Human-Computer Interaction
Visual Embellishments are Potential Distractions in Double-Column Reading
Songwen Hu, Chase Stokes, Marti Hearst et al. (+1 more)
Eye-catching graphics, such as circular figure labels and word-scale visualizations, are increasingly being placed directly within long-form text paragraphs. Some research has claimed that inline visualizations can help readers understand data-rich passages more clearly. However, research in the science of reading calls into question the introduction of images within the flow of text. In this work, we conduct an exploratory study of eye movement in both the presence and absence of visual embellishments. Using a high-resolution eye-tracker (EyeLink Portable Duo), we observed small mean increases in vertical and cross-column saccade rates among six participants, with substantial variation among readers and no detected difference in comprehension accuracy. Mean subjective ratings for the data-light circular-glyph passage indicated it is more distracting than its unembellished comparison passage. These exploratory observations differ from prior findings and motivate a larger, fully crossed study of inline graphics.
Read the full abstract ▾
All authors Songwen Hu, Chase Stokes, Marti Hearst, Cindy Xiong Bearfield
Software Engineering
A Governance-Aware Large Language Model Orchestrated Agentic Digital Twin for Transmission System Operator Control Room Decision Support
Costas Mylonas, Magda Foti and Emmanouel Varvarigos
Transmission system operators face rising complexity from renewable integration, reduced inertia, and tighter security margins. Large
language models offer natural-
language decision support, but their
hallucinations, uncontrolled tool use, and weak traceability conflict with control room requirements. This paper presents a governance-aware agentic digital twin for transmission grid control rooms. The large
language model only selects and parameterizes whitelisted analysis tools, and every proposed action passes through a governance layer that the model cannot bypass. The layer enforces four rules on every run. Only whitelisted tools execute. No run exceeds its step budget. No action with side effects executes without explicit operator approval. Every number in an answer is rendered by the layer from backend results with its unit, variable, and time. The rules are checked on a persistent audit trail for every run of a released 118-task
benchmark, which covers analytics, simulation, multi-step workflows, and twelve families of
adversarial inputs on a digital twin of the Greek transmission network. Across 590 runs of the primary model, tool selection reaches 96.5% and task success 93.7%, and all four rules hold without exception. In a separate three-repetition study across four large
language models, 1416 runs in total, the rules again hold on every run, with an approximate 95% lower bound of 99.8%. Removing the layer makes the same model execute all 45 approval-requiring runs without authorization and leaves only 39.2% of its answers with backend-supported numbers. Enforcement costs 12 to 16 milliseconds per request.
Read the full abstract ▾
Artificial Intelligence
Goal-driven Variant Categorization
Daniel Calegari and Daniel Amyot
Process discovery rarely yields a single coherent process structure. For analysis, a common step is to cluster process variants based on structural similarity and then assign business meaning to the resulting groups. Since these partitions are not derived from the organization's goals, analysts must manually interpret and consolidate variants into business-meaningful categories. This judgment-intensive step becomes increasingly difficult as the number and complexity of variants grow. In this paper, we propose a goal-driven approach to variant categorization that reverses this workflow. We first author an organization's goal model that predefines the categorization axis. Each variant is transformed into a textual narrative describing its behavior, and a Large
Language Model (LLM) interprets it in the context of the goal model and assigns the variant to the most appropriate category. LLM-based semantic
reasoning connects low-level process behavior with analyst-defined business goals. We instantiate this approach end-to-end and evaluate it on three public logs differing substantially in scale and behavioral diversity. Goal-model guidance yields partitions that differ from those produced by unguided induction and respond to controlled edits to the declared alternatives, at the cost of authoring a goal model.
Read the full abstract ▾
Earth and Planetary Astrophysics
Tectonics on early Earth driven by Earth's changing shape during tidal recession of the Moon
Simon J. Lock
Few fragments of Earth's earliest crust have survived, and so this formative period of our
planet's history is shrouded in mystery. To leverage the limited data available, it is important to understand all the processes that could have shaped our young
planet. Here, I explore a previously neglected aspect of early-Earth dynamics: rotation. Earth after the Moon-forming impact was rapidly rotating with its rotation rate slowing significantly during the first tens of millions of years after the impact as the Moon tidally receded. Using planetary structure and
tidal evolution calculations, I show that the initially distorted structure of Earth would likely have led to significant latitudinal variations in primary crustal thickness. Further, the subsequent changes in Earth's shape in response to its slowing rotation rate could have driven extensive tectonic activity and modulated the response of the crust to other forces. There was extension of the crust in polar regions and convergence in the equatorial regions at rates potentially comparable to those forming the Himalayas today in high-angular momentum scenarios. Depending on the rheology of the earliest crust, such motions could have led to burial of hydrated and weathered surface material, extensive volcanism, and degassing of Earth's interior. Phase changes due to increases in internal pressure could have led to delamination of parts of the crust, producing early mantle heterogeneities. Hydrated secondary melting of the crust and the production of evolved magmas could explain the emergence of a diversity of rock types within tens of millions of years of Earth's formation.
