Quantum Physics
Measuring Clock Precision Without an Ideal Time Reference
Pablo Á. Domínguez, Christopher Chubb and Mischa P. Woods
Characterising physical devices typically requires comparison with more precise reference standards. For clocks, however, this approach is problematic because no physically perfect time reference exists; real clocks have finite, model-dependent uncertainty. We address whether multiple imperfect clocks can collectively determine their precision without an absolute reference. This paper provides analytic bounds and protocols for quantifying clock precision using only multiple copies of the same clock, which need not be synchronised and may have manufacturing variations or environmental coupling. Beyond clock characterisation, these methods enable precise measurements f sub-period intervals and relativistic effects such as time dilation and Doppler shifts, providing tools for
quantum metrology and fundamental physics.
Read the full abstract ▾
High Energy Astrophysical Phenomena
Discovery of minute-scale polarization angle rotation in a supermassive black hole jet
Aristeidis Polychronakis, Ioannis Liodakis, Dmitry Blinov et al. (+7 more)
Relativistic jets from supermassive black holes oriented towards our line of sight called blazars have puzzled the astronomical community for decades. Their fast jets and preferential
alignment creates a fog of relativistic effects that obscures their true properties. Optical polarimetry - tracing the uniformity, direction, and
evolution of the
magnetic field in the jets, has revealed a unique phenomenon to blazars; coherent changes of the
magnetic field manifesting as rotations of the polarization angle ($\Psi$). The origin of those rotations has been debated for decades with several proposed models to explain them. We detected an extremely fast - minute time-scale - rotation in the polarization angle of $\sim136^\circ$ over the span of about 80 minutes, making it the fastest known rotation by almost an order of magnitude. Our optical polarization observations combined with
radio very long baseline interferometry allow us to reject different models of particle acceleration and strongly point to relativistic effects and the interaction of standing and moving shock waves as the origin of the extreme polarization variations in jets.
Read the full abstract ▾
All authors Aristeidis Polychronakis, Ioannis Liodakis, Dmitry Blinov, Anastasia Glykopoulou, Sebastian Kiehlmann, Karan Pal, Georgios F. Paraschos, Lena Debbrecht, Swati Ravi, Herman L. Marshall
Astrophysics of Galaxies
$\textit{N\=ap\=owaw\=a`enakaulua}$: a close quasar pair at cosmic dawn in the Aether survey
Silvia Onorato, Emanuele Paolo Farina, Roberto Decarli et al. (+32 more)
We present the discovery and first characterization of $\textit{N\=ap\=owaw\=a`enakaulua}$: the most compact known
quasar pair candidate in the Epoch of Reionization, with a projected separation of only $0.98^{\prime\prime}$ ($\approx 5.56$ pkpc at $z \sim 6.08$) and a velocity offset of $\Delta v \simeq 15 \text{ km s}^{-1}$. Using spectroscopy from JWST/NIRSpec IFU, as part of the $\textit{Aether survey}$, Gemini/GNIRS and GMOS we characterize the continuum and emission line properties of the two
quasars, enabling constraints on their black hole masses ($M_{\text{BH}} \simeq 1.5 \times 10^9$ and $4.0 \times 10^8\,M_\odot$),
accretion states ($\lambda_{\text{Edd}} \simeq 1.23$ and $1.07$), and systemic
redshifts ($z_{\rm sys} \simeq 6.0781$ and $6.0785$). We place this pair in the cosmological context by comparing the chance of occurrence to extrapolations of high-$z$
quasar clustering measurements at larger scales, resulting in a $\sim 2-3$ order of magnitude excess with respect to expectations. This excess hints at an underlying population of merger-triggered
quasar pairs at
cosmic dawn or to a denser-than-average
galactic environment. $\textit{N\=ap\=owaw\=a`enakaulua}$ thus represents a crucial signpost to investigate the mass assembly of the first massive
galaxies and black holes.
