longitudinal phase-space distortions, and space-charge induced aberrations. The PIC model represents an intermediate step between linear optics and combined, geometry-specific electromagnetic and particle-transport simulations. It resolves particle-level transport and beam self-fields while retaining an analytical field description that can be varied without committing to a particular DWA structure. More detailed effects may be introduced through externally generated field data or empirical corrections, including models of cell-to-cell coupling, providing a practical basis for increasingly realistic DWA field and beamline studies.
Read the full abstract

All authors Christopher M. Lund, Paul M. Jung, Jamiel Nasser, Morgan J. Maher, Julien Bancheri, Chau Giang Bui, Thomas Planche, Rick Baartman, Jan Seuntjens

Statistical Methodology

Parameterising Gaussian Graphical Models

Gaussian graphical models (GGMs) describe the dependence structure among jointly Gaussian random variables. However, the most common parameterisation of GGMs, the precision matrix, describes both the dependence and scale of the variables. This has been shown to lead to model selection methods that depend on the scale of the variables, despite graphical models being scale invariant. Even after standardising data to have unit sample variances, entries of the precision matrix can be on different scales leading to poor model selection. This paper decomposes the precision matrix into marginal variances and interpretable scale-invariant parameters - the partial correlations and variance inflation factors. This decomposition gives new insights into the precision matrix entries and the dynamics of model selection methods such as penalised likelihoods. In particular, it explains the observed phenomenon that penalties on the precision matrix perform poorly at selecting hub variables and motivates the necessity of data standardisation. It also shows why methods based on partial correlations have better hub detection properties. The effect of penalisation of different quantities on the estimation of marginal variances is then investigated and an interesting simplification of the log-likelihood is shown when using maximum likelihood estimation of the marginal variances.
Read the full abstract

Solar and Stellar Astrophysics

Discovery and Characterization of Three New High-amplitude delta Scuti Stars from TESS Observations

We report the discovery and detailed analysis of three new high-amplitude {\delta} Scuti (HADS) stars, TIC 408074920, TIC 189714989, and TIC 34137913, identified from Transiting Exoplanet Survey Satellite short-cadence observations. Fourier analysis reveals dominant radial modes accompanied by rich harmonic structures in all three stars, confirming their HADS classification. In addition, TIC 189714989 exhibits clear symmetric side peaks around the radial harmonics, indicative of amplitude and/or phase modulation. Broadband spectral energy distribution (SED) fitting constrained by Gaia DR3 parallaxes provides independent estimates of the stellar effective temperatures and radii. These are compared with results from stellar evolutionary and seismic modeling based on MESA and GYRE. For two targets, the SED-derived and seismic radii are mutually consistent, while TIC 408074920 displays a significant discrepancy, plausibly attributable to the different sensitivities of the two methods and additional systematic effects. These findings highlight the diversity of amplitude variability among classical HADS and emphasize the importance of the long-term monitoring that future missions such as PLATO will allow in order to explore the underlying physical mechanisms of this variability.
Read the full abstract

All authors Chenglong Lv, Ali Esamdin, J. Pascual-Granado, Guojie Feng, Haozhi Wang, Rivkat Karimov, Bolkin Juraev, Shuhrat A. Ehgamberdiev

Systems and Control

Non-Blocking Fault Current Limiting Control of Half-Bridge MMCs for MTDC Transmission

A key protection requirement in multiterminal HVDC (MTDC) transmission systems is to selectively isolate faulted areas using DC circuit breakers (DCCBs) during DC faults while keeping the rest of the systems in operation. However, in MTDC systems using half-bridge modular multilevel converters (HB-MMCs), the lack of inherent fault current suppression capability in HB-MMCs can cause converter blocking before fault clearance by DCCBs. This work presents an active fault current limiting (AFCL) control strategy for HB-MMCs to mitigate DC fault currents, thus avoiding converter blocking and reducing the breaking capability required of DCCBs. In addition, a systematic approach is proposed to design the AFCL control strategy under converter and network stability constraints. The effectiveness of the proposed strategy is demonstrated through PSCAD on two typical MTDC systems during DC faults.
Read the full abstract

Networking

WiSDoM: Wireless Sparse Decision Transformer with Mixture-of-Experts for Multi-Task Mobile Network Optimization