Read the full abstract ▾
Solar and Stellar Astrophysics
The "Other" Centuries-Long Record of Solar Magnetic Activity Cycles: Lessons From the Stars
Jason T. Wright, Travis S. Metcalfe, Kristof Petrovay et al. (+1 more)
Studies of space
climate and
solar magnetic activity and studies of
stellar dynamos and
magnetic cycles are separated by discipline and by the quality of the data we have to work with: the Sun gives us detailed knowledge of a single
star, while the
stars give us comparatively sparse data across a range of (sometimes poorly known) masses and ages. This review, a collaboration among scientists with expertise in
stellar astronomy, space
climate, and
solar physics, seeks to provide a guide to what the Sun's and
stars'
magnetic activity records can tell us about how
stellar dynamos operate on decadal, millennial, and
stellar-evolutionary timescales. It reviews the historical
Solar sunspot record, terrestrial cosmogenic isotope records, and decades-long chromospheric activity measurements of dozens of sun-like
stars across a range of masses, ages, and
metallicities near the
solar value. Particular emphasis is placed on the Maunder minimum, and the hunt for similar events in the
stellar record, especially the recently discovered grand minimum event ongoing in the
star HD 166620. These records show that
solar-like dynamos are a common feature of Sun-like
stars, and
magnetic grand minimum events akin to the Maunder minimum might be very rare, with only one identified in 3,000
star-years of
magnetic activity records. Together,
solar and
stellar dynamo studies are bringing a still-fuzzy picture into focus in which
stellar magnetic activity weakens as
stars spin down over billions of years towards a critical Rossby number near the
solar value, at which point
stellar cycles begin to be frustrated. These periods of frustration manifest as
magnetic grand minima, and increase in frequency until they become permanent and
stars enter a period of very low and constant
magnetic activity.
Read the full abstract ▾
All authors Jason T. Wright, Travis S. Metcalfe, Kristof Petrovay, Hisashi Hayakawa
Materials Science
Revealing epitaxial relationships at Ga$_2$O$_3$ interfaces with p-type oxides
Anna Sacchi, Krishna Acharya, Michelle A. Smeaton et al. (+5 more)
p-type
oxide contact layers such as Cr$_2$O$_3$ and NiO are attracting increasing interest in pn-heterojunctions with n-type monoclinic $\beta$-Ga$_2$O$_3$ for high-power electronic devices and other extreme environment applications. However, scientific understanding of their epitaxial relationships remains incomplete. In this work we investigate the epitaxial relation of Cr$_2$O$_3$ and NiO layers to (001) and ($\bar{2}$01) out-of-plane oriented Ga$_2$O$_3$ substrates. Surprisingly, we find that, for the most commercially relevant (001)-orientation of the Ga$_2$O$_3$ substrate, the epitaxial relationships are Cr$_2$O$_3$ (0001) and NiO (111) $\parallel$ Ga$_2$O$_3$ (101), both at the non-intuitive $\chi$ = 22.5 $^\circ$ angle with respect to the substrate normal. We explain this unusual discovery by the interfacial atomistic bonding dominated by oxygen sublattice equivalence of these Cr$_2$O$_3$ and NiO polar surface orientations to the tilted Ga$_2$O$_3$ (101), rather than Ga$_2$O$_3$ (001) substrate surface planes. Furthermore, we assign the in-plane orientation for Cr$_2$O$_3$ on ($\bar{2}$01)-oriented Ga$_2$O$_3$ as: Cr$_2$O$_3$ $[12\bar{3}0]$ $\parallel$ Ga$_2$O$_3$ $[010]$ with two in-plane rotational domains. Interface modeling confirms the in-plane orientation for Cr$_2$O$_3$/$(\bar{2}01)$ Ga$_2$O$_3$ and shows that strained O-terminated Ga$_2$O$_3$ $(\bar{2}01)$ surfaces have the lowest interfacial energy with Cr$_2$O$_3$ (0001). Beyond establishing the specific epitaxial relationships for Cr$_2$O$_3$ and NiO on Ga$_2$O$_3$, this work provides a systematic methodology for the unambiguous structural characterization of heterointerfaces involving materials with markedly different
crystal symmetries.
Read the full abstract ▾
All authors Anna Sacchi, Krishna Acharya, Michelle A. Smeaton, Renae N. Gannon, Vladan Stevanovic, Steven R. Spurgeon, M. Brooks Tellekamp, Andriy Zakutayev
Machine Learning
Efficient Mixture-of-Experts with Speculative Decoding via Expert Coactivation
Kumari Nishu, Han-Byul Kim, Santosh Chilkunda et al. (+13 more)
Mixture-of-Experts (MoE) models are increasingly deployed alongside Speculative Decoding (SD) to accelerate inference, but combining the two is challenging. SD improves the inference speed of dense models by verifying groups of
tokens in parallel. However, the inference speedup for SD with MoEs depends heavily on the number of
tokens being verified. Using more verification
tokens results in more experts being transferred from DRAM to the Neural Processing Unit (NPU), which increases the
memory transfer cost. This negatively impacts model runtime, as
memory transfer is typically the bottleneck in inference. In this work, we investigate the impact of MoE router design during training on the speed of MoEs with SD. We find that routers with high degrees of expert coactivation result in much faster runtimes, mitigating the impact of using more verification
tokens. Motivated by this observation, we assess the impact of various router design choices on expert coactivation and runtime using billion-parameter
transformer models. We find that combining a global load-balancing loss, shared experts, a consistency loss, and an autoregressive expert selection mechanism during training results in significantly stronger expert coactivation. This increased coactivation translates into higher overall runtime throughput: our exploration yields a model that improves throughput by 21% over MoE baselines, while maintaining on-par accuracy with the baseline MoE.
