Computer Vision
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Tuesday 8 September
Learned image matching has experienced significant progress in recent years, culminating in robust and accurate matchers such as RoMa, whose robustness is often attributed to its use of frozen DINO features. In a parallel development, feed-forward reconstruction models, such as VGGT, have been trained on ever-growing datasets to accurately regress dense 3D point maps and camera poses. The distinction between matchers and feed-forward reconstruction models has become increasingly blurred with the introduction of matching losses in models such as MASt3R and VGGT-$\Omega$. This raises a natural question: what do feed-forward 3D models know about image matching? In this work, we answer this question by analyzing three scenarios: (i) zero-shot matching of patch features, (ii) direct matching of 3D point predictions, and (iii) training a full matcher on top of the learned representations. We find that, despite performing poorly in zero-shot matching, especially in later layers, feed-forward reconstruction models provide strong representations for linear probing and full matching pipelines. We further show that, even without any training, their raw predictions alone enable competitive matching, albeit only under moderate viewpoint changes and modality gaps. Based on these insights, we retrain RoMa v2 by replacing its DINO backbone with VGGT-$\Omega$. Our resulting model, \ours, outperforms state-of-the-art matchers on a wide range of benchmarks, e.g. +8.1 mAA compared to RoMa v2 on WxBS.
Numerical Analysis
Constructive Tchakaloff results and well-conditioned quadrature through randomized least squares
We consider using randomized least squares to construct quadrature rules exact on a subspace of functions. Using new conditions that we call relative admissibility and reference weight concentration, we establish both that the quadrature weights from such a procedure concentrate close to their asymptotic values with prescribed and arbitrarily large probability, and this in turn provides useful finite-sample probabilistic bounds on the stability of the resulting quadrature rules for very general classes of possibly complex-valued functions. Our analysis both significantly generalizes the existing analysis of randomized least squares quadrature construction, and provides new bounds on stability for these rules. These results specialize to existence results for positive quadrature rules, when such rules can be theoretically expected. Because our procedures are formally algorithmic, our analysis is a substantive advance toward computationally constructive generalized Tchakaloff theorems.
Read the full abstract
Mesoscale and Nanoscale Physics
Intrinsic (valley) thermal Hall conductivity of phonons in graphene for integer quantum Hall phases
We calculate the intrinsic contribution to the (valley) thermal Hall conductivity of phonons in (Semenoff-gapped) graphene in the quantum Hall regime where bulk electron thermal transport is suppressed. The (valley) thermal Hall transport of phonons considered here originates from a phonon (valley) Hall viscosity induced by electron (valley) Hall conductivity and viscosity via geometric electron-phonon coupling. The so generated phonon (valley) Hall viscosity is identified with the emergent (valley) Hall viscosity recently introduced in Phys. Lett. A 595, 132096 (2026). While our calculations refer to graphene, they may very well be generalized to other Dirac materials like group-VI transition metal dichalcogenides. We discuss the prospect of measuring the phonon thermal Hall conductivity in graphene-based heterostructures and estimate the corresponding valley response induced by electrons in the inversion symmetry broken phase.
Read the full abstract
Cosmology and Nongalactic Astrophysics
Inductive Biases in Field-Level Cosmological Inference from Galaxy Catalogs
We perform field-level likelihood-free inference of the matter density parameter $\Omega_m$ from simulated galaxy catalogs using machine learning models with differing inductive biases. Using hydrodynamic simulations from CAMELS, we examine how observable choice and architecture govern cosmological information extraction. We consider galaxy positions and line-of-sight peculiar velocities, separately and jointly, and compare permutation-invariant Deep Sets, implemented with either multilayer perceptrons (MLPs) or Kolmogorov-Arnold Networks (KANs), to graph neural networks (GNNs), which explicitly encode spatial relations. We test in-distribution and out-of-distribution (OOD) performance across simulations with different subgrid galaxy-formation prescriptions. Deep Sets infer $\Omega_m$ from velocities alone with mean relative errors of approximately $18\%$ in-distribution and $\sim25\%$ OOD, with KANs and MLPs achieving comparable performance. In contrast, the same set-based approach does not yield useful $\sigma_8$ predictions in either in-distribution or cross-suite tests. Adding positions does not improve Deep Sets, while GNNs infer $\Omega_m$ with mean relative errors of about $10\%$ in-distribution and $10$–$17\%$ OOD. These results indicate that peculiar velocities provide the dominant source of $\Omega_m$ information for set-based models in this setting, while spatial information is most effectively used by architectures that explicitly encode galaxy-galaxy relations. Because the velocity inputs are exact simulated peculiar velocities, applications to survey data will require validation under realistic velocity-measurement noise, selection effects, and survey geometry.