Read the full abstract ▾
All authors Silvia Onorato, Emanuele Paolo Farina, Roberto Decarli, Klaudia Protušová, Elia Pizzati, Debora Pelliccia, Devin S. Chu, Larry Lindsey Kimura, Leinani Lozi, Emily Peavy, Fabrizio Arrigoni-Battaia, Eduardo Bañados, Aaron J. Barth, Silvia Belladitta, Manuela Bischetti, Sarah E. I. Bosman, Hyunseop Choi, Tiago Costa, Anna-Christina Eilers, Xiaohui Fan, Simona Gallerani, Thomas R. Geballe, Anniek J. Gloudemans, Eunchong Kim, Brian C. Lemaux, Weizhe Liu, Yoshiki Matsuoka, Chiara Mazzucchelli, Atsuko Nitta, Jan-Torge Schindler, Andrew W. Stephens, Hyewon Suh, Fabian Walter, Feige Wang, Jinyi Yang
High Energy Physics (Phenomenology)
Disentangling $V_{ud}$ from New Physics in Beta Decays with the FCC-ee
Lemonia Gialidi, Elie Hammou, Kamil Laurent et al. (+3 more)
Superallowed
nuclear $\beta$-decays enable the most precise determination of the CKM element $V_{ud}$, but they are only sensitive to an effective combination of $V_{ud}$ and New Physics contributions to left-handed charged currents, which low-energy data alone cannot disentangle. Here we determine $V_{ud}$ simultaneously with the relevant Wilson coefficients of the Standard Model Effective Field Theory (SMEFT) by combining, within the SMEFiT framework, a dedicated likelihood for superallowed transitions, including radiative and
nuclear-structure uncertainties, with LEP and LHC world data and HL-LHC and FCC-ee projections. Current
collider data and HL-LHC projections only partially lift the degeneracy between $V_{ud}$ and possible New Physics effects. Instead, the FCC-ee fully decouples $V_{ud}$ from New Physics, recovering SM-level precision while significantly tightening the constraints on the SMEFT Wilson coefficients. This conclusion is robust against variations in flavour assumptions and theory uncertainties. Our results highlight the complementarity between future
colliders and low-energy experiments, and pave the way towards a SMEFT-consistent interpretation of first-row CKM unitarity tests and the Cabibbo angle anomaly.
Read the full abstract ▾
All authors Lemonia Gialidi, Elie Hammou, Kamil Laurent, Vaisakh Plakkot, Juan Rojo, Jordy de Vries
High Energy Physics (Phenomenology)
Neutrino Direct Simulation Monte Carlo: Accurate modeling of out-of-equilibrium decaying cosmological relics
Kensuke Akita, Miguel Escudero, Oleksii Ihnatenko et al. (+1 more)
We consider nonstandard cosmologies in which decaying relics inject energy at MeV temperatures, driving
neutrinos out of
equilibrium. The subsequent
neutrino evolution shapes predictions for the
cosmic radiation density and
primordial light element abundances. Earlier work introduced
Neutrino Direct Simulation Monte Carlo ($\nu$DSMC), which describes
neutrino transport through Monte Carlo sampling of particle interactions. Here, we extend this method by jointly evolving the relic population,
neutrino spectra, electromagnetic
plasma, and
cosmic expansion, and coupling this
evolution to Big Bang nucleosynthesis. The framework includes direct and inverse decays of the relic,
neutrino flavor conversion,
quantum statistics, and pion production by energetic
neutrinos. To reveal the physical impact of these effects and independently test $\nu$DSMC, we compare several approaches to
neutrino evolution across a range of nonstandard cosmological scenarios. We find excellent agreement with independent solutions of
quantum kinetic or quasi-classical Boltzmann equations. Compared with these calculations, $\nu$DSMC includes a broader range of physical processes and is substantially faster for strongly nonthermal
neutrino populations. This makes broad, robust parameter scans for nonstandard cosmologies practical. We show that when departures from thermal
equilibrium are large, approximations that average over
neutrino momenta can give qualitatively incorrect predictions, including the wrong sign of the change in the effective number of
neutrino species and substantially incorrect
primordial nuclear abundances. We conclude that the extended $\nu$DSMC framework enables reliable cosmological tests of new physics by following
neutrino thermalization and its observable consequences.