Emerging 6G wireless networks are expected to operate across diverse deployment scenarios, where variations in network topology, user mobility, traffic demand, and radio conditions challenge the scalability of conventional radio resource management (RRM). While offline reinforcement learning (RL) methods have demonstrated strong decision-making capabilities, learning a single policy that performs consistently across heterogeneous wireless environments remains difficult due to conflicting optimization objectives and limited model specialization. These challenges become particularly pronounced in coordinated multipoint (CoMP) transmission, where selecting the optimal serving-cell combination requires sequential decision-making under evolving network conditions. This paper presents the Wireless Sparse Decision Transformer with Mixture of Experts (WiSDoM), a sparse multi-task offline RL framework for adaptive multi-cell selection. WiSDoM combines Decision Transformers (DTs) with a Mixture-of-Experts (MoE) architecture that dynamically activates specialized experts according to task characteristics. This MoE mechanism improves model capacity without proportionally increasing inference cost, mitigates negative transfer, and enables expert specialization across tasks. WiSDoM is trained jointly on diverse network configurations spanning multiple base station and user equipment densities, mobility levels, and scheduler policies. Experimental results show that WiSDoM consistently outperforms heuristic methods, single-task models, and conventional multi-task DTs, improving quality of experience (QoE) by up to 55% while activating approximately one-third of the parameters of its dense counterpart during inference. Furthermore, WiSDoM exhibits strong task generalization and efficiently adapts to unseen wireless scenarios through few-shot prompting without retraining or fine-tuning.
Read the full abstract

Astrophysics of Galaxies

ELVES-Dwarf. II. A Systematic Search for Satellite Systems of Dwarf Galaxies in the Local Volume

We present the Exploration of Local VolumE Satellites of Dwarf Galaxies (ELVES-Dwarf) survey, a systematic census of satellite systems around dwarf hosts in the Local Volume. Our final sample comprises 39 predominantly isolated hosts with stellar masses $10^{7}<M_\star<10^{10}\,M_\odot$, including 32 hosts searched uniformly in this work and 7 drawn from the literature. We search for satellite candidates within the projected virial radius of each host using $155~\mathrm{deg}^2$ of Legacy Surveys imaging data. We determine satellite membership using surface brightness fluctuation distances from Subaru/HSC, Magellan/IMACS, and Gemini/GMOS imaging, supplemented by literature TRGB distances and radial velocities. From 207 candidates, we confirm 39 satellites with $M_\star>10^5\,M_\odot$ around the 39 hosts. Above our fiducial completeness threshold of $M_\star\gtrsim10^{5.7}\,M_\odot$ and within the projected virial radius, 21 hosts have no confirmed satellites, 10 have one, six have two, and two have four, revealing substantial host-to-host scatter in satellite abundance. Overall, the observed satellite abundances and stellar mass functions are broadly consistent with predictions from the cosmological simulation TNG50 and galaxy formation models calibrated using Milky Way satellites. The projected radial distribution of the satellites is also consistent with theoretical expectations and with satellite populations around Milky Way-mass hosts. In contrast, the quenched fraction of satellites around dwarf hosts is substantially lower than around Milky Way-mass hosts, suggesting that environmental quenching is less efficient in dwarf halos. ELVES-Dwarf provides the first large, homogeneous, distance-confirmed sample of satellites around dwarf hosts and establishes a foundation for understanding galaxy formation and evolution in less-dense environments.
Read the full abstract

All authors Jiaxuan Li, Jenny E. Greene, Shany Danieli, Scott G. Carlsten, John Moustakas, Marla Geha, Masayuki Tanaka, Fangzhou Jiang, Ping Chen, Sufia Birmingham

Computational Engineering

Lightweight Adaptation of EEG Foundation Models for Stroke Motor Imagery Decoding: Domain Shift and Subject-Level Robustness

Motor imagery (MI) electroencephalography (EEG) decoding could support post-stroke rehabilitation, but models developed on healthy cohorts may not transfer reliably to pathological EEG. We evaluated whether Low-Rank Adaptation (LoRA) can efficiently adapt three pretrained EEG foundation models (i.e., LaBraM-base, REVE-base, and REVE-large) for binary left- versus right-hand MI decoding. Frozen-backbone head-only baselines and LoRA adaptation were evaluated using subject-wise five-fold cross-validation on the PhysioNet EEG Motor Movement/Imagery Dataset and a binary subset of the UET175 dataset comprising 30 stroke participants. On EEGMMIDB, LoRA increased accuracy to 0.822 for LaBraM-base and 0.957 for REVE-base. On UET175, all head-only models performed near chance. With LoRA, LaBraM-base remained near chance (0.499$\pm$0.009), whereas REVE-base reached 0.847$\pm$0.194 and outperformed REVE-large (0.806$\pm$0.178), indicating that increased model capacity alone did not improve stroke-domain adaptation. The strongest stroke configuration, REVE-base LoRA, was further evaluated using within-cohort leave-one-subject-out cross-validation (LOOCV), showing 0.952 mean accuracy, but subject-wise accuracy ranged from 0.586 to 1.000, revealing a small low-performing tail. Zero-shot transfer from EEGMMIDB to UET175 remained near chance (0.464$\pm$0.072). These findings show that healthy-benchmark performance does not ensure transfer to stroke EEG. Translation of EEG foundation models to pseudo-online or real-time rehabilitation BCIs should therefore include target-domain adaptation and subject-level assessment of temporal informativeness, spatial sensitivity, and physiological discriminability.
Read the full abstract