Read the full abstract
Robotics
Agentic AI-enabled Semantic Commissioning of a Cognitive Digital Twin for Reconfigurable Manufacturing
Rapid bespoke commissioning of the Cognitive Digital Twin (CDT) is a major challenge in reconfigurable manufacturing. Traditional digital twin (DT) construction methods primarily focus on geometric reconstruction, often neglecting the deep semantic integration and functional interoperability necessary for autonomous reasoning. This paper proposes an agent-based, AI-driven workflow to automate end-to-end CDT debugging. The system utilises LangGraph as a multi-agent orchestration engine to achieve dual-path synthesis: the semantic path extracts technical specifications from unstructured documents using Retrieval Augmented Generation (RAG), while the functional path autonomously discovers and binds to real-time industrial telemetry data using Model Context Protocol (MCP). Experimental validation in a robotic machining cell demonstrates that the system achieves a mean average accuracy (mAP) of 97.2% in perception and reduces the deployment cycle from several weeks to an average of 2 hours, marking a paradigm shift from manual scripting to autonomous orchestration.
Read the full abstract
Machine Learning (Statistics)
Recovery Theory for Projected Power Iterations in Permutation Synchronization
We study the projected power method for permutation synchronization: recovering $n$ unknown permutations of $m$ objects from pairwise relative permutations, observed independently with probability $p$ and independently corrupted by uniform permutations with probability $1-\pi_0$. Our main theorem establishes uniform one-step contraction for all estimates with optimally aligned block-error fraction at most $1/2-\epsilon$, for fixed $\epsilon>0$, including measurement-dependent estimates. When $np\ge C_0\log(n)$ and $m=o(np\pi_0^2)$, one update contracts the error by a factor $O(m/(np\pi_0^2))$, up to a floor of $O(e^{-c\min\{m\,np\pi_0^2,\,np\pi_0\}} + \log(n) / n)$. Thus, every initialization in this majority basin achieves $o(1)$ error after one update and maintains it uniformly over subsequent iterations. A reference-block spectral initializer enters the basin with probability $1-o(1)$, yielding end-to-end almost-exact recovery. If additionally $\min\{m\,np\pi_0^2,\,np\pi_0\} \ge C\log(n)$ for sufficiently large $C$, the method achieves exact recovery in finitely many iterations. We also establish prescribed-block recovery under weaker conditions for measurement-independent estimates and extend the guarantees to independent permutation-valued corruptions with uniform mean and to partial permutations with common visible support.
Read the full abstract
Populations and Evolution
Population Ecology of Tunes
How cultural repertoires maintain diversity under selection is a fundamental question in cultural evolution. We address this using thirteen years of weekly popularity data for approximately 20,000 Irish traditional tunes, fitting ecological birth-process models under neutral, frequency-dependent, and per-tune selection hypotheses. We find strong evidence that tunes differ in intrinsic fitness - some are systematically more likely to be learned than others. We find that 29% of the variance in fitness can be explained by a mixture of social and melodic features. Some tunes appear to be carried along via linkage due to the tradition of playing tunes in sets, analogous to selective sweeps in genetics. By measuring changes in fitness over time and comparing this with recordings we precisely identify the mechanism by which a long-dormant tune can become fit through a popular recording. Despite the directional selection, repertoire diversity increases, driven by the continual arrival of new compositions. These results demonstrate that selection and diversity can coexist in a cultural ecosystem, and establish Irish traditional music as a quantitatively tractable system for studying the evolution of cultural variants and understanding what makes a tune stand out.
Read the full abstract
Statistical Mechanics
Topological instability and reentrant crystallization in active solids
We extend the KTHNY theory of defect-mediated melting in two dimensions to the case of active solids exhibiting both non-reciprocal (odd) elasticity and Cosserat (micropolar) coupling. For perturbatively small values of non-reciprocity, melting still proceeds via defect unbinding, but the melting temperature shifts due to activity. However, at a threshold value of non-reciprocity, we discover a qualitatively distinct mechanism for 2D melting, which we call topological instability. This melting occurs through the proliferation of defect pairs at any temperature. We use a combination of field theory and simulations to characterize this zero-temperature transition in terms of the defect-pair fugacity. Surprisingly, for higher activity values, we find reentrant crystallization, where the quasi-long-range order survives even at temperatures for which the equilibrium crystal would melt. Although active solids present one realization of this topological instability, we envision core-energy-driven defect proliferation as a generic and unexplored route for two-dimensional melting.
