Computer Vision
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Wednesday 9 September
The pedestrian crossing intention task involves predicting whether pedestrians are likely to cross the road from the point of view of an autonomous vehicle. We introduce TrajFusionNet+, a novel transformer-based model for pedestrian crossing intention prediction. TrajFusionNet+ combines sequential and visual representations of pedestrian trajectory with a graph-based representation of the scene context in order to predict pedestrian crossing intention. The proposed architecture builds upon our previous model, TrajFusionNet, and comprises three branches: a Sequence Attention Module (SAM), which processes a sequential representation of past and predicted pedestrian trajectories; a Visual Attention Module (VAM), which utilizes a visual representation of the pedestrian trajectories by overlaying observed and predicted bounding boxes onto scene images; and a Graph Attention Module (GAM), which extracts pedestrian-centric graphs from segmented scene images and captures the relational dependencies between pedestrians and traffic elements. TrajFusionNet+ achieves improved state-of-the-art performance on the two most widely used pedestrian crossing intention datasets, PIE and JAAD. Furthermore, we introduce a new evaluation protocol in which models are trained jointly on the PIE and JAAD datasets but evaluated separately on each. Under this setting, TrajFusionNet+ demonstrates superior generalization compared to existing approaches.
High Energy Physics (Theory)
Alternative Boundary Conditions in Asymptotically Anti-de Sitter Spacetimes
Anti-de Sitter (AdS) spacetimes are not globally hyperbolic, so boundary conditions are generally needed for well-defined dynamics. Ishibashi and Wald (IW) determined boundary conditions yielding well-posed dynamics with conserved positive energy for scalar, electromagnetic, and linearized gravitational fields, using nonlocal potentials and individual spherical harmonic modes. In this paper, we determine which IW conditions are local in the primary fields and which are AdS-invariant. In four spacetime dimensions, the standard AdS-invariant boundary conditions set the rescaled magnetic field to zero for electromagnetism and the rescaled magnetic Weyl tensor to zero for Einstein gravity on the conformal boundary. We generalize these standard boundary conditions by setting arbitrary linear combinations of the electric and magnetic fields, or the corresponding Weyl tensors, to zero, and we conjecture that these exhaust the local, AdS-invariant, conservative boundary conditions in AdS$_{1,3}$. Boundary conditions with non-AdS-invariance and in different spacetime dimensions are also studied. The AdS-invariant conditions in AdS$_{1,3}$ extend to Yang-Mills theory and nonlinear gravity in a general Fefferman-Graham setting. We construct phase spaces for these general boundary conditions. The covariant phase spaces for general conditions are naturally associated with Lagrangians containing Pontryagin terms and, in gravity, a Gauss-Bonnet term. In gravity, except for the standard boundary condition, no nontrivial asymptotic symmetries exist and all conserved charges such as ADM mass vanish identically, as in a closed universe. Spacetimes with Killing fields including AdS then exhibit linearization instability in the sense of Fischer and Marsden and Moncrief. For nonstandard boundary conditions, the modified symplectic structure yields new expressions for the entropy formula of black holes.
Read the full abstract
High Energy Physics (Experiment)
The NOvA Test Beam Experiment
NOvA is a long-baseline neutrino oscillation experiment designed to study the neutrino mixing parameters, mass ordering, and CP violation in the lepton sector. A key component of the success of the experiment is a robust understanding of the systematic uncertainties associated with detector response and calibration. To address this, NOvA deployed a Test Beam experiment at the Fermilab Test Beam Facility, which collected data from April 2019 through July 2022. The NOvA Test Beam experiment used a 30-ton segmented liquid scintillator detector functionally identical to the NOvA Near and Far Detectors to analyze tagged particles produced from p-Cu collisions, with instrumentation capable of selecting and identifying electrons, muons, pions, kaons, and protons with momentum ranging from 0.4-1.5GeV/c. Analysis of the collected data provides a better understanding of the largest systematic uncertainties impacting NOvA's analyses, which include the detector response, energy calibration, and hadronic and electromagnetic energy resolutions.