Earth and Planetary Astrophysics

Three-dimensional circumplanetary flows in a PDS 70c-inspired system: hydrodynamic simulations with FARGO3D and analysis with FARGOpy

The detection of circumplanetary material around the forming planet PDS~70c provides a valuable setting for studying three-dimensional gas accretion during giant-planet formation. We present three-dimensional hydrodynamic simulations with FARGO3D for a locally isothermal protoplanetary disk, adopting system-motivated stellar, orbital, and disk parameters and planetary masses of $2$, $4$, and $8,M_{\rm Jup}$, all above the local thermal mass. The locally isothermal prescription provides a controlled prescribed-temperature limit that isolates the hydrodynamic organization of the flow, without modeling self-consistent thermal evolution or radiative observables. Gas enters the planetary environment through meridional flows from the upper disk layers and streams near the midplane. With increasing planetary mass, the flow changes from a vertically extended, envelope-like configuration to a more compact and stratified circumplanetary structure. The inferred CPD radius increases from $R_{\rm CPD}\simeq0.31$ to $0.40,R_{\rm Hill}$, while the flow remains sub-Keplerian, reaching $v_\phi/V_{K,p}\sim0.79$–$0.85$. Most enclosed gas mass lies within a few tenths of the Hill radius, whereas the outer Hill sphere is dominated by anisotropic inflow, outflow, and recirculation. The analysis was performed with FARGOpy, a Python post-processing framework developed in this work to reconstruct three-dimensional fields, compute integrated fluxes, and characterize circumplanetary kinematics from FARGO3D outputs. This study therefore provides both a hydrodynamic baseline for a PDS~70c-like environment and a reproducible framework for analyzing three-dimensional FARGO3D simulations.
Read the full abstract

Combinatorics

List coloring $C_3$-free planar graphs with a sparse matching of restricted lists

A graph $G$ is $k$-choosable if it has a proper coloring for every $k$-list assignment. While every $C_3$-free planar graph is $4$-choosable, some of them are not $3$-choosable, as constructed by Voigt. Hu and Zhu conjectured that if $G$ is a $C_3$-free planar graph and $X \subseteq V(G)$ induces a bipartite subgraph, then $G$ has a proper $L$-coloring whenever $|L(x)| = 3$ for $x \in X$ and $|L(v)| = 4$ for $v \in V(G) \setminus X$. As evidence, they proved the conjecture when $X$ is an independent set. We provide further evidence by proving the conjecture when the induced subgraph $G[X]$ is an induced sparse matching. This is the first result supporting the conjecture in which the set $X$ receiving smaller lists may induce a subgraph with edges.
Read the full abstract

All authors Stephen G. Hartke, Yupei Li, Joseph Pappe, Fares Soufan, Lin Tian, Zimu Xiang

Machine Learning (Statistics)

Exact Global MCMC with Denoising Diffusion

This work shows that diffusion models learned with standard denoising loss can provide effective global MCMC proposals for complex high-dimensional target densities. The method is motivated by the observation that sequentially applying a forward and reverse diffusion process defines a Markov chain with a target stationary distribution for an ideal denoiser trained on samples of the target distribution. This observation can be made exact for any denoiser by applying a Metropolis-Hastings step whose acceptance ratio includes the density of the forward and reverse paths of a discrete time SDE approximation. We therefore propose to train denoising diffusion models on locally convergent MALA samples to learn global MCMC proposals. We call the composition of the global denoiser-based path sampler and a local MALA sampler Denoising Diffusion Monte Carlo (DDMC). Experiments show that DDMC can provide global proposals with high acceptance across a variety of complex target densities. Our results offer preliminary evidence that the established scaling behavior of standard diffusion training transfers directly to exact sampling from high-dimensional unnormalized densities.
Read the full abstract

Astrophysics of Galaxies

Accretion-disk sizes in two quasars with interferometrically resolved broad-line regions at $z=2.3$ and $z=4.0$

We present the first direct comparison between accretion-disk (AD) and broad-line region (BLR) sizes in quasars at z > 2, combining continuum reverberation mapping with interferometric BLR constraints from GRAVITY and GRAVITY+. Using medium-band photometric monitoring with the MPG/ESO 2.2 m telescope, we measure inter-band continuum lags in SDSS J092034.17+065718.0 at z = 2.33 (J0920) and SMSS J052915.80-435152.0 at z = 3.96 (J0529), among the most luminous quasars known (Lbol ~ 10^48 erg s^-1). These are currently the only two quasars at z > 2 with spatially resolved BLRs and dynamical black-hole masses from interferometry. We detect significant continuum lags in both quasars, increasing monotonically with wavelength. The inferred UV disk sizes are RAD = 4.35 +0.78/-0.91 light-days for J0529 and RAD = 3.15 +0.50/-0.48 light-days for J0920. For J0920, accreting at lambda_Edd ~ 7-20, the disk size is consistent with standard thin-disk expectations despite its super-Eddington regime. For J0529, the disk size agrees with thin-disk predictions using the single-epoch black-hole mass, but implies disk inflation by a factor of a few if the GRAVITY+ dynamical mass, an order of magnitude lower, is adopted. UV continuum disk sizes therefore provide an independent physical scale constraining black-hole mass and accretion-rate models, particularly where BLR kinematics are dominated by outflows. The interferometric BLR sizes reveal pronounced radial hierarchies, with RBLR/RAD ~ 270 for J0529 (Hbeta) and ~115 for J0920 (Halpha). The successful lag detections at Lbol ~ 10^48 erg s^-1 show that continuum reverberation mapping remains feasible for the most luminous systems, opening a path to larger samples with surveys such as the Vera C. Rubin Observatory's LSST.
Read the full abstract