Read the full abstract ▾
All authors Kensuke Akita, Miguel Escudero, Oleksii Ihnatenko, Maksym Ovchynnikov
Quantum Physics
Characterising the precision of a clock without any external time reference
Pablo Á. Domínguez, Christopher Chubb and Mischa P. Woods
Determining a clock's precision usually requires comparison with a better time reference, yet no ideal reference is physically attainable. We show how a single clock can characterise its own precision using only a signal splitter and delayed trigger, whose uncertainty is explicitly accounted for. The resulting
estimator depends only on tick-count statistics. A purpose-built electronic
circuit validates the method without an external timekeeper or a second clock.
Read the full abstract ▾
Strongly Correlated Electrons
Emergent low-energy many-body Hilbert spaces in Chern bands
Chen-Xin Jiang, Zi-Xiang Hu and Bo Yang
We show that in Landau levels and more generic Chern bands, the most relevant many-body subspace selected by realistic two-body interactions alone (e.g., in the limit of vanishing bandwidth and band gap) depends strongly on the particle density and can be constructed systematically. Using the well-defined hierarchical structure of conformal
Hilbert spaces (CHS), each with a highest-density
vacuum state, we compute, for both fermionic and bosonic systems, the many-body
quantum metric that quantifies the interaction energy scales of these subspaces. For a large family of short-range, Coulomb-based interactions, particles preferentially occupy the many-body states in the smallest accommodating CHS. This allows us to describe the low-energy physics of the interacting system in terms of the appropriate anyonic degrees of freedom. Our results also explain why, in any Chern band (in
lattice or continuum systems), band mixing tends to be suppressed even in the limit of strong interaction; this suppression is stronger when the band deviates less from the ideal trace condition and when the two-body interaction is shorter-ranged.
Read the full abstract ▾
Quantum Physics
Probing the classical complexity of quantum dynamics experiments
Thomas Schuster and Andreas Elben
A confluence of recent works has shown that many
quantum circuits and dynamics are efficiently simulable by classical
algorithms that track local information, even when conventional complexity measures such as the
entanglement and magic are high. Here, we introduce a novel measure of complexity, the reactivity, to capture this new method of classical attack. Unlike conventional complexity measures, the reactivity does not capture a property of a
quantum state or operator in isolation, but rather a
quantum experiment as a whole. We provide numerical and rigorous evidence that
quantum experiments with low reactivity are simple by a host of measures: they are efficient to classically simulate, learn, and fast-forward. This motivates the search for
quantum experiments with high reactivity, which may evade these simplistic features. To this end, we introduce easily implementable experimental protocols—dubbed Pauli path spectroscopy—that allow one to efficiently measure the reactivity of any
quantum experiment of interest. Our protocols are applicable even when the experiment itself is beyond the reach of classical simulation.
Read the full abstract ▾
High Energy Physics (Theory)
Localizing the M-Theory Path Integral
Pietro Benetti Genolini, Florian Gaar, Jerome P. Gauntlett et al. (+2 more)
We propose a supersymmetric localization formula for the local perturbative contribution to the M-theory path integral, reducing it to an integral over equivariant flux data. Relative equivariant extensions of the eleven-dimensional Chern–Simons couplings determine the localized action, while for M2-brane backgrounds a tangent-cotangent character computes the one-loop determinant. This gives the first closed form expression for the complete perturbative ABJM partition function at general squashing and chemical potentials, and new analytic results for the ABJM topologically twisted index describing black holes. Applied to the non-toric $V^{5,2}$ theory, the same prescription yields a closed analytic Airy constant previously known only numerically. Our formulas reproduce all available independent field theory results, and provide analytic predictions at general parameters where no field theory answer is currently known. We derive an all-genus result for the Type IIA expansion which is an equivariant form of the usual topological string constant map result, and also extract integrated correlators.
Read the full abstract ▾
All authors Pietro Benetti Genolini, Florian Gaar, Jerome P. Gauntlett, Jaeha Park, James Sparks
High Energy Physics (Phenomenology)
Impact of Helicity-Isospin Convention Matching on the Extraction of $\alpha$ from $B^0\to(\rho\pi)^0$
Radek Žlebčík
We show that the $\rho\to\pi\pi$ helicity convention used in the time-dependent Dalitz-plot measurements of $B^0\to(\rho\pi)^0$ decays does not match the convention assumed in the subsequent isospin analysis. Once corrected, the tension between the direct determination of the unitarity-triangle angle $\alpha$ from these decays and its indirect determination from global fits decreases from $2.9\sigma$ to $2.1\sigma$. Combined with the $B\to\pi\pi$ and $B\to\rho\rho$ modes, the direct determination becomes $\alpha = (87.0^{+11.8}_{-4.7})^\circ$, compared with the original world average $(84.1^{+3.7}_{-3.0})^\circ$.