Read the full abstract
Astrophysics of Galaxies
The THRILS Factor: Investigating the properties of Little Red Dots (LRDs) at 3<z<6 with JWST/NIRSpec
JWST has uncovered a class of objects called LRDs, whose nature is still widely debated. In this work, we present a comprehensive spectroscopic analysis of nine LRDs in the Extended Groth Strip (EGS) field studied as part of THRILS and C3PO, both JWST Cycle 3 programs. These targets, photometrically selected based on their compact red appearance, are observed with deep spectroscopic exposures ($\geq8$ hours), enabling robust detections of broad Balmer lines, He I emission, and other spectral features characteristic of AGN activity. Using the [SII] $\lambda\lambda6716,6731$ doublet, we find electron densities ($n_e$) between $2.33 < \log (n_e) < 2.97 \ \mathrm{cm^{-3}}$, comparable to those in narrow line regions (NLR) of local AGN and high-$z$ galaxies. The spectroscopic depth further enables detailed characterization of broad Balmer line profiles. We fit both Gaussian and convolved exponential models to each source and find that five LRDs are statistically better described by the latter model. We measure optical depths $\tau_{\rm sc} = 0.56-0.91$, scattering fractions $f_{\rm SC} = 0.40-0.82$ which correspond to column densities log(N$_e$) $\sim$ 23.93-24.14 cm$^{-2}$, and covering fractions $c_f = 0.43-0.59$. These results indicate a clumpier broad line region (BLR) geometry that deviates from conventional LRD models, which predict covering fractions close to unity. Furthermore, these Compton-thick gas columns may explain the X-ray weakness of LRDs. We also find that THRILS LRDs are narrow-line dominated compared to literature AGN-dominated LRDs, and show that exponential profile fitting corrects for systematic overestimation of black hole masses from Gaussian-based measurements.
Read the full abstract
Mathematical Physics
A Time-Frequency Framework for GKP Codes
We develop a time–frequency framework for lattice GKP codes in which ideal codewords are realized in the modulation space $M^\infty$ and identified, through a vector-valued Zak transform, with a finite logical fibre over the continuous syndrome torus. Multi-window Gabor analysis then represents the logical vector by a finite block of adjoint-lattice coefficients. We prove that the normalized block map is an isometry, obtain an explicit recovering projection, and derive stable logical reconstruction. We further construct normalizable GKP approximants as lattice-envelope Gabor multipliers and establish weak-$*$ convergence and asymptotically isometric encoding. Finally, we recover displacement syndromes from phase relations between translated coefficient blocks and quantify their stability under additive perturbations.
Read the full abstract
Computer Vision
How Much Velocity Does Off-Ball Space Value Need? A Broadcast-Viewport Benchmark
Velocity-aware pitch control is standard, but under a broadcast viewport half the players are off screen and on-screen velocities come from a drifting calibration. We ask at which layer of broadcast off-ball analysis velocity changes the answer. Inheriting our off-screen imputation protocol (three Metrica matches, 44 m viewport, block-bootstrap CIs), we score four velocity regimes – none, viewport-legal observed, true-for-visible, true-for-all – against a velocity-aware ground truth at three layers: imputation, the control surface, and team verdicts. Velocity is nearly useless for imputation (-0.2 pp against a 12–14 pp velocity-free surface MAE), first-order for the surface (-1.5 to -1.8 pp, 11–15% of that MAE), and ten times smaller for verdicts (-0.12 to -0.19 pp). The velocity that matters is the visible channel: perfect occluded-player velocity adds 2–6% of the visible gain, and no last-seen decay policy we tested exceeds that. Omitting velocity blurs the surface (per-frame |e| 2.2–2.6 pp) with small time-averaged bias (per cell <=0.4 pp), whereas imputation error is a structured bias against the defending team's deep zone (5–9 pp). At a fixed velocity window, a noise ladder of eleven jitter settings, including sigma_v-matched pairs, is ordered to first order by one velocity-noise axis sigma_v with break-even ~1 m/s; eleven SoccerNet-GSR clips from one match through our pipeline measure sigma_v=1.65 m/s yet recover 24–36% of the benefit: 43% of the variance is frame-common, which the surface tolerates, and the residual is heavy-tailed and clustered, which Gaussian controls matched on component RMS do not reproduce (+0.03 vs. +0.36). The share of velocity-free error that velocity removes grows with viewport width (7% at 36 m, 21% at 60 m): fix imputation on tight shots, velocity on wide ones. Code and logs are released.
Read the full abstract
Systems and Control
Dynamic Centrality Measures for Water Distribution Network Hydraulics
Water distribution networks (WDNs) are susceptible to various failures, including but not limited to human errors, cyber-attacks, and network modifications, necessitating the vulnerability analysis of WDNs. Graph-theoretic centrality measures—as a main class of centrality measures—aim to rank the network components solely based on their influence (i.e., criticality) on the WDN topology in the case of input changes, while overlooking the WDN dynamics. To overcome such a limitation, this paper uses a control-theoretic centrality measure to identify the network's most and least influential pipes in WDNs, by simultaneously incorporating the dynamics and topology of the WDN. First, given a WDN modeled by nonlinear differential-algebraic equations (NDAEs) consisting of transient flow dynamics as differential equation (DE) and conservation of water mass as algebraic equation (AE) and considering the pipe flow rates as states of the state-space (SS) representation, we extract a linearized system modeled by linear ordinary differential equations (LODEs) around the equilibrium (i.e., steady) flow rate vector. Second, treating pipe flow rates as SS nodes (states), we introduce a node centrality-based measure, namely vulnerability vector (VV), to rank the network pipes based on their influence on the dynamics and topology of the WDN in the case of input changes. In particular, the network's most and least influential pipes can be identified through such a centrality-based approach. This enables water engineers to understand the WDN's vulnerability better and effectively prioritize the maintenance and operational efforts on the most influential pipes within the WDN.