All authors Francisco Pozo Nuñez, Eduardo Bañados, Swayamtrupta Panda, Amit Kumar Mandal, Bozena Czerny, Jochen Heidt

Strongly Correlated Electrons

Engineering tunable $p$-wave magnetism in antiferromagnetic bilayers

We propose a symmetry-guided route to engineer tunable $p$-wave magnetism in a bilayer system composed of two AB-stacked antiferromagnetic square lattices with their collinear moments nonparallel to each other. We show that an in-plane relative shift between the two layers selectively breaks the symmetries protecting spin degeneracy while preserving time-reversal-related constraints, thereby generating odd-parity spin splitting in a fully compensated magnetic state without spin-orbit coupling. The direction and the magnitude of the resulting $p$-wave spin-splitting, together with the associated spin responses, can be continuously tuned by the in-plane displacement, providing a potential knob for control. We further map the bilayer system onto an effective bond-modulated square lattice and investigate the corresponding Hubbard model within a mean-field framework. The calculated phase diagram reveals extended regions where $p$-wave magnetism with coplanar or chiral spin textures emerges spontaneously. Our results establish a minimal equilibrium platform for realizing and manipulating $p$-wave magnetism in two-dimensional systems.
Read the full abstract

All authors Yu-Han Lin, Jin-Wei Dong, Ziqiang Wang, Sen Zhou

Sound

Cleaner Speech, Weaker Generalization: Revisiting Pitt-Derived Benchmarks for Alzheimer's Disease Detection

Speech-based Alzheimer's disease (AD) detection increasingly relies on speech-enhanced and curated versions of the Pitt Corpus, where speech enhancement, sample selection, and demographic balancing are often treated as beneficial preprocessing steps. However, whether these transformations improve real-world AD detection or instead affect model generalization and prediction behavior remains unclear. In this work, we revisit the role of speech preprocessing and dataset curation across widely used benchmarks for speech-based AD detection. We evaluate the speech quality of different datasets, the cross-dataset generalization of multiple deep learning models under matched and mismatched enhancement settings, and the behavior of several recent large audio-language models (LALMs). Experimental results show that across multiple supervised speech models, speech-enhanced datasets often improve in-domain performance while reducing robustness in cross-domain evaluation. Matched enhancement between training and test data alleviates, but does not eliminate, this degradation. LALMs show a similar sensitivity: enhanced datasets induce stronger class imbalance and prediction shifts than unprocessed data. These results suggest that speech preprocessing and dataset curation can substantially influence downstream AD detection behavior, indicating that “cleaner” speech datasets are not necessarily more reliable for real-world AD detection.
Read the full abstract

All authors Luqi Sun, Shreeram Suresh Chandra, Lin Zhang, You-Jin Li, Brian MacWhinney, Yu Tsao, Emily Mower Provost, Berrak Sisman

Artificial Intelligence

The Irreversibility Budget: Fleet-Level Risk Accounting and Admission Control for Agent Operating Systems

Fleets of LLM agents now externalize effects that cannot be fully undone: they move money, deploy code, delete data, and disclose information. Current controls check one effect at a time, so a fleet of individually authorized agents can overdraw its principal's risk under a shared trigger while every local gate stays correct. We propose the irreversibility budget, a cumulative account of residual value-at-risk that a trusted runtime maintains for each principal across agents, workflows, and tenants. Treating irreversibility as a first-class resource, the runtime charges each effect its residual loss below the agent and denies the marginal effect once the aggregate would overdraw the budget. Getting the price right is hard, because effects are heterogeneous, adversarially declared, and correlated. We perform a controlled study in which per-effect gates admit fleet-level overdraws of up to 48 times the tenant's risk limit while the budget holds every correctly charged run within that limit. Conservative, dependency-aware pricing remains the central open problem for a deployable design.
Read the full abstract