Read the full abstract
Audio and Speech Processing
Language Orthogonalization of Self-Supervised Speech Representations for Cross-lingual Parkinson's Detection
Self-supervised speech models (S3Ms) provide powerful representations for Parkinson's disease (PD) detection, making cross-lingual transfer attractive for languages lacking labeled patient speech. However, these representations also encode language identity, which can confound this transfer: without target-language PD speech, classifiers may separate languages rather than pathology, yielding high specificity but low sensitivity on target patients. We propose language orthogonalization, a closed-form ridge residualization of S3M features against external VoxLingua107 language embeddings, fitted using only healthy-control (HC) speech. By removing language-predictable components while retaining pathology-related variation, it produces a less language-dependent geometry in which HC representations concentrate while PD representations disperse. Across five S3M backbones, three speech tasks, and three target languages, our method consistently improves cross-lingual PD-detection performance while correcting the high-specificity/low-sensitivity failure.
Read the full abstract
Statistical Mechanics
Gap-controlled thermalization in a chain of alternating nonlinear springs: an SSH model version of the Fermi-Pasta-Ulam-Tsingou problem
In classical anharmonic lattices, the resonance structure driving nonlinear mode mixing is reshaped by band gaps. However, it remains unclear whether a spectral gap hinders or promotes long-time thermalization. Here we study a Fermi-Pasta-Ulam-Tsingou (FPUT) chain of equal masses joined by alternating dimensionless spring constants $1\pm\Delta\kappa$, so that the chain is dimerized with strength $\Delta\kappa$, the classical vibrational analogue of the Su-Schrieffer-Heeger (SSH) model, and quantify how the acoustic-optical gap controls relaxation under $\alpha$-type (cubic) nonlinearity. Tracking modal energies and spectral entropy via long-time symplectic simulations, we find that dimerization delays long-time relaxation. At $\Delta\kappa=1/2$, set by the onset of the first umklapp process that allows two zone-boundary acoustic phonons to fuse into a zone-center optical phonon, the three-wave acoustic-acoustic-optical channel closes. The effective acoustic-to-optical scattering rate reveals an interplay between the dynamic resonance vertex and the kinematic resonance manifold, which evolve differently with the dimerization parameter. Above the threshold $\Delta\kappa=1/2$, energy redistribution can nevertheless continue through higher-order resonances and quasiresonances. Yet specific boundary conditions and initial conditions can inhibit equipartition. These results identify the phononic gap as a tunable, momentum-selective filter for nonlinear energy transport, suggesting a general route to controlling energy transfer and equipartition in dimerized or topologically gapped nonlinear lattices and in flexible mechanical systems.
Read the full abstract
Materials Science
Light-Driven Ultrafast Control of Time-Reversal Symmetry in SrTiO3
Light-driven control of material symmetry enables the engineering of phenomena forbidden in equilibrium. Although electromagnetic fields have been used to break time-reversal symmetry, its dynamic control on ultrafast timescales in nonmagnetic insulating oxides remains virtually unexplored. Here, we demonstrate simultaneous magnetic symmetry lowering and timereversal-symmetry breaking in cubic SrTiO3 driven by an off-resonant, elliptically polarized THz pulse. Second harmonic generation (SHG) polarimetry reveals emergent SHG circular dichroism inconsistent with third-order nonlinear processes under cubic m3m symmetry. The SHG polar patterns exhibit mirror-symmetry breaking, indicating a transient reduction to tetragonal 4/mm'm' magnetic symmetry. Furthermore, the SHG response scales linearly with the angular momentum of THz pulse, providing direct evidence for THz-field-induced timereversal-symmetry breaking. These results establish a direct pathway for ultrafast symmetry control in non-magnetic oxides, enabling dynamic manipulation of emergent phases on ultrafast timescales.
Read the full abstract
Computation and Language
The Mutations of Machine Speech
Algorithmic outputs now populate the digital environments through which contemporary life is organized. The role of law in facilitating and constituting (rather than merely responding to) these processes is gaining increasing traction across scholarly accounts. This inquiry traces the evolution of algorithmic outputs attending to their legal underpinnings and social implications, surfacing the mutations of machine speech. The first mutation redefined speech as data to be queried: search engines transformed the web from a space of information retrieval into an economic regime of algorithmic visibility. The second mutation reframed speech as engagement: social media platforms fused moderation with amplification, turning expression into a metric of attention, governed by corporate architectures. The third mutation emerges in conversational systems and interfaces, where generative text displaces information retrieval, bringing with it dense technolegal entanglements and profound epistemic consequences. Scholars of freedom of expression, informational privacy, and communication studies have long grappled with these dynamics, yet their implications for broader legal thought have also become urgent. This piece seeks to organize and clarify the evolving debate around algorithmic speech, making this critical but often fragmented discourse more accessible to wider legal and interdisciplinary audiences. In doing so, it bridges the gap between observing technological transformation and critically assessing the constitutive role of law within it, offering a conceptual resource for researchers, students, policymakers, and practitioners navigating and contesting this evolving landscape.