Read the full abstract ▾
Astrophysics of Galaxies
Balmer Break Measurements Trace Departures from the Star-Forming Main Sequence at $1<z<4$ with JWST/NIRSpec
Suprabhas Narisetty, Yongda Zhu, Eiichi Egami et al. (+4 more)
We present JWST/NIRSpec PRISM observations of
galaxies at $1<z<4$ to study the relation between Balmer-break strength ($D_{\rm Balmer}$) and recent
star formation. Using NIRSpec spectra from the JADES Transient Survey in GOODS-S, we measure $D_{\rm Balmer}$ and hydrogen recombination-line equivalent widths, and infer
galaxy properties through spectral energy distribution (SED) fitting of the accompanying photometry. We find a clear anti-correlation between $D_{\rm Balmer}$ and the rest-frame equivalent width of H$\alpha$+[\ion{N}{2}], such that
galaxies with stronger breaks show weaker nebular emission and lower recent
star formation activity. $D_{\rm Balmer}$ also correlates with mass-weighted
stellar age, supporting its use as an empirical tracer of evolved
stellar populations at these
redshifts.
Star formation rates inferred from H$\alpha$, Pa$\gamma$, and SED fitting are broadly consistent, with scatter likely driven in part by
dust attenuation, line-measurement uncertainties, and the different timescales probed by each indicator.
Galaxies with stronger $D_{\rm Balmer}$ also tend to lie further below the
star-forming main sequence at fixed
stellar mass, indicating that $D_{\rm Balmer}$ is linked to recent suppression of
star formation. In contrast, we find only a weak relation between $D_{\rm Balmer}$ and SED-derived
dust attenuation, suggesting that
dust is not the main driver of the observed trends. These results show that $D_{\rm Balmer}$, combined with recombination-line measurements, is a useful empirical diagnostic of
galaxies transitioning away from the
star-forming population at
cosmic noon.
Read the full abstract ▾
All authors Suprabhas Narisetty, Yongda Zhu, Eiichi Egami, Stefano Carniani, Francesco D'Eugenio, Christa DeCoursey, Christopher N. A. Willmer
Astrophysics of Galaxies
MUSE spectroscopy of the compact dual AGN in the $z=3.273$ radio-loud gravitational lens MG B2016+112
Jianghao Huyan, Júlia M. Sisk-Reynés, Daniel A. Schwartz et al. (+4 more)
Dual active
galactic nuclei (AGN) are unique laboratories for studying how
galaxy mergers trigger
accretion onto supermassive black holes (SMBHs). However, at
redshifts $z>1$, only few dual AGN are confirmed at projected separations $\gtrsim$ 1 kpc for the two SMBHs. We present archival MUSE spectroscopy of the three resolved lensed images of the
radio-loud system MG B2016+112. At z = 3.273 and at a projected separation of 175 pc, this the most compact confirmed dual AGN at z > 1. We detect Ly$\alpha$, C IV, N V, He II, and C III] emission across images A$--$C after correcting for ISM absorption at $z \sim 0$. We find that the Ly $\alpha$, C IV, N V lines are affected by saturated absorption at $z \sim 3.27$. The He II / C III] emission-line ratios are consistent between images A and B but differ significantly from image C. We fit the C IV emission regions of images A$--$C by allowing for up to two Gaussians and use broad component to place lower limits on their respective SMBH masses: $\gtrsim 2 \times {10}^{7} M_\odot$, $\gtrsim 2.3 \times {10}^{7} M_\odot$, and $\gtrsim 1.5 \times {10}^{7} M_\odot$. Correcting for their strong
lensing magnifications suggests Eddington ratios of $\lesssim$ 0.47, 0.51, and 0.04, respectively. Together with previous VLBI and X-ray analyses, our results provide strong, independent spectroscopic evidence that MG B2016+112 hosts two AGN. Our findings motivate deep, spatially resolved near-IR spectroscopy to independently revisit our mass estimates and determine the physical conditions of the host
galaxy, which has likely undergone a major merger.