Artificial Intelligence

Autoresearch for Marketplace Catalogs: From Legacy Forms to AI-Native Matching

Two-sided service marketplaces are moving from deterministic request-form intake to AI-native probabilistic matching, enabled by large language models (LLMs) that infer intent, preferences, and latent constraints from natural language. Relying on inferred intent rather than fixed-form fields forces these platforms to regenerate the provider-side preference taxonomy underwriting matching, search, and pricing: attributes interpretable to service providers while remaining a useful signal for marketplace decisions. We present an autoresearch loop that generates this taxonomy, one occupation at a time, and has been deployed in production at a major U.S. consumer services marketplace since April 2026, spanning 132 occupations. Instead of one global hierarchy, the loop treats each occupation as an independent generation problem and runs iterative propose-evaluate-keep refinement cycles. Each candidate tag set is scored by a recalibrated six-rubric LLM-as-judge framework, and a 7-critic panel of distinct personas contributes weighted penalties to an adjusted score, with no hard vetoes. A separate parity-mapping stage maps legacy request-form Q&A pairs back to the generated taxonomy, yielding both a coverage signal and an interface for human quality assurance; it does so by first inferring the provider attribute each legacy question was meant to measure, rather than translating questions to tags literally.
Read the full abstract

All authors Kartik Ravisankar, Hojat Abdolanezhad, Daniel Capo, Sang Su Lee, Shishir Dash, Vijay Anand Raghavan

Rings and Algebras

An Algebraic Framework for Data Systems: Classification and Optimization of Algebraic Structures for Data Organization and Coding

Modern data systems are commonly studied through computational and information-theoretic methods, while their algebraic properties remain largely unexplored. This paper introduces a mathematical framework for modelling data systems using algebraic codes over finite groups, organized by an axiom signature: each algebraic axiom of the data domain is shown to be sufficient for a specific coding-theoretic capability, and each implication is given its precise reading (MDS parameters do not require a field; quotients require normality, not commutativity). Three types of results are presented. Proved: a Gilbert-Varshamov existence bound over arbitrary finite groups, commutative or not; explicit direct-sum codes over the elementary divisors of any finite abelian group; that transactions preserving a validity predicate form a submonoid of the full transformation monoid, which is not a group; that schema-preserving transformations form a finite group whose orbits are counted by Burnside's lemma; and that the constraint-preserving bijections of a data system form a finite group whose orbits partition the valid configurations. Demonstrated: an explicit code construction over $\mathbb{F}_5$, with known codes over $\mathbb{Z}_4$ and $(2^U,\triangle)$ cited to situate the framework. Proposed: an encoding efficiency measure Ef and an optimization functional $\Phi$ with tunable weights, both evaluated at declared coding instances. Weight-independent comparison holds under componentwise dominance; in general the weight simplex decomposes into polyhedral chambers on which the induced order is constant.
Read the full abstract

2609.00273 · 31 Aug · 26 pages, 7 tables, 22 references. Presented as a poster at ICM 2026, Philadelphia, PA, USA. Revised: assumptions clarified, theorems corrected, Gilbert-Varshamov bound added; classifier gives sufficient conditions; ranking weight-independence withdrawn

Computer Vision

CrossFeat: Bridging Imaging Modalities in Feature Descriptor Space

Most advances in keypoint descriptions address monomodal settings, where image variations arise from viewpoint, illumination, or contrast changes. Multimodal scenarios involve images produced by fundamentally different sensing processes, such as multispectral imaging, RGB-depth, satellite imagery, or medical imaging, causing the same structures to appear differently. A common solution to cross-modal description is to train descriptors for each modality pair, which requires retraining whenever the modalities change, or to train large models, which incur a significant increase in runtime. Instead, we propose CrossFeat, a framework that enables an existing monomodal descriptor to operate across modalities. Our method learns a crossing function in descriptor space that maps features from one modality to a representation compatible with another. To preserve the structural information captured by the original descriptor, CrossFeat introduces a geometry-appearance disentanglement such that only appearance is altered while the geometric properties are preserved. Experiments across multiple domains and datasets demonstrate improved performance in multimodal matching.
Read the full abstract

Applied Physics

A fully coupled electromagnetic-thermal-mechanical model for metal-insulated HTS high field magnets

Ultra high field REBCO magnets operate under strongly coupled electromagnetic, thermal and mechanical conditions, where screening currents, localized heating, thermal expansion and Lorentz forces can modify both the structural state and the critical current density of the conductor. In this work, a coupled electromagnetic, thermal and mechanical model is developed for a metal-insulated nested REBCO insert designed for a 40 T class SuperEMFL magnet. The existing electromagnetic formulation resolves the non-uniform screening currents in the REBCO tapes. The thermal model is extended from an explicit Finite Difference Method (FDM) to an implicit Backward Euler scheme with Picard iteration, while a new axisymmetric mechanical FDM solver based on BiCGSTAB is introduced to calculate displacements, strains and stresses in the coil windings and G10 spacer regions. Thermal expansion and Lorentz force contributions are included, and the calculated longitudinal mechanical strain is coupled back to the electromagnetic model through a strain dependent critical current density, which also depends on temperature, magnetic field, and its orientation. A literature-informed Parabolic-Weibull model is used to model reversible and irreversible strain degradation of the REBCO conductor. The numerical methods are benchmarked, and the resulting framework provides a computationally efficient approach for investigating temperature gradients, thermo-mechanical stresses, strain-dependent critical current degradation and quench behaviour in full scale nested high field REBCO magnets.
Read the full abstract