Read the full abstract ▾
All authors Kumari Nishu, Han-Byul Kim, Santosh Chilkunda, Maxwell Horton, Arnav Kundu, Mohammad Samragh, Lauren Hannah, Mohammad Sekhavat, Nikhil Bhendawade, Manuel Ciosici, Iman Mirzadeh, Keivan Alizadeh Vahid, David Harrison, Irina Belousova, Mehrdad Farajtabar, Minsik Cho
Cosmology and Nongalactic Astrophysics
Stress-testing the spatially flat Universe: nonparametric spatial curvature determination from DESI DR2
Hannah C. Turner, Suhail Dhawan, Sunny Vagnozzi et al. (+2 more)
We present a nonparametric determination of spatial
curvature using only late-time geometrical probes. Under a Gaussian Process smoothness prior, we jointly constrain the dimensionless expansion history and $\Omega_K$, combining three Type Ia
Supernovae (SNeIa) samples with either DESI DR2 line-of-sight
baryon acoustic oscillations (BAO) or
cosmic chronometers (CC) data. From our baseline SNeIa+BAO analyses we find $\Omega_K=0.219^{+0.110}_{-0.111}$, $\Omega_K=0.330^{+0.098}_{-0.150}$ and $\Omega_K=0.260^{+0.124}_{-0.127}$ for the PantheonPlus, DES Dovekie and Union3 SNeIa samples respectively. These nonparametric results mildly favour an open Universe, supporting recent hints in this direction, and improve on the sensitivity of a previous application of the method by up to a factor of $\sim 2.5$. We find consistent constraints combining SNeIa and CC, whereas restricting SNeIa data to the
redshift range covered by BAO somewhat weakens the constraints and shifts the PantheonPlus central value to $\Omega_K<0$. Our results depend only on late-time relative distances and expansion rates, making them insensitive to the absolute distance scale calibration, while relying neither on
Cosmic Microwave Background data nor on a parametric dark energy model.
Read the full abstract ▾
All authors Hannah C. Turner, Suhail Dhawan, Sunny Vagnozzi, Alice Townsend, Peter Massey
Algebraic Geometry
Local monodromy, Behrend function and Hilbert scheme of points on $\mathbb{C}^3$
Yunfeng Jiang
We prove a criteria for the Behrend function values on the
Hilbert scheme of $n$-points on $\C^3$ from local monodromy and the Denef–Loeser's Lefschetz trace formula of jet spaces. If the power two, three, or four of the monodromy is unipotent, it reduces respectively to a Hessian parity formula, a cubic-complement formula, or a quartic formula. This gives a Milnor fiber and monodromy based proof of nonconstancy of the Behrend function on $\Hilb^n(\C^3)$ after Jelisiejew-Kool-Schmiermann. We also calculate two ideals in $\Hilb^9(\C^3)$ and $\Hilb^{12}(\C^3)$ so that the Behrend function is not $\pm 1$, answering a question of Ricolfi.
Read the full abstract ▾
Computational Physics
TubeLab: Interactive inverse design of wind instrument bores with hard spectral constraints
Carlo Andrea Rozzi and Andrea Ferroni
We present TubeLab, a browser-based acoustic simulator and inverse design tool for wind instrument bores without tone holes. The direct problem is solved by a Transfer
Matrix Method. Inverse design - finding a bore correction that shifts selected resonance frequencies toward specified musical targets - is formulated as a Tikhonov-regularized least-squares problem whose Jacobian is computed by a single adjoint backward pass per mode. Modes whose frequencies must be preserved are handled by a saddle-point formulation that enforces hard equality constraints regardless of the regularization strength. Resonance frequencies agree with state-of-the-art finite-element calculations to within 1.21 cent over eleven geometries and 132 mode pairs. Experimental validation against tap-tone measurements of 16 numerical control-machined didgeridoos spanning six distinct bore designs yields 9.6 cent RMS deviation over 189 matched mode pairs. Replicate specimens of the same nominal bore disagree with each other by 4.5 cent RMS, so about 36% of the residual variance is fabrication scatter.
Read the full abstract ▾
Materials Science
Mobility-lifetime relation links photodegradation in spin-coated and gravure-printed organic solar cells
Maria Saladina, Paul Weitz, David Holzner et al. (+24 more)
The degradation mechanisms of organic
solar cells (OSCs) have been studied primarily in
spin-coated, laboratory-scale devices, whereas scalable processing modifies the device architecture, active-layer morphology, and underlying charge-transport and recombination properties. Whether these changes also alter how
solar cells degrade remains unclear. Here, we compare
spin-coated and fully roll-to-roll-compatible gravure-printed PM6:Y12
solar cells during $\sim$1000 h of continuous illumination. Despite distinct initial properties and degradation signatures, the loss of power-conversion efficiency systematically follows the mobility-lifetime product $\mu\tau$. Remarkably, ageing of the printed devices increases the recombination lifetime while strongly reducing charge-carrier mobility, showing that a longer lifetime alone does not imply improved device performance. The mobility reduction is accompanied by decreased PM6 lamellar order, whereas the additional open-
circuit voltage loss originates predominantly from increased non-radiative recombination. Dark recovery further reveals a metastable contribution specific to the printed architecture. These results identify the mobility-lifetime product as a unifying physical descriptor for photodegradation, linking ageing-induced microscopic changes to macroscopic performance loss across
spin-coated and scalable printed OSCs.