Read the full abstract
Quantum Physics
Certifying bipartite entanglement on a superconducting processor from a corrected QAOA cost layer
The Quantum Approximate Optimisation Algorithm is the canonical near-term heuristic, yet on real superconducting hardware its behaviour is set by device physics, not by the algorithm. Hardware demonstrations rarely prove the intended coupling is realised, that genuine entanglement is prepared, or what controls it. We report a pre-registered single-device study on the nine-qubit superconducting processor Red at the Barcelona Supercomputing Center, using its five-qubit component. Across six calibrations we certify bipartite entanglement of a corrected QAOA cost layer on the good coupler at depth one by two-qubit tomographic negativity: 0.077, BCa 95% CI [0.065, 0.091], excluding zero every day, while the coupling-off twin stays separable and the degraded coupler equals zero. We then measure the device physics directly: in a pre-registered manipulation on a fixed Bell preparation (not the QAOA layer), reducing the CZ pulse amplitude causally reduces the certified negativity along a positive, well-fit slope (R^2 = 0.83-0.96), every rung-level interval excluding zero and surviving Holm and Benjamini-Hochberg correction. This is a local linearisation of a peaked response, not a global monotone law: the negativity peaks at conditional phase |zeta| = pi, which the nominal amplitude overshoots. The J=0 twin is a separate off-state control and the fringe visibility is constant, so the effect is coherent. The average gate fidelity is non-monotone in the drive and is reported only as a diagnostic. A per-circuit statevector guardrail excludes a diagonal compilation fault we disclose and repair. A randomized-measurement witness is consistent with entanglement but underpowered. The certificate collapses by depth three, and a pre-registered sixteen-instance depth-three ensemble shows no optimisation signal (one-sided Wilcoxon p = 0.15). We claim no advantage; the instances are classically trivial.
Read the full abstract
Operator Algebras
Infinite twisted $C^*$-tensor product and symmetric states
The twisted $C^*$-tensor product was exhaustively investigated by the authors in two previous papers. In this new context, generalising the usual tensor product and the Fermi one, we analyse the possibility of studying the set of symmetric states, that is those invariant under all finite permutations, in the setting of infinite twisted $C^*$-tensor products. As a preliminary result, we recognise that such an investigation can proceed only when the bicharacter, involved in the construction of such twisted tensor products, is hermitian. Otherwise, there is no natural action of the finitary symmetric group on the infinite twisted chain. Furthermore, even if the investigation of symmetric states is certainly meaningful for all twisted systems based on hermitian bicharacters, it is shown that it can be fruitfully carried out in three cases only. In these cases, we can provide the "genuine" version of the celebrated De Finetti Theorem already established by Hewitt and Savage for the general classical case, Stormer for the usual tensor product and Fidaleo for Fermi models. Very surprisingly, it emerges that one more twisted model can be treated exhaustively: it corresponds to the chain twisted by the so-called Klein four-group, and its Klein bicharacter unique up to equivalence.
Read the full abstract
Systems and Control
Strategic Dispatching Equilibrium under Competition in Multi-Regional Ride-Hailing Markets
This paper studies competition between ride-hailing companies in a multi-regional market through a static non-cooperative game, in which each company allocates its fleet across regions to maximize profit. Under the assumptions of the proposed framework, we establish an equilibrium existence result on a restricted feasible set by analyzing the asymptotic quasi-concavity of the payoff functions. A numerical study of a two-company, two-region market shows that relative fleet size strongly affects equilibrium dispatching strategies: the smaller company tends to concentrate its fleet in one region and serves both regions only after its fleet share (in percentage) exceeds a threshold. In the numerical setting considered here and under the Logit demand specification, more balanced fleet sizes are associated with higher total industry profit and higher total served demand than strongly asymmetric fleet configurations.