Read the full abstract ▾
All authors Jianghao Huyan, Júlia M. Sisk-Reynés, Daniel A. Schwartz, Varsha P. Kulkarni, Anna Barnacka, Adi Foord, Charlotte A. Eades
High Energy Physics (Theory)
Cardy random matrix theory
Marco Ambrosini, Alexandre Belin, Jan de Boer et al. (+1 more)
We construct and study a “Cardy random
matrix model”: a random
matrix theory whose only input, in a maximal-ignorance prescription, is the constraint of $\mathbb{S}$-modular invariance. We regularize this constraint by coarse-graining over energy windows of fixed width, and obtain a potential for a single random
matrix whose
eigenvalues play the role of the scaling dimensions of primaries in a 2d CFT. The potential drives the coarse-grained density of states to the Cardy profile above a gap at the black hole threshold. We perform Monte Carlo simulations to confirm this behavior and find good agreement for the coarse-grained density of states, while at the same time revealing GUE statistics for nearby
eigenvalues. We identify the scaling limit in which these phenomena emerge. In this limit, the model realizes a hierarchy of scales: at the macroscopic scale given by the window size, the occupations are frozen and the bootstrap constraint rigidly fixes the spectrum to the Cardy profile, while at the mesoscopic scale (inside a window) the Vandermonde repulsion rearranges the
eigenvalues as in a random
matrix theory. It is only at the microscopic scale of the mean level spacing that the individual levels are resolved. Accordingly, our random
matrix model displays an effective Thouless time, set by the inverse energy width of the coarse-graining window. Our model provides a first step towards solving the
matrix/tensor model dual to 3D
quantum gravity on a computer.
Read the full abstract ▾
All authors Marco Ambrosini, Alexandre Belin, Jan de Boer, Julian Sonner
Atomic Physics
Microwave Power-to-Frequency Transduction in a Magnetically Initiated Rydberg Dissipative Time Crystal
Darmindra Arumugam
Microwave power-to-frequency transduction is demonstrated using the emergent oscillation frequency of a magnetically initiated Rydberg dissipative time
crystal (DTC). In a room-temperature 87Rb vapor, a ~14 G
magnetic field establishes a self-sustained DTC near 21 kHz, while a near-resonant 8.325 GHz field coupling the 63D3/2 to 64P3/2
manifold continuously shifts the autonomous oscillation frequency. The resulting field-to-frequency transfer function is strongly nonlinear, with a total fundamental-frequency excursion of ~6.7 kHz and peak responsivity approaching 1 kHz/(mV/cm). The nonlinear transition is substantially steeper than a simple quadratic saturation response and is captured by a self-consistent mean-field picture in which microwave-driven Rydberg-population redistribution modifies the collective resonance condition. Higher DTC harmonics preserve the same normalized transfer function while exhibiting approximately linear growth of absolute responsivity with harmonic order. These results establish an intrinsic many-body atomic route to RF amplitude-to-frequency conversion, enabling continuous frequency-domain sensing through an emergent atomic frequency without an externally engineered RF feedback oscillator, repeated optical spectral scans, or resolved Autler Townes splitting.