Read the full abstract ▾
All authors Maria Saladina, Paul Weitz, David Holzner, Chen Wang, Robert Eland, Svitlana Taranenko, Christopher Wöpke, Toni Seiler, Fabian Eller, Doyoung Sun, Alexander Ehm, Bekcy Joseph, Peli̇n Çi̇loğlu, Carmen Tretmans, Andreas Willert, Harald Kuhn, Roderick C. I. Mackenzie, Jan-Frederik Pietschmann, Martin Stoll, Yana Vaynzof, Dietrich R. T. Zahn, Safa Shoaee, Eva M. Herzig, Arved Carl Hübler, Thomas Heumüller, Christoph J. Brabec, Carsten Deibel
Cosmology and Nongalactic Astrophysics
Model-Independent Measurement of Baryon Gas Fractions through Galaxy-Galaxy Lensing and the Kinematic Sunyaev-Zel'dovich Effect
R. Henry Liu, Uroš Seljak, Simone Ferraro et al. (+52 more)
Baryon feedback is a leading source of systematic uncertainty for cosmology from weak
lensing, but measurements of the gas distribution around
galaxies have largely relied on parametric profile models or simulation-calibrated frameworks. We present model independent measurements of the radial gas fraction profile around
galaxies, combining
galaxy-
galaxy lensing and kinematic Sunyaev-Zel'dovich (kSZ) effect. We introduce a method that applies the same radial $\Delta\Sigma$ aperture filter to both the
galaxy-
galaxy lensing shear field and the velocity-weighted kSZ temperature maps; their ratio directly yields $\Delta\Sigma$ filtered gas fraction $f_{\rm gas}(R)$, the ratio of ionized gas to total matter as a function of projected radius. We apply this approach to DESI DR2 Bright
Galaxy Survey (BGS, $\bar{z} \approx 0.26$) and Luminous Red
Galaxy (LRG, $0.4 < z < 1.1$) samples, using ACT DR6 component-separated CMB maps for kSZ, and the HSC Year 3 shear catalog for weak
lensing. After correcting for ACT beam suppression using a simulation-calibrated compensation factor, we detect
baryon depletion relative to the
cosmic mean
baryon fraction at SNR = 17.2 (BGS) and SNR = 13.8 (LRG bin 1). Comparison with six hydrodynamical simulations from the Illustris, IllustrisTNG, SIMBA, and FLAMINGO suites shows that the fiducial FLAMINGO run most closely tracks the observed radial gas distribution in both samples, with Illustris-1 and TNG300-1 bracketing the data from below and above at small radii. We find a mild discrepancy between the kSZ measurements in the South and North
Galactic Caps in one LRG
redshift bin, which may indicate residual systematics. We show that the ratio is robust to splits by
stellar mass and to the removal of close pairs. These results establish $\Delta\Sigma$ filtering of kSZ versus weak
lensing signal as a model-independent probe of the
baryon distribution around
galaxies.
Read the full abstract ▾
All authors R. Henry Liu, Uroš Seljak, Simone Ferraro, Boryana Hadzhiyska, Frank J. Qu, Bernardita Ried Guachalla, Emmanuel Schaan, Jessica Nicole Aguilar, Steven Ahlen, Anton Baleato Lizancos, Florian Beutler, Davide Bianchi, David Brooks, Aurelio Carnero Rosell, Francisco Javier Castander, Todd Claybaugh, Andrei Cuceu, Axel de la Macorra, Jaime E. Forero-Romero, Enrique Gaztañaga, Satya Gontcho A Gontcho, Gaston Gutierrez, Klaus Honscheid, Dragan Huterer, Mustapha Ishak, Stephanie Juneau, Tanveer Karim, David Kirkby, Anthony Kremin, Ofer Lahav, Martin Landriau, Laurent Le Guillou, Martine Lokken, Marc Manera, Aaron Meisner, Ramon Miquel, John Moustakas, Seshadri Nadathur, Jeffrey A. Newman, Hernan Enrique Noriega, Will Percival, Ignasi Pérez-Ràfols, Francisco Prada, Corentin Ravoux, Graziano Rossi, Lado Samushia, Eusebio Sanchez, Christoph Saulder, David Schlegel, Michael Schubnell, Hee-Jong Seo, Małgorzata Siudek, Gregory Tarlé, Benjamin Alan Weaver, Rongpu Zhou
Combinatorics
Collision Positivity for Three-Variable Symmetric Monomial Inequalities
Jian Sun
Let \(\lambda\succ\gamma\succ\mu\) be equal-degree exponent partitions with at most three parts, and let \begin{equation*} P_{\lambda,\gamma,\mu} = J_\lambda+J_\mu-2J_\gamma, \end{equation*} where \(J_\nu\) denotes the symmetric monomial orbit sum associated with \(\nu\). We prove a necessary and sufficient collision criterion for the positivity of this three-point majorization difference in three variables. For nonnegative integer exponent partitions, \begin{equation*} P_{\lambda,\gamma,\mu}(x,y,z)\ge0 \qquad(x,y,z>0) \end{equation*} if and only if \begin{equation*} P_{\lambda,\gamma,\mu}(t,1,1)\ge0 \qquad(t>0). \end{equation*} Thus the positivity of a genuinely three-variable symmetric
polynomial of this form is completely determined by its one-variable restriction to the locus where two variables coincide. The
theorem gives a uniform and effectively checkable criterion for an infinite class of symmetric
polynomial inequalities. For a fixed integer chain, the global three-variable problem is reduced to a single univariate
polynomial inequality, which can often be verified exactly by factorization, Sturm's
theorem, or other one-variable methods. The criterion extends well beyond the classical Schur family, produces explicit inequalities outside the Schur pattern, and yields genuine strengthenings and refinements of Schur's inequality.