Read the full abstract
Robotics
Actuator Dynamics Curricula for Narrow-Viability Tasks in Legged Robot Learning
Reinforcement learning has produced capable controllers across a broad range of legged-robot tasks, but a subset of these tasks fail to converge under standard training: those for which most exploration trajectories terminate before producing useful gradient signal. To address such tasks we introduce the Actuator Dynamics Curriculum, a procedure that initializes joint stiffness at a high value and anneals it toward the system-identified value as completed episode lengths grow. Using a cart-pole system as a representative example, we show that higher closed-loop joint natural frequency under critical damping enlarges the viability kernel of the underlying Markov Decision Process, increasing the fraction of initial states from which the task is feasible. We validate the kernel monotonicity on the cart-pole and apply the curriculum to a quadrupedal-to-handstand transition on the Boston Dynamics Spot, a narrow-viability task where training under fixed identified stiffness plateaus at a policy that never completes the transition. The trained policy executes the transition in simulation across 10 seeds and transfers to hardware. More broadly, our results suggest that simulated actuator dynamics is a useful axis along which to design curricula for tasks in which exploration is bottlenecked by termination conditions rather than by reward signal.
Read the full abstract
Computer Vision
Learning Global Camera Poses from Noisy View-Graphs for Structure from Motion
Camera pose estimation is a key step in 3D reconstruction and view-synthesis pipelines. We present a deep, global Structure-from-Motion framework based on learned view-graph aggregation. Our method employs a permutation-equivariant, edge-conditioned graph neural network that takes noisy pairwise relative poses as input and outputs globally consistent camera extrinsics. The network is trained without ground-truth supervision, relying solely on a relative-pose consistency objective. This is followed by 3D point triangulation and robust bundle adjustment. Our approach is efficient, scalable to more than a thousand images, and robust to graph density. We evaluate our method on MegaDepth, 1DSfM, Strecha, and BlendedMVS. These experiments demonstrate that our method achieves superior rotation and translation accuracy compared to deep track-centric methods while registering more images across many scenes, and competitive results compared to state-of-the-art classical pipelines, while being much faster.
Read the full abstract
Networking
Prototyping QoE-Aware Rate Adaptation in Cellular Networks with Commercial Applications
Prior work has shown that QoE-aware resource sharing for real-time interactive video can support up to three times more simultaneous sessions at acceptable quality compared to rate-fair allocation. However, the required capabilities (QoE-targeted encoding, runtime spatial complexity estimation, and rich application-network APIs) are not yet available in commercial deployments. In this paper, we take an evolutionary approach: we design a system that delivers QoE-aware resource allocation using only capabilities that can be assembled in a lab today. We extend the utility-based allocation framework to the radio resource domain by introducing composite spatial complexity, which combines a session's video spatial complexity with its time-variant spectral efficiency into a single resource demand function. To operate with commercial real-time video streaming applications that use rate-based congestion control and lack capability to measure QoE, we use external tooling for QoE measurements. We develop an incremental reallocation algorithm with per-interval limits that encode both the congestion control algorithm's speed constraint and that spatial complexity estimates are reliable only near the current rate. The resulting prototype combines external QoE measurements with congestion-signal-based rate steering and does not require modification to commercial applications. We chart an evolution path from this prototype toward full QoE-aware resource sharing, mapping emerging standards (IETF SCONE, CAMARA, Media over QUIC) to the progressive capabilities they enable.
Read the full abstract
Cryptography and Security
Audio Deepfake Detection Using Temporal Coherence Analysis
The proliferation of AI-generated audio (so-called "deepfake" audio) poses significant threats to information integrity, from voice cloning fraud to synthetic music copyright disputes. We present a temporal coherence analysis framework built upon Contrastive Language-Audio Pretraining (CLAP) embeddings that spans speech, instrumental music, and music with vocals. By computing pairwise cosine similarities between audio segment embeddings and extracting statistical features from the resulting distributions, we train lightweight ensemble classifiers that reliably distinguish authentic from synthetic audio. Our work provides an interpretable, computationally efficient alternative to common deep learning methods while still achieving competitive performance across speech and music domains. Further, we reveal two notable empirical findings about audio deepfakes: (1) a feature-label inversion phenomenon in which 21 of 29 statistical features reverse their discriminative direction between training and in-the-wild deployment, and (2) a speech–music direction reversal in which entropy discriminates in opposite directions for speech and music deepfakes.
Read the full abstract
Econometrics
Two Margins in Difference-in-Differences with a Continuous Treatment
This paper studies difference-in-differences with staggered adoption and a continuous, time-invariant dose. Each cohort-time comparison contains two margins. The level margin is the average treatment effect at realized doses. Under level parallel trends it equals the level contrast between the treated cohort and not-yet-treated controls. The response margin is the within-cohort slope of the outcome change on dose. It uses no controls, and its causal interpretation requires a response parallel trends assumption and a restriction on selection on gains. We show that the continuous-dose OLS coefficient in each cohort-time comparison is a convex combination of the response index and the level contrast per unit of mean dose, with a mixing weight that depends on the not-yet-treated share. We provide estimators of both margins, joint inference across cohort-time comparisons and event-time aggregates, and a covariate-adjusted extension. In an application to hydraulic fracturing, the level leads reject a joint zero restriction, whereas the response-index leads do not. The continuous-dose OLS coefficient draws primarily on the level margin. We report the level and response margin separately.