Read the full abstract ▾
Computer Vision
SynDORBench: Evaluating LVLM Perceptual Robustness Under Physically Constrained Visibility Conditions
Jeremy Stephen Gabriel Yee, Zhengkui Wang, Zhiyuan Zhang et al. (+5 more)
Large
vision-
language models (LVLMs) have demonstrated remarkable performance on
multimodal reasoning benchmarks, yet their perceptual reliability under physically constrained imaging conditions remains poorly understood. Existing evaluations predominantly assume ideal visual inputs and therefore fail to characterize how camera distance, illumination, viewpoint, and pixel density fundamentally affect semantic recoverability. We introduce SynDORBench, the first physically grounded
benchmark for evaluating LVLM perceptual robustness under DORI-calibrated conditions aligned with human visual capability standards. SynDORBench comprises over 54k question–answer pairs generated through a controllable synthetic pipeline that systematically varies viewing distance, lighting, camera geometry, and action pose according to physically interpretable pixel-density regimes. To support scalable low-visibility supervision, we further propose a discernibility annotation framework that propagates human perceptual labels using mask-conditioned statistical features and ensemble learning. We evaluate 16 open-source LVLMs, a commercial LVLM baseline, and YOLO11x across human-presence classification and action recognition tasks under progressively degraded visibility conditions. Our results reveal that perceptual failure in LVLMs is strongly governed by pixel density and physical imaging constraints rather than model scale alone. Surprisingly, several compact open-source LVLMs outperform larger commercial baselines and substantially exceed YOLO11x robustness under long-range and low-light conditions. SynDORBench establishes a new
benchmark paradigm for physically grounded
multimodal evaluation, enabling systematic analysis of LVLM reliability under real-world perceptual constraints and direct comparison against human visibility thresholds.
Read the full abstract ▾
All authors Jeremy Stephen Gabriel Yee, Zhengkui Wang, Zhiyuan Zhang, Avinash Anand, Timothy Liu, Benedict Chan, Aik Beng Ng, Simon See
Solar and Stellar Astrophysics
Measuring the Expansion of Solar Magnetic Fields with Multiline Inversions from Sunrise III
Juan Carlos Trelles Arjona, María Jesús Martínez González, Christoph Kuckein et al. (+34 more)
Magnetic fields in the
solar atmosphere are expected to expand with height, forming
magnetic canopies that couple the photosphere and chromosphere. Except for sunspots, direct measurements of this expansion remain scarce because they require simultaneous
magnetic diagnostics over a broad range of atmospheric heights. We analyze four consecutive high-resolution spectropolarimetric raster scans of an emerging flux region observed with the Sunrise III/SCIP instrument. Simultaneous inversions of five photospheric Fe I lines and two chromospheric Ca II lines provide a quasi-continuous reconstruction of the
magnetic field from the deep photosphere to the lower chromosphere. We apply a constant-flux method that follows the same
magnetic flux through different atmospheric layers to quantify
magnetic expansion. All three pores analyzed exhibit an overall increase in
magnetic area with height, approximately doubling their cross section between the lower photosphere and the highest atmospheric layers sampled by the inversions. The expansion remains broadly stable throughout the 48-minute observing sequence. We further characterize the expansion with a quadratic parameterization and investigate its relation to pore properties. The three pores exhibit similar core
magnetic field strengths but different expansion behaviors, indicating that the
magnetic core strength alone does not determine the expansion. However, the present sample is too small to establish whether these differences are related to the thermodynamic or
magnetic properties of the pores and their surroundings. These results demonstrate the potential of simultaneous multiline inversions for investigating the three dimensional
magnetic topology of
solar magnetic structures and provide the first direct measurements of the height-dependent
magnetic expansion of
solar pores.