Read the full abstract
Materials Science
On the physical origins of switching diversity in Cu-embedded SiO$_x$ memristive devices
Resistive switching devices with sub-stoichiometric SiO$_x$ and pancake-like Cu nanoparticles (Cu-PCs) exhibit distinct macroscopic current-voltage characteristics classified as capacitive or gradual (interface-type switching) and abrupt or resistive (filamentary-type switching), motivating an analysis of the microscopic processes underlying this diversity. It is proposed that the device defect landscape is largely shaped by two charged defect types, mobile oxygen vacancies and immobile Cu-related defects, whose distributions jointly govern interfacial and bulk transport. An effective one-dimensional cloud-in-a-cell simulation framework is employed to reproduce the phenomenological picture of both interface-type and filamentary-type switching by incorporating the dominant coupled ionic and electronic processes underlying these mechanisms. The model includes oxygen-vacancy drift-diffusion, Schottky-limited injection at the metal/oxide interfaces, and bulk trap-assisted transport via Poole-Frenkel conduction, with Cu-PCs near the top interface treated effectively. A simulation-based parametric study varying voltage stress, sweep rate, and oxide thickness is used to examine how these factors rebalance voltage partitioning and the spatiotemporal electric field distribution, thereby altering vacancy redistribution and the relative contributions of interface- and bulk-limited conduction. Using representative, physically motivated parameter sets informed by prior device-level studies, the simulations accurately reproduce the characteristic $I$-$V$ signatures of seven different experimentally observed switching responses. Overall, the findings help to link microscopic defect landscapes and transport processes to experimentally measured macroscopic responses within a single, self-consistent modeling framework.
Read the full abstract
Machine Learning
RiVaT-Fuse: Reliability-Calibrated Variational Tensor Fusion with Matrix-Valued Trust for Multimodal Prediction under Modality Uncertainty
Multimodal prediction from images and structured metadata requires integrating complementary evidence whose reliability can vary across samples and latent directions. A single confidence weight per modality cannot capture this directional variation. We propose RiVaT-Fuse, a reliability-calibrated variational tensor fusion framework that formulates fusion as sample-wise latent-state estimation. For each image-metadata pair, the fused representation minimizes a quadratic objective combining agreement with modality embeddings, structured cross-modal interactions, and regularization. Positive-definite, low-rank-plus-diagonal trust matrices are conditioned on learned state descriptors and metadata completeness, allowing modality contributions to vary across latent directions. Additive, multiplicative, and relational interactions model cross-modal dependencies within the latent estimation objective. The resulting system admits a unique solution computed through a differentiable linear solve. A first-order analysis with fixed trust operators relates latent sensitivity to system conditioning and perturbations in modality embeddings and interactions. The framework further incorporates a state-binned entropic surrogate for conditional distributionally robust learning and task-coupled quadratic prediction heads. We instantiate RiVaT-Fuse on mBRSET, pairing retinal images with clinical and demographic metadata for diabetic retinopathy grading, diabetic macular edema detection, and referable-status prediction. Comparisons with unimodal and representation-level fusion baselines assess the predictive utility of the complete framework across these related clinical tasks.
Read the full abstract
Functional Analysis
A General Proof of the Fong-Tsui Conjecture
We present a general proof of the Fong-Tsui conjecture for bounded operators on arbitrary complex Hilbert spaces. Specifically, we show that $|T|\leq|\operatorname{Re}T|$ implies that $T$ is self-adjoint. The argument combines a positive inverse of a Sylvester map with a spectral cutoff determined by the norm of the positive defect $|\operatorname{Re}T|-|T|$. A local vanishing lemma reduces the analysis to the classical squared self-adjointness criterion, while positivity of the defect yields a global norm contradiction. We formulate the argument as an abstract four-operator vanishing principle, without compactness, trace, or separability assumptions. We also establish a quantitative stability estimate: if $|T|\leq|\operatorname{Re}T|+\varepsilon I$ and $0\leq\varepsilon\leq|T|$, then $|\operatorname{Im}T|\leq6|T|^{7/8}\varepsilon^{1/8}$. The constant is independent of the dimension, and the exponent is not claimed to be optimal. Large language models (LLMs) were used to assist with proof development, algebraic calculations, numerical checks, and auditing of the arguments.