Read the full abstract ▾
Systems and Control
Expressive Power of WSTL Formulas for Learning to Rank
Ruya Karagulle, Gustavo A. Cardona, Necmiye Ozay et al. (+1 more)
Weighted Signal Temporal Logic (WSTL) is increasingly used as a scoring function in learning-to-rank problems of trajectories with safety guarantees, where its weighted quantitative semantics serve as a parametrized utility function. Despite the growing interest, prior work only assumes the expressiveness of WSTL formulas and empirically demonstrates its utility in capturing diverse preferences. This work focuses on the correctness of this assumption and asks whether using WSTL formulas as scoring functions is theoretically justified. We formalize two concepts: first, rank-realizability, which asks whether all rankings of a given signal set are achievable by varying weights, and, second, rank-capacity, the maximum signal set size for which the formula is rank-realizable. We propose a Mixed-Integer Linear Program to decide rank-realizability, and derive constructive lower bounds for rank-capacity. Experiments on a robotic
navigation task show that while a practical WSTL specification may fail to be rank-realizable on a set with similar trajectories, its rank-capacity exceeds the size of the trajectory set. Analysis of Boolean-equivalent formulas reveals that formula structure affects expressivity and that rank-capacity can be increased without altering qualitative semantics.
Read the full abstract ▾
All authors Ruya Karagulle, Gustavo A. Cardona, Necmiye Ozay, Cristian-Ioan Vasile
Computation and Language
Toward Personalized Sleep Guidance from Wearable Data Using Language Models
Yusheng Tan, Running Zhao, Sofia Angel et al. (+14 more)
Sleep monitoring using wearable data has shown promise for personal health, yet large
language model (LLM)-based summarization and question answering remain insufficient for personalized sleep guidance. Training specialized models, however, often requires costly expert annotation. Moreover,
privacy and accessibility concerns motivate lightweight, local deployment for end users. We present a two-stage framework to address these challenges. Specifically, in Stage~1, a multi-
agent LLM pipeline reasons structured sleep guidance from unannotated wearable records, enabling scalable
dataset construction. Stage~2 distills guidance
reasoning trajectories into small
language models (SLMs) through supervised fine-tuning and integrates a training-free Best-of-$N$ selection strategy to enhance inference. Experimental results demonstrate our method outperforms commercial general and medical LLMs and open-source models. Human evaluation further supports the quality of the generated guidance and the feasibility of personalized sleep guidance with SLMs.
Read the full abstract ▾
All authors Yusheng Tan, Running Zhao, Sofia Angel, Ninghui Hao, Ash Arian, Nikita N. Dulin, Jay Lin, Ou Zhu, Faiza Shaik, Xinxing Yang, Bonnie W. Leung, Katie Roster, Arlene Ruiz de Luzuriaga, Kenneth Lee, Alejandra Lastra, Habibul Ahsan, Guihong Wan
Robotics
VLPSA: Vision-Language-Poisson-Safe Actions for Full-Body Safety of Learned Policies
Meg Wilkinson, Emily Fourney, Joel W. Burdick et al. (+1 more)
Vision-
language-action (VLA) models enable increasingly general-purpose robotic manipulation, but such learned policies do not provide safety guarantees for collision avoidance—especially in environments outside of training distributions. This work presents
Vision-
Language-Poisson-Safe Actions (VLPSA), a safety filtering framework that provides full-body safety for VLA policies in cluttered and dynamic environments without retraining. VLPSA synthesizes Poisson Safety Functions (PSF) online from perception data, yielding a Control Barrier Function (CBF) that is enforced through a CBF-QP safety filter over the full body and any grasped object, treated as an extension of the final
robot link. To enable real-time deployment while maintaining fine spatial resolution in critical task regions, VLPSA combines dual resolutions of this PSF using Boolean CBF compositions. We evaluate VLPSA on SafeLIBERO against safety-filtering baselines, where it achieves the highest collision avoidance rate among the evaluated methods, increasing collision avoidance from 23.1% for the base $\pi_{0.5}$ policy to 91.2% while surpassing its task success rate. We further deploy VLPSA on a Franka FR3 in cluttered scenes with dynamic obstacles and human interference, demonstrating real-time full-body safety during manipulation tasks.
Read the full abstract ▾
All authors Meg Wilkinson, Emily Fourney, Joel W. Burdick, Aaron D. Ames
Materials Science
Laser-Ablated Au Electrodes with Preferred Orientation and Flat Interface on Oxides
Ambrose Seo, Richard Lai, Sujan Shrestha et al. (+5 more)
Gold (Au) thin-film electrodes deposited on
oxides by pulsed
laser deposition (PLD) exhibit distinctive structural properties compared with thermally evaporated Au films. While thermally evaporated Au forms polycrystalline films with randomly oriented grains, PLD-grown Au films show a strong preferred [111] orientation perpendicular to the substrate surface, likely due to the energetic nature of the
laser-ablated Au plume. High-resolution transmission electron microscopy reveals atomically flat interfaces between the PLD-grown Au electrodes and epitaxial
oxide thin films, separated by nanometer-scale gaps comparable to interlayer spacings in van der Waals materials. The highly ordered microstructure and flat interfacial morphology result in enhanced adhesion of the Au films to
oxide surfaces. These results demonstrate that PLD can produce both high-quality Au electrodes and
oxide thin films within a single deposition platform, offering improved interfacial control and potential benefits for a wide range of
oxide electronic and functional devices.