Read the full abstract ▾
All authors Juan Carlos Trelles Arjona, María Jesús Martínez González, Christoph Kuckein, Sergio Javier González Manrique, Manuel Collados, Sami K. Solanki, Andreas Lagg, Achim Gandorfer, Jose Carlos del Toro Iniesta, Yukio Katsukawa, Pietro Bernasconi, Thomas Berkefeld, Alex Feller, Tino L. Riethmüller, Alberto Álvarez-Herrero, Masahito Kubo, H. N. Smitha, David Orozco Suárez, Bianca Grauf, Michael Carpenter, Alexander Bell, Valentín Martínez Pillet, Francisco Javier Bailén, Julian Blanco Rodríguez, Juan Sebastián Castellanos Durán, Edvarda Harnes, Johannes Hölken, Francisco A. Iglesias, Ryohtaroh T. Ishikawa, Yusuke Kawabata, Takuma Matsumoto, Takayoshi Oba, Azaymi L. Siu-Tapia, Hanna Strecker, Dušan Vukadinović, Hirohisa Hara, Toshifumi Shimizu
Computer Vision
A Surgical Foundation Model Reveals Task-Dependent Label Efficiency
Florian Philipp Stilz, Lorenzo Arboit, Vinkle Srivastav et al. (+6 more)
Developing label-efficient models is a central challenge in
surgical AI due to the high cost and scarcity of expert annotation. While self-supervised foundation models adapt well to new tasks with minimal data, how label efficiency varies across different
surgical tasks remains largely unexplored. Here, we introduce SURGE, a
surgical foundation model trained on SurgSpectrum-30M+, the largest pretraining
dataset comprising over 30 million frames, with checkpoints released to enable further research. We systematically evaluate label efficiency across 5 task categories and 15
benchmarks. These range from temporal and spatial scene understanding to fine-grained
reasoning tied to instrument-anatomy interactions and safety-critical maneuvers. SURGE outperforms prior state-of-the-art on all
benchmarks, even surpassing task-specific models on complex
reasoning tasks. Crucially, we reveal a task-dependent scaling behavior: while scene understanding tasks saturate with minimal supervision, fine-grained
reasoning tasks continue improving with substantially larger annotation budgets, providing a blueprint for allocating expert effort in complex domains. Code: https://github.com/CAMMA-public/SURGE
Read the full abstract ▾
All authors Florian Philipp Stilz, Lorenzo Arboit, Vinkle Srivastav, CAMMA International Surgical Partners, Jacques Marescaux, Sergio Alfieri, Pietro Mascagni, Nassir Navab, Nicolas Padoy
Software Engineering
Hypothesis-awkward: Property-Based Testing Strategies for Awkward Array
Tai Sakuma, Ianna Osborne and Peter Elmer
Hypothesis-awkward is a collection of Hypothesis strategies for Awkward Array. Awkward Array can represent a wide variety of nested, variable-length, mixed-type data. Many tools that process Awkward Arrays are widely used and actively developed. The unit test cases of many of these tools list predefined input samples. In practice, such samples can cover only a small portion of the vast combinatorial space of Awkward Array instances. Hypothesis, a Python property-based testing library, generates test data that can make test cases fail and automatically explores edge cases. The main strategy of Hypothesis-awkward generates nearly all possible Awkward Arrays, with options to control the layout, data types, missing values, masks, and other array attributes. Property-based tests written with these strategies have begun to run in the continuous integration of Awkward Array; to date, they cover a small part of Awkward Array and have found 37 bugs in it, as well as one in Hypothesis itself. These strategies make the tests of Awkward Array and of the tools that use it more reliable.
Read the full abstract ▾
Nuclear Theory
Global equilibrium at order $\mathcal{O}(\hbar^2)$: pseudo-gauge ambiguity, Maxwell relations, and thermodynamic consistency
Asaad Daher, David Wagner, Masoud Shokri et al. (+1 more)
We study massive
spin-$1/2$ Boltzmann particles in global thermodynamic
equilibrium with rotation and acceleration, consistently including
quantum corrections up to order $\mathcal{O}(\hbar^2)$. For such a system, the fluid velocity becomes an independent thermodynamic variable and fundamental thermodynamic relations have to be extended. Employing the Wigner-function formalism, we show that the standard mass-shell condition $k^2 = m^2$ is modified at order $\mathcal{O}(\hbar^2)$ by contributions quadratic in the thermal vorticity. Incorporating this modified mass-shell constraint, we derive the (net) particle-number current, energy-momentum tensor,
spin-current tensor, and associated thermodynamic quantities in the kinetic, canonical, and de Groot–van Leeuwen–van Weert (GLW) pseudo-gauges. We explicitly demonstrate that, while global quantities, like the total particle number or total energy, are independent of the choice of pseudo-gauge, local quantities, like the particle-number density or the energy density, are pseudo-gauge dependent. We then examine the thermodynamic relations in each pseudo-gauge and demonstrate that thermodynamic consistency requires that the thermodynamic Maxwell relations are satisfied. We find that these relations hold in the kinetic-theory pseudo-gauge, whereas they are in general violated in the canonical and GLW pseudo-gauges. These findings suggest a preference for using the kinetic-theory pseudo-gauge in applications such as
spin hydrodynamics.