Read the full abstract
Machine Learning
DR-LabStack: Design and Implementation of a Clinician-Facing Web System for Diabetic Retinopathy Prediction
Pretrained diabetic retinopathy (DR) prediction models differ in their input fields, serialization formats, preprocessing requirements, and output semantics. Making these models accessible through a common clinical interface therefore requires explicit coordination between the user interface and the inference service. We designed and implemented DR-LabStack, a React-Flask web system integrating four externally developed pretrained models: RuleFit, Pruned RuleFit, Elaborative XGBoost, and Two-level Ensemble. A shared form retrieves ordered model features, renders model-specific numerical and categorical controls, and constructs a positional input vector. Backend adapters load heterogeneous artifacts and apply the ensemble's accompanying scaler, while a common JSON response supports binary classification display alongside method and source information. Functional evaluation on September 8, 2026 used copied application files and real model artifacts in a documented isolated environment. All four models loaded and exposed their 14-, 6-, 8-, and 25-field contracts. Sixty-two Flask test-client requests characterized service behavior; 12 limited-vector checks confirmed invocation-path and threshold consistency. Twenty-four browser-component scenarios with mocked transport verified input ordering and result rendering and characterized input-validation behavior. The resulting system demonstrates a reusable interaction and serving workflow for heterogeneous DR models. The contribution is web-system design, integration, and software functionality; clinical effectiveness and clinician usability require separate evaluation.
Read the full abstract
Combinatorics
Settling the total domination-annihilation conjecture for graphs with minimum degree two
The total domination number $\gamma_t(G)$ of a graph $G$ is the minimum cardinality of a set $D\subseteq V(G)$ such that every vertex of $G$ has a neighbor in $D$. The annihilation number $a(G)$ is the largest integer $k$ for which the sum of the $k$ smallest degrees of $G$ is at most $|E(G)|$. A well-known conjecture, originating from Graffiti.pc and later formulated explicitly by Desormeaux, Haynes, and Henning, asserts that $\gamma_t(G)\le a(G)+1$ for every connected nontrivial graph $G$. The conjecture is known for graphs of minimum degree at least three and for several classes of graphs having vertices of degree one or two. In this paper we settle the minimum-degree-two case. More precisely, we prove $\gamma_t(G)\le a(G)+1$ for every connected graph $G$ with $\delta(G)=2$. The proof combines two sharp bounds on the total domination number with an estimate for the annihilation number. Moreover, in some specific cases, the stronger inequality $\gamma_t(G)\le a(G)$ holds.
Read the full abstract
Cosmology and Nongalactic Astrophysics
$w_0w_a$ or $bc$: DESI constraints on relative baryon-CDM perturbations
We carry out a search for the relative baryon-cold dark matter (CDM) density and velocity modes in the DESI full-shape (FS) data, and investigate whether DESI's preference for evolving dark energy can be affected by these relative perturbations, which are a known contaminant of the baryon acoustic oscillations. For the first time, we use a holistic effective field theory (EFT) treatment of the power spectrum and bispectrum both for the standard terms (without relative perturbations), modeled in EFT to one-loop order, and for the relative perturbations. For the latter we implement an efficient computation of the one-loop power spectrum that allows us to scan over different cosmologies. We obtain robust constraints on the relative density and velocity bias parameters $b_{\delta_{bc}}$, and $b_{v^2_{bc}}$ for all the DESI DR1 tracers, finding e.g. $b_{\delta_{bc}}=5 \pm 6$, $b_{v^2_{bc}}=-0.031\pm 0.023$ (at 68% CL) for the LRG2 sample in a DESI FS-only analysis within $\Lambda$CDM, supplemented by CMB priors on the baryon density and spectral tilt. Assuming instead a $w_0w_a$CDM cosmological model, and combining DESI FS with the CMB primary anisotropies, CMB lensing, and Pantheon+ supernovae, our LRG2 constraints read $b_{\delta_{bc}}=2 \pm 5$, $b_{v^2_{bc}}=-0.032\pm 0.022$. Adding the relative velocity divergence biases we find $b_{\delta_{bc}}=-17_{-17}^{+19}$, $b_{\theta_{bc}}=-760_{-680}^{+700}$, $b_{v^2_{bc}}=-0.032\pm 0.024$ for the same sample. The constraints on the dark energy equation of state parameters are virtually unchanged when marginalizing over the baryon-CDM modes. We conclude that DESI's preference for evolving dark energy is robust with respect to the baryon-CDM relative perturbations.
Read the full abstract
Mathematical Physics
Henstock–Kurzweil Gauge Integral in the Non–Gaussian Regime: A Machine–Verified Construction
We develop a machine-checked construction of non-Gaussian functional integrals using the Henstock–Kurzweil gauge integral and Chernoff product approximations. The central object is a finite family of bosonic modes with action S(phi) = (1/2) phi^T A phi + lambda * sum_i phi_i^4, where A is positive definite. We prove that the one-mode integral I(omega, j, lambda) is finite, strictly positive, monotone and infinitely differentiable in the coupling lambda on [0, infinity). Its derivatives are given by convergent integrals of phi^{4k} with the same weight, not by the divergent perturbative series. The M-mode influence functional factorises into one-mode integrals and is bounded by its Gaussian value. A Chernoff / Lie–Trotter splitting handles the non-commutativity of the free and non-Gaussian generators. All statements are formalised in Lean 4 with Mathlib; the accompanying file HkNonGaussian.lean is free of sorry and uses only the standard axioms propext, Classical.choice, Quot.sound. Four illustrative applications are worked out at the level of explicit formulas: the Duffing oscillator, local volatility (CEV) in finance, Wilson–Cowan neural fields, and non-Gaussian quantum reservoirs. The construction is completely direct and does not use Wick rotation, Wiener measure, zeta-regularisation or analytic continuation back from imaginary time.