Read the full abstract
General Relativity and Quantum Cosmology
Shadows and gravitational perturbations of black bounces
We investigate axial gravitational perturbations and shadow formation in black-bounce geometries supported by anisotropic fluids within general relativity. Two distinct classes of solutions are analyzed: a symmetric Simpson–Visser-like black bounce and an asymmetric deformation of the Reissner–Nordström metric, which can feature either a bounded interior region or an unbounded wormhole-like extension. Our findings indicate that configurations with horizons, whether symmetric or asymmetric, exhibit single-barrier effective potentials and quasinormal frequencies nearly indistinguishable from those of standard black holes. In contrast, horizonless symmetric configurations may develop multiple potential barriers, giving rise to gravitational-wave echoes whose amplitude and separation depend sensitively on the bounce parameter $a$ and the density parameter $\rho_0$. Horizonless asymmetric solutions, on the other hand, exhibit regularized effective potentials that do not produce echoes and remain phenomenologically close to the Reissner–Nordström case. From the optical perspective, both classes of solutions with horizons exhibit a single relevant photon sphere and shadow profiles nearly indistinguishable from those of their black-hole counterparts, while horizonless symmetric configurations can sustain multiple photon rings. Horizonless asymmetric configurations may instead support either a single photon ring or none. Finally, we recover the standard eikonal correspondence between quasinormal modes and the instability of circular null geodesics for configurations with horizons. For horizonless geometries, however, the emergence of multiple potential barriers prevents a straightforward extension of this relation, although the strong similarity between the wave and optical sectors suggests that a meaningful phenomenological correspondence may still persist.
Read the full abstract
Machine Learning
Efficient Fairness Auditing Across Guidance Scales in Text-to-Image Diffusion Models via Causal Abstraction
Fairness auditing of text-to-image diffusion models often requires generating large numbers of images across sampling configurations, making comprehensive evaluation computationally expensive. We propose a causal-abstraction-based audit instrument for efficiently evaluating fairness under interventions on the classifier-free guidance scale. Given a fixed prompt and a target feature function, we represent the diffusion process as a low-level structural causal model and construct a corresponding high-level model over abstract denoising states. We characterize the projected causal structure, establish identifiability of the fairness-relevant interventional query, and provide sufficient conditions under which the high-level model preserves this query. A probabilistic transformer implements the high-level model as an amortized predictor of target-feature distributions across guidance scales. Experiments evaluate distributional fidelity, fairness-query accuracy, and computational efficiency. We present two auditing demonstrations: one using standard Stable Diffusion 1.5 and another using StayFair, a fairness-enhanced Stable Diffusion model, to examine their behavior across guidance scales.
Read the full abstract
General Relativity and Quantum Cosmology
Space-Based Lock-In Determination of Newton's Gravitational Constant
Inspired by a recent ground-breaking space-based measurement of Newton's gravitational constant, we outline a proposal for a general layout of an optimized space-based approach that combines the advantages of a largely disturbance-free environment with lock-in techniques, null coordinates and the method of covariances. The method of covariances, and cross-covariances, is used for an optimized statistical analysis of data originating from two distinct experimental channels. One of two distinct measurement channels relies on an approximately constant acceleration in the gravitational field of static sources masses, while the second channel involves a lock-in measurement based on test and source masses with oscillating coordinates. Conceivable mission parameters are analyzed in terms of the optimum scale for the source masses and launch capabilities of currently available space launch systems. A candidate location for a largely disturbance-free position in geostationary orbit is identified.
Read the full abstract
Numerical Analysis
A Stress-Based Estimator for Pressure and Stress Recovery from Velocity Measurements
Non-invasive pressure field estimation from velocity measurements is a longstanding engineering problem. We propose, analyze, and test a pressure-recovery method that computes a full stress field from velocity measurements, and leaves the pressure estimation as a cheap post-processing step. The method relies on a stress-velocity first order formulation of the Navier-Stokes equations, and we show that the formulation accounts for deviations from incompressibility in the measured velocity field by construction. In addition, we theoretically establish the convergence of the finite element (FE) approximation scheme, the stability of the stress recovery with respect to finite-resolution velocity measurements, and then validate this theory numerically. Our results show that the proposed estimator is robust in convective flow regimes and remains accurate at reduced spatial resolution, improving upon state-of-the-art pressure-recovery strategies.