Read the full abstract ▾
All authors Ambrose Seo, Richard Lai, Sujan Shrestha, Tina Tong, Joseph W. Brill, Menglin Zhu, Chris Chae, Jinwoo Hwang
Numerical Analysis
Update to: Spectral sets: Numerical range and beyond
Michel Crouzeix and Anne Greenbaum
We use the results of Lorist and Schwenninger, A solution to Crouzeix's
conjecture, arXiv:2608.03841v2, https://arxiv.org/abs/2608.03841, to update the 2019 paper of Crouzeix and Greenbaum, Spectral sets: Numerical range and beyond, SIAM J.~
Matrix Anal.~Appl., 40(3):1087-1101. For all regions considered in the 2019 paper, we are able to show that the region is a $K$-spectral set with a bound on $K$ that is smaller than that established in the 2019 paper. In particular, we show that various annular regions are $2$-spectral sets and that a more general convex region with a circular hole or cutout is a $4$-spectral set.
Read the full abstract ▾
Differential Geometry
A Differential-Geometric Framework for Risk-Optimal Asset Reallocation
Georgios Leventidis and Evangelos Melas
We develop a differential-
geometric framework for risk-optimal portfolio reallocation. A long-only portfolio is represented as a point of the probability simplex, endowed with a positive-definite Riemannian metric combining
market covariance risk with position-dependent concentration risk. The cumulative risk of a rebalancing trajectory is identified with its Riemannian length, so the least-risk transition between a current allocation and a Markowitz target is a geodesic. We compare this route with direct linear rebalancing and projected gradient ascent. When only
market covariance risk is priced, the metric is constant and flat, and the geodesic is exactly the straight-line path. Once position-dependent risk is introduced, the geometry becomes curved and the geodesic weakly dominates competing paths with the same endpoints. A Fisher-Rao concentration term produces modest but systematic savings, while an endogenous crowding metric creates non-convex risk ridges that geodesics can bypass through temporary diversification. Numerical experiments, Monte Carlo transitions, and regression analysis show that the largest gains occur when the direct path crosses strongly crowded regions. The framework provides a general
geometric formulation of transition management and can accommodate richer risk metrics and transaction-cost structures.
Read the full abstract ▾
Analysis of PDEs
On the Zero-Mass Problem of Bandle, Levine, and Zhang
Mohamed Jleli and Bessem Samet
We study the critical behavior of the inhomogeneous semilinear heat equation \[ u_t-\Delta u=|u|^p+w(x) \qquad\text{in }(0,\infty)\times\mathbb{R}^N, \] where \(N\geq3\), \(p>1\), and \(w\in L^1(\mathbb{R}^N)\) is nontrivial and has zero total mass: \[ \int_{\mathbb{R}^N}w(x)\,dx=0. \] This zero-mass case was posed as an open problem by Bandle, Levine, and Zhang in 2000 and has remained unresolved since then. We provide a complete answer to this problem by proving nonexistence of global weak solutions for \[ 1<p\leq\frac{N}{N-2}. \] Combined with the known supercritical existence result for sufficiently small data, this shows that the critical exponent separating the nonexistence and existence regimes remains \[ p_c=\frac{N}{N-2}, \] the same as in the positive-mass case.
Read the full abstract ▾
Astrophysics of Galaxies
The ALMA View of High-Redshift Galaxy Formation
Renske Smit and Rebecca A.A. Bowler
The advent of routine operations with the Atacama Large Millimeter/submillimeter Array (ALMA) in the last decade has led to a revolution in the direct study of the interstellar medium (ISM) in `normal' high-
redshift galaxies in the rest-frame far-infrared (FIR). This review summarizes the observational literature on z>6.5 sources observed with ALMA and NOEMA (NOrthern Extended Millimeter Array). The main findings are as follows: Cool gas consistently scales with
star-formation in a wide range of
galaxy environments, similar to local relations, suggesting fundamental processes of
star-formation stay relatively constant over 13 billion years of
cosmic time. Significant metal enrichment is present in
galaxies just 300-400 million years after the Big Bang. There is a trend to higher obscured fraction of
star formation with
stellar mass and SFR already in place at z~7, with measurements of the
cosmic SFR density showing that >10 percent of the SFR density is obscured at this epoch. The estimated
dust masses compared to the
stellar mass suggest that rapid
dust enrichment occurs, likely from
supernova with little
dust destruction and/or rapid growth in the ISM, which is in agreement with maximal predictions from models. In individual sources cold disks are already in existence, with their gas disks being more extended and smooth than their observed
stellar counterparts.
Read the full abstract ▾
Metric Geometry
Isoperimetric Inequalities in the $\ell_{\infty}$ cube and torus
Ben Baker, Joseph Briggs, Manuel Fernandez V et al. (+1 more)
We prove an exact isoperimetric inequality for cubes and
tori under the $\ell_{\infty} metric. As a corollary, we recover the celebrated grid edge-isoperimetric inequality of Bollobás-Leader.
Read the full abstract ▾
All authors Ben Baker, Joseph Briggs, Manuel Fernandez V, Chris Wells
Computation and Language
Apollo Restore: A Foundation LLM for Historical Greek Optimized for Fill-in-the-Middle Restoration of Ancient Greek Texts
Hope McGovern, Anna Dolganov, Samuel Belkadi et al. (+3 more)
We present Apollo Restore, a 24-billion-parameter large
language model for restoring lacunae—physical gaps—in fragmentary Ancient Greek texts. Fine-tuned from Mistral Small with a fill-in-the-middle objective, Apollo Restore reconstructs missing spans without requiring oracle knowledge of their length. To our knowledge, it is the first large-scale decoder model for historical Greek, and the first for any ancient Mediterranean
language. Evaluated as in prior work, on short gaps of up to ten characters, Apollo Restore places the correct restoration among its top twenty candidates for 80.6%/54.6%/61.0% of documentary-papyrus, literary-papyrus, and stone-inscription lacunae, exceeding the strongest published models by $1.6\times$/$2.6\times$/$1.4\times$. Prior evaluation protocols, however, inflate scores through a bias toward trivially short gaps; under a length-balanced metric Apollo Restore's advantage over the strongest published models grows to $2.3\times$/$3.5\times$/$1.6\times$ and degrades gracefully, even given incorrect length hints. In a blind study, 20 expert papyrologists, epigraphists, and philologists strongly preferred Apollo Restore to the strongest baseline and judged its performance at least as good as human restorations in 77% of cases. Apollo Restore also improves the published reading of PHerc. 1667—a papyrus roll carbonised in the eruption of Vesuvius in 79 CE and digitally unrolled and edited after Apollo Restore's training data was compiled. Apollo Restore is an output of the Decoding Antiquity initiative to build specialized LLMs for historical
languages and manuscripts, led by the Austrian Academy of Sciences.