All authors Anang Dadhich, Nikola Jerance, Tara Benkel, Philippe Fazilleau, Enric Pardo

High Energy Physics (Phenomenology)

Two-Stage Machine Learning Strategy for Scalar and Vector Leptoquark Discrimination at the LHC

We present a machine learning framework for the characterization of leptoquark (LQ) signals at the Large Hadron Collider, focusing on the discrimination between scalar (SLQ) and vector (VLQ) hypotheses. The method is based on a two-stage inference pipeline that combines a classifier trained to separate Standard Model backgrounds from a mixed LQ signal with a second classifier designed to distinguish between SLQ and VLQ scenarios, using the signal yield inferred from the first-stage classifier to guide the corresponding scalar and vector mass hypotheses. A test statistic is constructed from the classifier outputs and interpreted using reference probability density functions. The approach is applied to realistic LHC final states with hadronically decaying tau leptons, multiple jets, and missing transverse momentum, and its performance is assessed using simulated pseudo-experiments. We show that the proposed strategy provides a robust and statistically consistent procedure to discriminate between SLQ and VLQ signals across a wide range of masses and signal stregths. We find that the spin identification power closely follows the discovery potential, demonstrating that determining the spin nature of a newly discovered LQ does not require substantially larger datasets than those needed for discovery itself.
Read the full abstract

All authors Ernesto Arganda, Martín de los Rios, Andres D. Perez, Rosa M. Sandá Seoane, Alejandro Szynkman

Algebraic Geometry

A holomorphic (2, 2)-Theorem for Abelian varieties of CM-type

For complex tori with algebraic normalized period matrix, in particular for Abelian varieties of CM-type, the map from the second cohomology group of the analytic sheaf of the second Milnor K-groups to the rational cohomology classes of Hodge type (2,2) is surjective (up to denominators).
Read the full abstract

Algebraic Geometry

Thom series in negative relative codimension

We develop a theory of Thom series for contact function singularities. Quadratic stabilization $\sigma_q$ of a contact singularity $\eta\in J^k(n,1)$ gives contact singularities $\sigma_q^{i}\eta\in J^k(n+i,1)$. We study the stable Thom polynomials (i.e.\ the Thom polynomial in quotient variables) of $\sigma_q^{i}\eta$ as $i$ increases. To this end, we define the Thom series of a contact function singularity $\eta$, which for any given $i$ is a linear combination of Schur polynomials $s_\lambda$. For large enough $i$ the value of the Thom series at $1-(n+i)$ is the stable Thom polynomial of $\sigma_q^i\eta$ and for small $i$ it determines its coefficients outside the kernel of a specialization map. We prove that the Thom series has two main properties: 1) as $i$ increases, the partitions $\lambda$ follow a simple stabilization pattern, and there is a finite set of $\lambda$ which generates the support of the entire Thom series via this stabilization; 2) the coefficients of the $s_\lambda$ are polynomials in $i$ with explicit degree bounds. We describe the precise relationship between unstable and stable Thom polynomials of contact function singularities and Legendre Thom polynomials and we carry out computations of all three, as follows. We compute the complete family of stable Thom polynomials for function singularities with $\gamma\leq 6$. We compute unstable and Legendre Thom polynomials for several families of binary and ternary singularities. We define a class of multi-binary singularities, and compute their Thom polynomials. We discuss second order Thom-Boardman classes, and indicate difficulties that arise beyond function singularities. The results of the paper will be used in a companion paper, where Thom polynomials will be applied to problems in enumerative geometry.
Read the full abstract

Cryptography and Security

Delegation Without Trust: An Empirical Gap Analysis of Identity, Authorization, and Runtime Governance in Multi-Agent LLM Systems

Autonomous LLM agents increasingly act on a user's behalf: they hold credentials, call tools and services, and spawn sub-agents that act further on their behalf. This turns a long-standing distributed-systems question – who is authorized to do what, on whose authority – into an urgent and largely unsolved problem, because the component driving each agent is a language model an adversary can hijack. We argue that agent security must be evaluated under an untrusted-model assumption: a correct system is one in which a fully prompt-injected agent still cannot exceed the authority explicitly delegated to it. Against this standard we make three contributions. First, we give a threat model for multi-agent delegation centered on four adversaries – confused deputy, token theft and replay, prompt-injection privilege escalation, and compromised sub-agents – and derive eight security requirements a governed agent system must meet. Second, we show the gap is real: a default agent runtime modeling common practice (broad bearer credentials, authorization gated inside the model) fails all four threats, and across four widely used frameworks – LangGraph, CrewAI, AutoGen, and the Model Context Protocol (MCP) authorization model – three provide no built-in confinement and one only partial; no existing standard alone covers the requirement set. Third, we implement and adversarially evaluate an authorization broker that closes the gap. It blocks all four threats; it resists 11 direct attacks on its design and accepts 0 of 200,000 forged tokens; it confines a compromised sub-agent to its delegated task (a mean of 1.5 reachable actions versus all 8,100 under bearer delegation, across 2,000 randomized scenarios); and it enforces at microsecond cost (about 2.6 microseconds per decision), negligible against model inference. These principles are also realized in production in VotalAI's LLM Shield.
Read the full abstract