Read the full abstract
Computation and Language
Analyzing Traditional and Neural Approaches to Multilingual Readability Assessment
Transformer-based models excel at Automatic Readability Assessment (ARA), yet feature-based models remain in active use because their predictions tie back to linguistic properties. This matters because readability labels are subjective and rater-dependent, so high accuracy on noisy ground truth may reflect surface patterns rather than the linguistic structure that defines difficulty. We test whether transformers internalize the same features as traditional models across Arabic, English, French, Hindi, and Russian using the ReadMe++ dataset. Shapley Additive Explanations (SHAP) identify the features driving traditional classifiers, which we then use as TCAV concept sets to probe multilingual XLM-R and language-specific encoders. Transformers recover surface-length, syntactic, and lexical-diversity signals, and reflect the ordinal CEFR structure of the traditional models. Alignment varies by model family, language, and layer, with language-specific encoders tracking traditional models more clearly than XLM-R. High linear separability does not always imply directional influence, limiting linear probing for count-based readability features.
Read the full abstract
History and Overview of Mathematics
Perspectives on the unit distance problem
This is a survey on an old open problem in combinatorics called the unit distance problem, and the field of mathematics around it, called incidence geometry. What do we know about the problem? Why is it difficult? How does it connect with other parts of math?
Distributed Computing
Composable CXL Memory as a Kubernetes-Native Shared Memory for LLM Serving
We present a Kubernetes Dynamic Resource Allocation (DRA) driver that makes composable CXL memory a schedulable cluster resource, and evaluate the resulting shared-memory tier for cross-node KV-cache reuse in LLM serving. The driver composes CXL regions on demand, materializes them as DAX devices on each participating host, and injects them into pods under a single Container Device Interface (CDI) name so that pods on different nodes access the same physical region. A shared-memory connector for vLLM/llm-d uses that region as a KV-cache tier with a slot directory embedded inside the shared medium, which eliminates the need for an external metadata service. On a two-node cluster with a 512\,GiB CXL appliance and Qwen2.5-7B-Instruct, cross-node prefix reuse reduces TTFT by 5.5$\times$–36.6$\times$ at an external hit rate of 95.4–99.5\,%, while node-local tiers (GPU prefix caching, CPU-DRAM offload) fall back to full recompute. The sharing gap, defined as the latency ratio between cross-node and same-node reuse, is 1–4%, indicating that cross-node reuse incurs little additional latency relative to same-node reuse on our testbed. Both replicas run full engines; the study demonstrates memory disaggregation rather than prefill/decode disaggregation. We report this as a feasibility study rather than a performance evaluation.
Read the full abstract
Computer Vision
Two-Parameter Flow Map Learning for Continuous-Time Diffeomorphic Image Registration
Diffeomorphic image registration is central to medical image analysis, enabling anatomically consistent alignment across subjects. Most learning-based diffeomorphic methods model autonomous ODEs(ordinary differential equations) by parameterizing a stationary velocity field and recovering deformations via scaling-and-squaring. While non-autonomous ODEs with time-dependent velocities increase expressiveness, existing approaches rely on numerical integration to implicitly enforce flow structure that entangles model expressiveness with discretization accuracy. We propose a framework to directly learn the continuous-time solution of a non-autonomous ODE formulated as a two-parameterflow map. By enforcing cocycle consistency, a fundamental structural property of time-varying flows, we learn the flow maps without time discretization and velocity integration during training. The framework recovers diffeomorphic mappings at inference using a small number of compositions. Our proposed framework seamlessly incorporates standard registration backbones and improves alignment accuracy consistently across nine datasets while preserving diffeomorphic structure. Notably, the proposed method achieves an average Dice improvement of 2.1% on brain MRI benchmarks, a 12% TRE reduction on lung CT, and a 2.6% Dice gain on cardiac MRI and ultrasound datasets.
Read the full abstract
Metric Geometry
The Nieuwland number of the Octahedron
We prove that the Nieuwland number of the Octahedron is $3\sqrt{2}/4$. Along the way, we provide an elementary proof that Rupert's problem for the Octahedron admits no solution in which the outer projection is a hexagon.