Read the full abstract ▾
All authors Hope McGovern, Anna Dolganov, Samuel Belkadi, Guillaume Kunsch, Dimitris Vlitas, David A. Smith
Computational Physics
Stochastic consensus dynamics for decentralized decision systems
André L. M. Vilela, Caio B. L. Silva, Kenric P. Nelson et al. (+2 more)
A fundamental mechanism underlying collective phenomena in social, technological, and economic systems is decentralized decision-making, in which the behavior of individual
agents follows from local interactions in the absence of central coordination. Consensus formation is a central feature of such systems, with relevance to blockchain networks, distributed artificial intelligence, autonomous multi-
agent systems, distributed control, and collective decision networks. We investigate consensus formation using a
stochastic consensus model on random networks and examine how initial conditions, network connectivity, and system size shape the emergence of unanimous states. We show that the
stochastic dynamics strongly amplify small initial majorities, progressively suppress the competing state, and drive the system toward a predictable collective outcome. Network connectivity primarily controls the efficiency of this process: increasing connectivity accelerates the propagation of local agreement and reduces the likelihood that fluctuations reverse the initially dominant state, although these gains gradually saturate in highly connected networks. System size produces a complementary effect. Larger networks require more individual updates to reach unanimity, but they are also increasingly reliable in selecting the state favored by the initial majority. Finite-size analysis shows that the range of initial conditions associated with uncertain outcomes becomes progressively narrower as the network grows, decreasing approximately with the inverse square root of the system size. These results reveal a collective amplification mechanism by which weak initial asymmetries become increasingly decisive in large decentralized networks, and they provide a simple framework for understanding the efficiency, predictability, and reliability of consensus formation in distributed decision systems.
Read the full abstract ▾
All authors André L. M. Vilela, Caio B. L. Silva, Kenric P. Nelson, Emilio Cobanera, Gaogao Dong
Commutative Algebra
The Bourbaki Degree of Line Arrangements
Abbas Nasrollah Nejad and Samira Nouri
We study the Bourbaki degree of line arrangements in the projective plane and its interaction with the intersection
lattice and the syzygies of the gradient ideal. We show that the Bourbaki degree is obtained by evaluating the reduced characteristic
polynomial at the initial degree of the first syzygy module. This leads to a sharp upper bound and to refinements involving the two largest intersection multiplicities, with consequences for the global Tjurina number, the freeness defect, and Terao's
conjecture, including an improvement of Dimca's numerical criterion when the smaller exponent is at least five. We also establish addition–deletion formulas and show that the defect from the sharp bound is monotone under line addition. Finally, we prove that the Bourbaki degree is determined by the intersection
lattice for arrangements with at most eight lines, while examples with isomorphic intersection
lattices and different Bourbaki degrees exist for every number of lines at least nine.
Read the full abstract ▾
Computation and Language
Vox-Infinity: Benchmarking the Limits of Long-Context Spoken Language Models
Xize Cheng, Wenxu Jia, Chenyuhao Wen et al. (+4 more)
Long-context understanding remains a fundamental challenge for large
language models, as excessively long inputs often lead models to forget salient information. This issue is even more pronounced in the
speech domain, where audio, as a low-compression modality, requires substantially more
embeddings than text to preserve both semantic content and acoustic cues. To address this challenge, we introduce Vox-Infinity, the first
benchmark specifically designed to evaluate long-context understanding in spoken
language models. Vox-Infinity systematically extends audio history along two dimensions: turn count and turn duration. It covers a diverse range of representative scenarios with varying interaction structures and semantic complexity. Crucially, Vox-Infinity provides explicit answer-provenance annotations and organizes samples according to the amount of historical context required to resolve each query, enabling precise and length-aware evaluation. Extensive evaluations of seven representative spoken
language models reveal a clear overall recency effect: models generally achieve higher accuracy when answer-supporting evidence is closer to the query, but struggle to retrieve and use evidence located farther back in the dialogue history. Cases and
datasets are available at https://vox-infinity.github.io.
Read the full abstract ▾
All authors Xize Cheng, Wenxu Jia, Chenyuhao Wen, Dongjie Fu, Zehan Wang, Xinyu Zhang, Tao Jin
History and Overview of Mathematics
Mathematical folklore
Govind Menon and Akshay Venkatesh
We present reflections on an experimental seminar studying the human and cultural roots of modern mathematics.