Lattice Field Theory

Long-distance hybrid spin-dependent and hybrid-quarkonium mixing potentials from lattice gauge theory

We present $SU(3)$ lattice gauge theory results for the $\mathcal{O}(1/m_Q)$ hybrid spin-dependent and hybrid-quarkonium mixing potentials. The only existing lattice computation of these potentials is limited to quark-antiquark separations smaller than approximately $0.5\,\text{fm}$. In this work, we extend the lattice results up to around $1.25 \, \text{fm}$ using large-volume, high-statistics simulations at several lattice spacings and gradient flow times, providing the first lattice data in this previously unexplored regime.
Read the full abstract

All authors Vilija de Jonge, Paul Pütz, Antonio Vairo, Marc Wagner

Instrumentation and Methods for Astrophysics

The Hopkins Ultraviolet Telescope: A Historical and Personal Retrospective

This manuscript documents the inception and development of the Hopkins Ultraviolet Telescope (HUT) project, which was a component of the Astro Observatory payload that flew on two Spacelab shuttle missions in the 1990s. The discussion places the HUT project in the context of the times and highlights the importance of its acceptance by NASA in 1980 to the growth and development of astrophysics at the Johns Hopkins University. Personal recollections from the author, including a detailed description of the real time flight operations during each mission, provides an insiders view. The document includes a link to an online historical archive of supporting materials compiled in support of this effort.
Read the full abstract

Artificial Intelligence

The Answer Is Not the Argument

Chain-of-thought monitoring is proposed for AI oversight, yet evaluations often provide monitors with a trusted reference answer. We ask whether answer access improves verification of the reasoning or mainly supplies information about its conclusion. We collected 237 naturally generated, step-numbered solutions to 79 Humanity's Last Exam physics questions and independently labelled final-answer correctness and the first false step. Eight LLM monitors evaluated the traces with varying access to the reference answer. Certification raised mean balanced accuracy from 0.637 to 0.796, but its effect on error detection depended strongly on the conclusion: recall increased by +0.299 on wrong-answer traces, while there was no evidence of improvement on correct-answer traces containing a reasoning error (-0.083, 95% CI [-0.196, +0.030]). We then held the reasoning trace fixed in a seven-monitor certificate-congruence intervention. Replacing the true certificate with the trace's own incorrect conclusion reduced flagging by 0.659 (95% CI [0.602, 0.711]) and left flagging 0.389 below the answer-blind level. Conversely, a conflicting false certificate increased flagging of clean traces by 0.580, with 82.9% of newly flagged cases assigning the alleged error to an interior reasoning step. Trusted-answer access can therefore make monitoring appear substantially stronger because aggregate performance combines independent reasoning verification with a powerful certificate-conclusion consistency signal.
Read the full abstract

All authors Will Yeadon, Sergio Juárez, Paul Mackay, T. J. Dowling, Elise Agra, Oto-obong Inyang, Arin Mizouri, Craig P. Testrow

Earth and Planetary Astrophysics

A JWST transiting survey of FGK stellar limb darkening: empirical evidence for quadratic laws and atmospheric model comparisons

We present a study of stellar limb-darkening using JWST transit observations of seven exoplanets orbiting FGK host stars, spanning 4200-6800 K. The wide wavelength coverage and high S/N of NIRISS/SOSS and NIRSpec/PRISM enable precise constraints on the wavelength-dependent limb-darkening. Using Bayesian model selection, we find that the quadratic limb-darkening law is statistically preferred over higher-order laws for most FGK stars, recovers consistent intensity profiles for $>$95% of the stellar disk area, and introduces a minimal bias in the derived transit depths of only $\sim$14 ppm (1$\sigma$) for 6/7 targets. This contrasts with previous studies relying on stellar models. We compare the empirically derived quadratic coefficients to predictions from the PHOENIX, MPS-ATLAS, MURaM, and Stagger stellar atmosphere grids. We introduce a quadratic limb-darkening parameterization in terms of limb intensity ($\ell$) and curvature at mid-$\mu$ ($\delta$), finding that empirical FGK limb darkening is generally more linear than models predict ($\delta \lesssim 0.1$). We identify wavelength-independent offsets between data and model quadratic coefficients, minimized by adopting $\mu_{\rm min} = 0.2$ in intensity calculations; we attribute this in part to models overpredicting limb-darkening near the limb where the plane-parallel approximation breaks down. For spherical PHOENIX models, we derive a $\mu$ rescaling method using a $\tau = 1$ photospheric radius. With these corrections, residual offsets are minimized and all stellar models achieve statistically acceptable fits. We provide recommended limb-darkening offset priors for use in JWST transit analyses, enabling more accurate constraints on exoplanet transmission spectra while accounting for residual stellar model uncertainties.
Read the full abstract