Computer Vision
Shedding Light: A Benchmark for Evaluating Lighting Understanding in Generative Image Models
Accurate modelling of illumination is central to realistic image synthesis and scene understanding. Yet, there is little exploration into whether image generative models are good at this task or whether physical plausibility remains a key challenge for them. Clearly, significant progress has been made in realistic image synthesis, but do models truly understand lighting in a physically accurate manner? To answer this question, this work proposes a benchmark to assess the lighting understanding and harmonisation capabilities of generative models. Our key insight is that evaluating lighting understanding for such models only requires testing how well they insert novel objects into real photographs whilst maintaining consistent illumination. To do so, we use a multi-illumination dataset with images containing simple objects serving as “light probes”, and prompt models to inpaint the same object onto the original image, then compare the generated results against the ground-truth light probes. We then estimate the lighting direction, colour and radiance distribution from the inpainted probes, providing a quantitative measure of illumination accuracy and photometric realism. Our work establishes a scalable evaluation protocol to systematically assess how well generative models capture and reproduce real-world lighting, offering a foundation for benchmarking the photometric accuracy of any future models. All code and data are available at https://lvsn.github.io/SheddingLight/ .
Read the full abstract
Solar and Stellar Astrophysics
UX Men: Detached Eclipsing Binary as a Benchmark Candidate
We present ultra-precise fundamental parameters for the detached eclipsing binary UX Men based on 31 TESS sectors ($\sim$882 d) and radial velocities from HARPS, FEROS, CORALIE, and UCLES. The light curves were detrended for spot-induced variability and modelled independently with JKTEBOP and ALLESFITTER, while radial velocities were extracted via TODCOR and fitted with a Keplerian orbit. Spectral disentangling yielded individual component spectra analysed with iSpec to determine $T_{\rm eff}$ and $\mathrm{[M/H]}$. Multi-band photometry was used to determine the fluxes, flux ratios and ultimately to measure photometric $T_{\rm eff}$. We obtain the masses $M_{\rm A}$ = $1.2300 \pm{0.0012}$ M$_{\odot}$, $M_{\rm B}$ = $1.1946\pm{0.0012}$ M$_{\odot}$ and radii $R_{\rm A}$ = $1.3605\pm{0.0031}$ R$_{\odot}$, $R_{\rm B}$ = $1.2801\pm{0.0069}$ R$_{\odot}$ with sub-percent precision. Uncertainties were found to be dominated by systematic effects coming from (weak) stellar activity. $T_{\rm eff,A}$ = $6302\pm{77}$, $T_{\rm eff,B}$ = $6292\pm65$ & Metallicity estimates are $[M/H]_{\rm A} = 0.02\pm{0.13}$ dex and $[M/H]_{\rm B}= 0.07\pm{0.10}$ dex. Isochrone fitting yields a consistent age of $\tau$ = $2.75\pm{0.13}$ Gyr. The system exhibits a $\sim$ 673-day periodicity in eclipse-timing variations, likely due to stellar activity, with no evidence for a third body. Results are independent of the modelling code used, but without proper treatment of activity-induced systematics the uncertainties would be underestimated. UX Men is a high-precision benchmark binary providing stringent tests of stellar models and a robust calibration source for PLATO. Expanding the sample of well-characterised detached eclipsing binaries will strengthen the empirical foundations for stellar physics and the calibration of future photometric missions
Read the full abstract
Systems and Control
Structural Sign Herdability in Temporal Networks: A Sufficient Condition via $\pi_p$-Graphs
In this letter, we study the herdability of temporally switching directed networks. A temporal network is modeled as a switched system with a fixed switching sequence, which imposes more restrictive herdability conditions than those of conventional switched systems. By exploiting the relationship between temporal walks and the entries of the controllability matrix, we derive sufficient conditions for herdability. We further show that the magnitude of edge weights influences the sign pattern of the controllability matrix, thereby affecting herdability. Consequently, herdability in temporal networks depends not only on the network topology and switching durations, but also on the magnitude of the edge weights. Motivated by this observation, we establish equivalent graph-theoretic conditions for structural sign ($\mathcal{SS}$) herdability in temporal networks. In particular, we introduce the union multigraph of temporal subsystems and propose the notion of a $\pi$-graph. We show that the existence of a $\pi_p$-graph, which is a temporally evolving $\pi$-graph, is sufficient to guarantee $\mathcal{SS}$ herdability. Illustrative examples are provided to demonstrate the proposed results.