Nuclear Theory
Evidence for a Continuous Hadron–Quark Transition in Cold Dense Matter
Yong-Jia Huang and Bikai Gao
Statistical evidence for a continuous
hadron-
quark transition is found in this work. Confronting microscopic descriptions connecting the Parity Doublet Model and the Nambu–Jona-Lasinio model with observationally constrained non-parametric equations of state,
Bayesian model comparison decisively favors a boundary-free crossover over the conventional first-order Maxwell construction ($\Delta\ln\mathcal{Z} \sim +7.20$) and fixed-boundary crossover ($\Delta\ln\mathcal{Z} \sim +5.16 $). The preferred crossover decouples the intermediate stiffening from the intrinsic stiffness of each phase, allowing both sectors to exhibit physical self-consistency. The
chiral-invariant nucleon mass is large, $M_0 = 834^{+28}_{-92}\text{ MeV}$, as expected for a substantial
baryon mass surviving
chiral restoration. While the
quark sector accommodates a small pairing gap $\Delta_{\rm CFL} = 88^{+83}_{-59}\text{ MeV}$ consistent with perturbative QCD limits, avoiding the excessively large gaps $\Delta_{\rm CFL} \gtrsim 200\text{ MeV}$ required by the other constructions. These results demonstrate that a realistic unified description represents a continuous transition deviating substantially from isolated effective models of
hadrons and
quarks. Exploring genuine phase boundaries is therefore necessary through the emergence of spinodal instabilities within unified frameworks.
Read the full abstract ▾
Astrophysics of Galaxies
LoTSS of dual AGN: enhanced detectability and prospects for a systematic search
Q. D'Amato, F. Mannucci, E. De Rubeis et al. (+25 more)
Context: Dual active
galactic nuclei (DAGN) - systems of two accreting supermassive black holes at kpc-scale separations - trace
galaxy mergers and are the progenitors of gravitational-wave sources targeted by
pulsar timing arrays and the upcoming
Laser Interferometer Space
Antenna. However, their
radio properties remain poorly constrained, particularly at high
redshift. Aims: We demonstrate the detectability of high-
redshift DAGN at 144 MHz through sub-arcsec imaging of the LOFAR Two-metre Sky Survey (LoTSS) and characterize their low-frequency
radio properties. Methods: We cross-matched 92 spectroscopically confirmed DAGN (z > 0.3, projected separation r_p < 30 kpc) with the LoTSS DR2 catalog and reprocessed international LOFAR
telescope (ILT) data for three systems (0.59 < z < 2.39), achieving ~0.3 arcsec resolution at 144 MHz. Multi-frequency analysis and brightness-temperature diagnostics were used to investigate the
radio emission. Results: We obtained high-quality ILT images for all three systems. In one source at z = 1.49, both AGN are detected at >5 sigma, making it one of the highest-
redshift and faintest
radio-detected DAGN. Brightness temperatures confirm AGN-powered emission even in sub-mJy sources. In the other two systems only one AGN is detected; for one, the new data constrain the low-frequency spectrum of a compact steep-spectrum source, indicating that its morphology is driven by jet-ISM interaction rather than the companion
galaxy. For the 92 objects sample, the LoTSS detection rate is ~24%, significantly higher than for matched single AGN (~9-10%), suggesting an enhanced detectability of DAGN in the LoTSS. Conclusions: Low-frequency sub-arcsec
radio observations provide an efficient,
dust-unbiased method to detect and characterize DAGN. This pilot study establishes a
benchmark for the forthcoming all-sky ~0.3 arcsec iLoTSS survey and future SKA observations.
Read the full abstract ▾
All authors Q. D'Amato, F. Mannucci, E. De Rubeis, I. Prandoni, A. De Rosa, M. Scialpi, M. Kunert-Bajraszewska, A. Janiuk, A. Krauze, L. Battistini, E. Bertola, C. Bracci, A. Chakraborty, M. Ceci, C. Cicone, G. Cresci, A. Marconi, C. Marconcini, E. Nardini, M. Parvatikar, A. Peca, K. Rubinur, P. Severgnini, C. Spingola, L. Ulivi, G. Venturi, C. Vignali, M. V. Zanchettin
Fluid Dynamics
Applications of a moving surface drag model: From Stokes waves to offshore wind farms
Manuel Ayala, Zein Sadek, Raul. B Cal et al. (+2 more)
Offshore wind farms operate within the marine atmospheric boundary layer, where ocean waves can modify air-sea momentum transfer, turbine inflow, and wake recovery. Wave-resolved simulations can capture phase-dependent wind–wave interactions, but their computational cost and implementation complexity limit their use in large atmospheric and wind-farm domains. This study describes applications and tests of the recently proposed moving surface drag (MOSD) model, a flat-bottom wall-stress formulation for horizontally resolved moving waves, to a sequence of increasingly complex wind–wave and offshore wind-energy applications. The model is first evaluated for turbulent flow over Stokes-like waves using laboratory measurements, where it reproduces mean velocity profiles and captures the qualitative structure of wave-induced motions, although their amplitude is underpredicted. MOSD is then applied to broadband wave spectra fields and offshore wind-farm boundary layers, showing good agreement with mean velocity profiles from wave-resolved simulations across different wave ages, grid resolutions, and turbine spacings. Finally, exploratory offshore wind-farm simulations using MOSD under conventionally neutral conditions demonstrate its application at wind-farm scale. Large-scale waves leave detectable signatures in near-surface velocity spectra, but these signatures decay with height and do not produce clear peaks in temporal spectra of aggregate turbine power. Overall, the results demonstrate that the MOSD model provides a tractable framework for incorporating resolved wave effects in wind-wave and offshore wind-farm simulations.
Read the full abstract ▾
All authors Manuel Ayala, Zein Sadek, Raul. B Cal, Dennice. F Gayme, Charles Meneveau