Read the full abstract
Human-Computer Interaction
Integrating Multi-Source Feedback in Computational Design
In real-world design practice, evaluations rarely rely on a single source of judgment. Designers routinely combine expert opinions, empirical studies, and computational models, each with distinct strengths and limitations. While machine learning offers methods to integrate multiple feedback sources, these approaches remain largely inaccessible to designers without technical expertise. In this paper, we explore how to integrate multiple feedback sources, primarily through: (1) a practical approach for multi-source integration, and (2) its implementation in MUSE, a no-code tool that allows designers to combine and balance diverse sources. Our technical findings show that independent modeling of multiple evaluation sources enables exploration across heterogeneous feedback, accommodates different evaluation speeds, surfaces disagreements between sources, and supports an adaptable evaluation setup that designers can reconfigure during their process. In a visualization design study, participants navigated their own judgments alongside simulator feedback, reporting a perception of enhanced confidence and flexibility. Our results highlight the viability of multi-source integration to support computational design, offering a step toward bridging the gap between advanced optimization methods and design practice.
Read the full abstract
Computer Vision
Infra-Bench CLS: A Global, Open-Source Benchmark for Critical Infrastructure Classification with Earth Observation Foundation Models
Critical infrastructure location data is often incomplete and unevenly distributed globally, especially in developing regions. Earth observation foundation models are proposed as a new step in enabling us to more efficiently understand the natural and built environment, raising questions as to their effectiveness in performing challenging downstream tasks. Yet, foundation models remain largely untested for detecting and classifying the facility-scale critical infrastructure that underpins a range of important societal and economic functions. Subsequently, Infra-Bench CLS is introduced as a benchmark to test foundation models on 18,756 Sentinel-1 SAR and Sentinel-2 multispectral facility-scale critical infrastructure asset images covering seven continents and 13 infrastructure classes, with results reported for the 10 retained classes. Using linear probing and fine-tuning for two training dataset levels (1.0x and 0.3x), seven foundation models are evaluated (SatlasPretrain S2, SatlasPretrain S1, CROMA, Prithvi-EO-2.0, AlphaEarth Foundations, OlmoEarth v1.1-Base, and DINOv3 ViT-L/16). When comparing macro F1 scores to a ResNet-18 supervised baseline of 39.2 percent, the best foundation model achieved 57.9 percent, a 48 percent improvement. Top performing classes were airports (F1 85.3 percent), train stations (F1 82.1 percent), and data centers (F1 77.6 percent). By contrast, many of the power sector classes perform poorly (F1 27.5-46.2 percent). These findings suggest foundation models can enable superior critical infrastructure classification, but future work should evaluate performance on higher-resolution imagery, particularly for poorly performing sectors, such as power.
Read the full abstract
Earth and Planetary Astrophysics
Detecting White Dwarf Exoplanets in the Roman Era
White dwarfs (WDs) provide an unparalleled opportunity for detecting and characterizing exoplanets because of the large planet-to-star size ratio. However, only one transiting WD exoplanet is confirmed to date, and occurrence rate constraints remain loose because the rapid transits and weak phase curve modulations of WD exoplanets require high-precision, high-cadence photometry. The Galactic Bulge Time Domain Survey (GBTDS) of the Nancy Grace Roman Space Telescope will provide exactly this by observing thousands of WDs at 12.1-minute cadence. The expected WD exoplanet yield of this survey is not yet quantified. Here, we predict the GBTDS WD exoplanet yield over its first three high-cadence seasons, through both transiting and non-transiting (phase curve) channels. We run Monte Carlo detection probability simulations over a grid of planet radius, orbital period, and host properties for a Besançon synthetic population of WDs. Jupiter-sized planets dominate the transit yield, with a period-averaged effective survey size of $\overline{N}_\mathrm{eff}\approx8$, requiring an underlying occurrence rate of $\overline{\eta}_1\approx12\%$ for one expected detection, followed by $\overline{N}_\mathrm{eff}\approx2.6$ ($\overline{\eta}_1\approx38\%$) for Neptune-sized and $\overline{N}_\mathrm{eff}\approx1.8$ ($\overline{\eta}_1\approx55\%$) for sub-Neptune-sized planets. Source blending is the major observational challenge in the crowded GBTDS fields, removing $\approx95\%$ of the raw predicted yield. Phase curve provides marginal detections, adding at most $\approx4\%$ to the transit yield. A null detection by GBTDS would constrain $\eta_\mathrm{WD}$ to $< 28\%$ for Jupiter-sized planets at $95\%$ confidence. GBTDS will deliver the first WD exoplanet occurrence-rate constraints beyond the solar neighborhood and identify prime targets for future characterization.