All authors David K. Sing, Joshua D. Lothringer, Jeff A. Valenti, Natalie H. Allen, Katherine A. Bennett, Carlos Gascón, Mei Ting Mak, Patrick McCreery, Sagnick Mukherjee, Lakeisha M. Ramos Rosado, Stephen P. Schmidt, Kevin B. Stevenson, Daniel P. Thorngren, Le-Chris Wang

Fluid Dynamics

Creeping flows through confined arrays of cylinders

Hair-covered appendages serve a variety of purposes in Nature, from chemical sensing and particle capture on the antennae of a crustacean to drag generation on bristled wings. At low to intermediate Reynolds numbers, these finite porous media experience three flow regimes. At low Reynolds numbers, the flow goes around the porous structure; this is the paddle or rake regime. As the Reynolds number increases, so does the relative flow rate through the array. If the fluid exits the structure mostly laterally, the flow is in the deflection regime. If the fluid exits downstream, the flow is in the sieve regime. Confining structures, such as the animal body or larger hairs, have been hypothesized to focus the flow on the hair-covered region. We investigate the influence of confinement on the flow through and around an array of cylinders, using a combination of experiments and numerical simulations. Experimentally, we vary the cylinder spacing, channel dimension, and flow rate and measure the velocity field using Particle Image Velocimetry. After comparing the results of finite element analysis with the experimental data, we numerically investigate a broader range of system geometries and flow parameters. Our results show that the confinement focuses the flow in the array and shifts the domains of existence of the three regimes. We present an analytical model that relies on the permeability of rectangular slits to predict the relative flow rate through and around the array. The model is in quantitative agreement with the numerical results, demonstrating that the flow through the array increases with increasing confinement, while the flow angle decreases. These results should provide insight into the morphology of hairy surfaces and have implications in the design of bio-inspired flow sensors and filters.
Read the full abstract

All authors Sean Bohling, Sri Savya Tanikella, J.P. Raimondi, Nathan Jones, Emilie Dressaire

Optimization and Control

An Inexact Augmented Lagrangian Method for $(L_0, L_1)$-Smooth Convex Optimization

Augmented Lagrangian methods are among the most effective approaches for solving constrained convex optimization problems. However, classical complexity analyses of first-order methods applied within the augmented Lagrangian framework usually rely on the assumption that the objective function has a Lipschitz continuous gradient. This assumption excludes an important class of generalized smooth functions whose gradients may grow unboundedly. In this paper, we study an inexact augmented Lagrangian method for solving linearly constrained convex optimization problems with $(L_0,L_1)$-smooth objective functions. We show that the augmented Lagrangian subproblems preserve the $(L_0,L_1)$-smooth structure, with parameters depending on the penalty coefficient. This property allows us to employ recent accelerated first-order schemes designed for generalized smooth optimization instead of classical smooth optimization methods. In particular, we combine the inexact augmented Lagrangian framework with a two-stage acceleration procedure based on clipped gradient descent and accelerated optimization.
Read the full abstract

Instrumentation and Methods for Astrophysics

Simulation and development of astrophotonic H-band nulling beam combiner for the CHARA Array

Nulling interferometry is a promising method for direct detection and characterization of faint stellar companions such as exoplanets and faint binary sources. The dominant constraint on instrument performance is instability noise from vibrations, thermal expansions, and other systematic sources. Photonic instruments have footprints of a few centimeters and enable advanced light control architectures. Astrophotonic implementations can include active phase control via Mach-Zehnder interferometers (MZIs) to actively minimize residual optical path errors at the location of beam combination. The compact size of photonic chips also makes them a promising solution for space-based applications such as the upcoming Habitable Worlds Observatory. The Center for High Angular Resolution Astronomy (CHARA) Array is equipped with the world's longest baseline in the near-infrared at 330 meters long. CHARA therefore has the most sensitive angular resolution in the near-infrared. A nuller at CHARA would have access to a novel parameter space essential for exoplanet discovery and characterization. The first step in preparing a nulling beam combiner for CHARA is to develop compelling, realistic, and well-defined science cases. Simulation suites quantify the technical requirements to reach such science goals in their respective complex systems. The CHARA Array Response Model (CHARM) is the first-ever simulator built specifically for CHARA, opening up new pathways for instrument simulation. CHARM is a flexible simulation tool of an H-band, self-calibrated astrophotonic nulling beam combiner with active phase control for the CHARA Array. In this paper, we present an update on the development of our photonic nulling beam combiner for CHARA. We include early results from CHARM, demonstrating that the photonic nuller has the potential to detect and characterize a wide range of bright stellar companion systems.
Read the full abstract