Read the full abstract
High Energy Astrophysical Phenomena
Low-Mass Magnetic Monopoles in the Galaxy: Simulations and Comparison with Ultra-High-Energy Cosmic-Ray Data
Origin and composition of ultra-high energy cosmic rays are still uncertain, particularly the rare events exceeding the Greisen-Zatsepin-Kuzmin cutoff and whose apparent arrival directions point to the Local Void. This work is an in-depth investigation of the possibility that such cosmic rays contain low-mass magnetic monopoles predicted in several recent theoretical models. Using a custom extension to CRPropa, a state-of-the-art code for modeling cosmic ray propagation, monopoles are tracked in the galactic environment under realistic assumptions on mass, magnetic charge, and initial phase-space distributions. Our simulations demonstrate that relic MMs follow filamentary “Galactic magnetic funnels”, producing anisotropic arrival directions at Earth. The expected arrival directions are concentrated in a small region of the sky, which has significant implications for the design and interpretation of future searches. A comparison of the simulated arrival directions with the Pierre Auger Observatory dataset shows that the null hypothesis - no monopole contribution - is statistically preferred for the majority of cases. Using a profile-likelihood analysis that incorporates both directional and energy information, we set 90% C.L. upper limits on the integral MM flux of $< 2.1\times10^{-23}$ $\mathrm{cm^{-2}s^{-1}sr^{-1}}$.
Read the full abstract
Combinatorics
A 50-Vertex Cubic Counterexample to the Domination-versus-Edge-Domination Conjecture
Baste, Furst, Henning, Mohr, and Rautenbach conjectured that every finite regular graph of positive degree satisfies \(\gamma(G) \leq \gamma_e(G)\), where \(\gamma\) is the domination number and \(\gamma_e\) is the edge domination number, equivalently the minimum cardinality of a maximal matching. We show that the conjecture is false already for cubic graphs. The counterexample is a previously public 50-vertex cubic graph that had been used to refute the stronger independent-domination inequality \(i(G) \leq \gamma_e(G)\). For this graph we prove \(\gamma(G) = 16 > 15 = \gamma_e(G)\). The equality \(\gamma_e(G) = 15\) has a short counting proof, and a dominating set of order 16 is displayed explicitly. For the lower bound \(\gamma(G) \geq 16\), we give a self-contained exact reduction: after fixing which of the 20 clause vertices lie in a putative dominating set, the remaining problem is a finite set-cover problem on the 30 literal vertices. We enumerate all \(2^{20} = 1,048,576\) clause subsets, derive two explicit lower bounds, and solve exactly the 5,931 residual cases by a recurrence stated in the paper. The complete case counts and minima are displayed, and a short standard-library Python implementation is included in an appendix. A separate 893,049-node proof-tree certificate and a direct graph search provide independent verification. Thus the regular-graph conjecture is disproved. Combined with Gupta's recent theorem that every cubic graph on at most 48 vertices satisfies the conjectured inequality, the example is order-minimal among cubic counterexamples.
Read the full abstract
Astrophysics of Galaxies
A Novel Approach to 3D Dust Mapping of the Central Molecular Zone
The 3D distribution of dust and gas in the Milky Way's Central Molecular Zone (CMZ) is key to understanding gas inflows towards the Galactic Centre (GC), the process of star formation in this extreme environment, and the propagation of energetic cosmic rays originating from Sgr A*. However, while recent efforts have combined datasets in a Bayesian framework to estimate the near/far positions of individual molecular clouds in the CMZ, conflicts between different methodologies still remain and we are still lacking a comprehensive, model-independent map of all of the gas and dust in the CMZ, which is critical to address key science questions. Here we develop a new methodology to infer the 3D dust distribution of the CMZ. The key idea of the method is to use stellar proper motions to get probabilistic information about the unknown stellar distances through a model of the distribution of star positions and velocities of the nuclear stellar disc (NSD), co-spatial with the CMZ. Taking stellar proper motions and extinctions as input, the latter adopted as a proxy of the dust column density, the method returns the 3D dust distribution. It is non-parametric, makes no assumptions on the functional form of the dust distribution, and is fundamentally distinct and largely independent of all existing methods. We show that the method can robustly and effectively reconstruct the mock 3D CMZ structure by testing it on a range of mock dust distributions, both analytically generated and taken from hydrodynamical simulations. Finally, we discuss the prospects for applying the method to real data.
Read the full abstract
Machine Learning
From Connectivity to Rewards: Dense Reward Learning with Directed State Graphs
The integration of graphs with Goal-Conditioned Hierarchical Reinforcement Learning (GCHRL) has received increasing attention, as graphs naturally encode task hierarchies for effective subgoal sampling. However, existing methods often overlook intrinsic connectivity information, failing to fully leverage the underlying topology for efficient learning. Most graph-based GCHRL methods use the graph as a stochastic sampling tool rather than as an environmental model that encodes connectivity and state-accessibility information. This limitation is particularly acute in quasimetric environments, where the inherent asymmetry of state transitions poses a fundamental challenge to stable policy learning and robust path planning. In this paper, we address these problems by introducing a state connectivity model designed to predict pairwise state connectivity strength in asymmetric environments. We transform these connectivity strengths into scalar auxiliary dense rewards, providing continuous guidance across multiple hierarchical levels. We demonstrate that our proposed framework, Graph-Guided Quasimetric Dense Reward (G2QDR), can theoretically be integrated into any existing GCHRL architecture, and the state connectivity model is efficiently implemented via a neural network trained on a directed state graph generated during exploration. Empirical results across a wide range of sparse reward environments indicate that, in general, G2QDR can enhance the performance of baseline GCHRL approaches with acceptable computational overhead.