Read the full abstract
Probability
Gaussian Approximation in Wasserstein-p: Markovian Martingale Differences and Stochastic Approximation
We establish nonasymptotic Gaussian approximation bounds in higher-order Wasserstein distance $W_p$, $p\geq2$, for deterministic matrix-weighted sums of multivariate Poisson martingale differences generated by uniformly ergodic Markov chains. Under an $L^{(2+\eta)p}$-moment condition, $\eta>0$, we obtain the explicit bound $$O\left(p^3\|A\|_4^2+pd^{1/4}\|A\|_2^{1/2}\|A\|_4^2\right),$$ where $A\in\mathbb{R}^n$ collects the moment sizes of the weighted increments. When these sizes are of order $n^{-1/2}$, the bound achieves the first optimal $O(n^{-1/2})$ rate for fixed $p$ and $d$. Consequently, we also obtain the first optimal $O(n^{-1/2})$ $W_p$ Gaussian approximation rate for multivariate additive functionals of uniformly ergodic Markov chains. Our analysis develops two techniques for addressing the interplay between higher-order Wasserstein distance and temporal dependence. First, building on the Ornstein–Uhlenbeck relative-score approach of Fang and Koike (2023), we formulate the bound in terms of antisymmetric Stein couplings while retaining the conditional tensor structure. Second, we develop a refresh-then-maximal coupling that combines an independent first-step resampling, which preserves the desired Stein identity, with a subsequent maximal coupling that provides effective control of the coupling increment. We apply these bounds to nonlinear Markovian stochastic approximation by decomposing normalized errors into matrix-weighted Poisson martingale sums and remainders. This yields the first sharp $W_p$ Gaussian approximation rates for both last iterates and Polyak–Ruppert averages. In particular, the optimized rates are $O(n^{-1/3})$ for polynomial-stepsize last iterates and $O(n^{-1/4})$ for Polyak–Ruppert averages, with an $O(n^{-1/2})$ rate for harmonic-stepsize last iterates under an additional spectral condition.
Read the full abstract
Statistical Mechanics
Exact asymptotic equivalence between precisions of record-times and integrated currents
We establish an exact equivalence between the fluctuations of integrated currents and the temporal statistics of current record events in continuous-time Markov jump processes. Specifically, we show that, under minimal assumptions, the Fano factor of an integrated current coincides exactly with the squared coefficient of variation of the inter-record waiting times. This result is obtained by showing that inter-record times induce a renewal structure, since each record event resets the system to a fixed state, and that their generating function can be expressed exactly in terms of a combinatorial organization governed by Narayana numbers. The equivalence extends previous results obtained for unicyclic networks to arbitrary finite-state graphs and provides a unified representation of current fluctuations in terms of record-time observables. The results are derived within a purely probabilistic framework and clarify the structural origin of current precision in terms of first-passage-like event statistics.
Read the full abstract
Machine Learning
Unthrottling the Tanh Jacobian in SAC: A Negative Result on Bang-Bang Control and MetaDrive
Soft Actor-Critic (SAC) represents a continuous policy as an unbounded Gaussian that is squashed by tanh. The Jacobian of that map is $\partial a/\partial u = 1-a^2$, which vanishes as $|a|\to 1$. A natural concern is that this throttle starves the actor of critic signal exactly where extreme actions (full brake, full throttle) are optimal. We test a minimal intervention that restores the missing signal: one extra term in the actor loss whose gradient on the pre-tanh mean is the detached action-gradient of $Q$, with no gain parameter. On a minimum-time double integrator whose optimum is bang-bang at the action bounds, vanilla SAC already reaches near-optimal return ($-31.6$ vs. a calibrated optimum of $-30.3$) across ten paired seeds. An ungated bypass does saturate the policy (99% of eval steps with $|a|\ge 0.9$) and collapses return to $-195.5$. A gated bypass that fires only on the flat shoulder $|a|\in[0.9,0.999]$ also fails, and does so without leaving a saturated policy. Warm-started MetaDrive fine-tuning shows the same pattern: the bypass does not improve return, and where collision rate falls it is typically traded for out-of-road departures. Auto-tuned entropy coefficient rises against the bypass, which is a push toward the tails. The Jacobian effect is real. Treating it as a bug to be undone is not free, and on the tasks studied here it is not helpful. Saturating a bound is not the same as solving a problem whose optimum lives on that bound.