Read the full abstract
Cryptography and Security
Big Enough to Break Out: Tracking the Rising Capability of LLM Penetration-Testing Agents
Large language model (LLM) agents are increasingly applied to penetration testing, but we still know little about what they can do or how they fail. We compare two PentestGPT-based systems: a legacy human-in-the-loop system running the open-weight Kimi K2.5, and a newer autonomous system running Claude Opus 4.8. Across three public targets, the autonomous system solves all three, including the two the legacy system never finishes. The legacy result is the more surprising of the two. Even on the machines the legacy system fails to solve, it completes about half the subtasks, while running on ordinary university GPUs with no provider guardrails. We can describe the trend but not explain it, since model, harness, autonomy, and memory architecture all change together. Its direction still points to the next question: what will limit these agents as they take on more complex tasks? The usual answer is long-horizon memory, the loss of access to earlier findings during long attack chains. We test it by adding a coverage-memory layer to both systems, and neither improves outcomes. In the legacy stalled runs we could review, the limiting factor appeared to be planning and commitment rather than lost memory: agents held the evidence for a route forward and never turned it into a concrete exploitation hypothesis, which may suggest that offensive capability will advance with agents' ability to plan rather than with better memory. The same subtask scoring that tracks this capability is available to defenders, who can measure it as it rises instead of waiting to meet it in the field.
Read the full abstract
Computers and Society
Designing Technology for Social Wellbeing in Built Environments: A Conceptual Framework
Digital technologies are deployed in urban built environments with the aim of supporting social dimensions. However, the research offers no unified guidance for such digital technology design. Additionally, evidence shows that week social wellbeing contributes to mental and physical health outcomes which suggests that the technology designed to strengthen social wellbeing could also function as a form of health promoting and preventive intervention. This chapter addresses this research gap by developing a conceptual framework that supports the design of digital technologies for social wellbeing in built environments. We propose a conceptual framework composed of three core components which are drawn on the synthesis of selected empirical studies on technologies embedded in built environments for social wellbeing. First, a social wellbeing dimensions model that identifies what digital technology could address. Second, a digital technology contribution matrix that distinguishes the types of contributions a digital technology could make. Third, levels that maps the scope at which technology could support social wellbeing. This conceptual framework could help researchers, practitioners and policymakers to design and guide digital technology interventions that target social wellbeing in the built environment.
Read the full abstract
Machine Learning
Counterfactual Marginalisation: Framework for Evaluating Robustness to Nuisance Variables
Machine learning models can achieve strong test performance while relying on demographic or acquisition-related shortcuts. We propose counterfactual (CF) marginalisation as a test-time evaluation procedure for assessing robustness of classification models to such variables. Given a CF image generator, we intervene on nuisance parent variables such as age or sex, generate CF versions of each test image, and average predictions over a target intervention distribution. This produces intervention-aware predictions that marginalise demographic effects while preserving patient-specific latent information. We use these predictions to define metrics for CF risk, calibration, stability and worst-case sensitivity. We demonstrate this framework's utility for quantitative robustness evaluation.
Read the full abstract
Mesoscale and Nanoscale Physics
Thermal Stability of Encapsulated Molecular Structures with Extended OH-Hydrogen-Bond Chains
Using a coarse-grained model, we performed numerical simulations of the dynamics of linear molecular chains adsorbed on a flat substrate (on the surface of an h-BN crystal). It is shown that molecules containing benzene rings and hydroxyl groups in their structure can form stable hydrogen-bond chains OH$\cdots$OH$\cdots$OH. Such chains can be formed by phenol C$_6$H$_5$OH, 4-phenylphenol C$_6$H$_5$–C$_6$H$_4$OH, paracetamol CH$_3$C(O)NHC$_6$H$_4$OH, and 4-hydroxybenzanilide C$_6$H$_5$C(O)NHC$_6$H$_4$OH molecules. The dissociation of these chains occurs at temperatures above $T_1$=160, 210, 300, and 390K, respectively. Coating such molecular systems with a hexagonal boron nitride sheet (van der Waals encapsulation) significantly enhances their thermal stability. Such encapsulated molecular structures retain hydrogen-bond chains up to temperatures of $T_2$=510, 860, 600, and 1260K, respectively. The performed simulations allow us to conclude that chains of these molecules encapsulated by h-BN sheets can be used to create anhydrous proton-exchange membranes capable of operating at high temperatures. The most promising are chains of 4-hydroxybenzanilide molecules, since their hydrogen-bond chains possess the highest thermal stability.