Read the full abstract ▾
All authors Asaad Daher, David Wagner, Masoud Shokri, Dirk H. Rischke
Physics and Society
Vertices that belong to every minimum dominating set of a graph and their connection with transportation sharing systems with study cases in Campo de Gibraltar area
Mohammad Farhan, Dorota Kuziak, Iztok Peterin et al. (+1 more)
This study addresses a theoretical model regarding equity and accessibility challenges in designing shared transportation systems (such as micro-mobility networks) by applying graph-vertex domination setting. The work focuses on identifying dominating forced vertices, that represent nodes belonging to every dominating set of a graph of the smallest possible cardinality, and which correspond to critical, non-negotiable station locations essential for maintaining system efficiency and coverage. From a theoretical perspective, in the paper it is first demonstrated that determining whether a given vertex is a dominating forced vertex is co-NP-hard, establishing the computational infeasibility of exact identification in large networks. To analyze graph structures, sharp theoretical bounds on the maximum number of dominating forced vertices are established, proving that their count is bounded above by one-third of the order of the graph, and provide complete structural characterizations for graphs achieving this bound, as well as, trees with no dominating forced vertices. To overcome computational limits in practical urban settings, the study uses an iterated greedy metaheuristic framework to generate minimal dominating sets and approximate critical forced nodes based on their appearance frequency across iterations. The methodology is validated on strong grid graphs and applied to real-world road network models of Algeciras and La Línea de la Concepción, two cities in the area of Campo de Gibraltar, Spain, which successfully pinpoints candidate location points for scooter-sharing stations across both cities.
Read the full abstract ▾
All authors Mohammad Farhan, Dorota Kuziak, Iztok Peterin, Ismael G. Yero
Numerical Analysis
Finite-Data Error Bounds for Approximating the Koopman Operator: Sampling Measures, Super-Polynomial Convergence and Regularization
Daniel Fassler, Rachel Morris, Jason Bramburger et al. (+1 more)
The Koopman operator is a well-established framework for lifting nonlinear dynamical systems to an infinite-dimensional space where dynamics are linear. Extended Dynamic Mode Decomposition (EDMD) is a widely used method in data-driven dynamics as it provides a Galerkin approximation of the Koopman operator on the finite-dimensional span of a prescribed dictionary of observable functions. In particular, EDMD only requires samples from the dynamical system, describing the Koopman operator without knowledge of the underlying dynamics. While many studies analyze asymptotic convergence results for EDMD in terms of data, the question of more practical finite-data convergence results remains incomplete in general settings. In this work, we prove bounds for the finite-data EDMD Galerkin approximation of the Koopman operator by a rate proportional to the inverse of the root of the number of samples (the Monte Carlo rate). These results apply to multiple classes of problems, i.e., for discrete or continuous dynamics, including
stochastic systems. In certain more structured setting, we demonstrate super-
polynomial rates are achievable both theoretically and computationally. Finally, we derive recovery guarantees for an EDMD variant in the undersampled regime inspired by compressed sensing techniques.
Read the full abstract ▾
All authors Daniel Fassler, Rachel Morris, Jason Bramburger, Simone Brugiapaglia
Machine Learning
Relational Compression: A Framework for Relational Fidelity in Constrained Representations
Yaniv Shulman
What should a compressed representation preserve when the information of interest lies in relationships among elements rather than in the elements themselves? We formulate relational compression in the classical source-description-reconstruction sense, but with relational structure itself as the fidelity-bearing content. Each instance specifies the source relation, retained description, reconstructed or evaluated relation, fidelity criterion, and constrained resource. We use this interface to situate selected methods from graph summarization, spectral sparsification, similarity-preserving representation, and relational
distillation within a common formulation while keeping their different reconstruction and resource assumptions explicit. We develop finite-codeword collision as one concrete realization. Same-codeword probability yields a relational geometry linking pair-specific
alignment and separation to aggregate Rényi-2 occupancy and the spherical geometry of categorical assignments, with exact objective correspondences to squared-Euclidean centroid reconstruction and normalized graph association and cut. Graph and image studies illustrate complementary routes within the finite-codeword family: graph- and teacher-defined relational requirements act directly on equality or collision, while reconstruction acts through a joint decoder. Together, these results illustrate how distinct relational requirements can be formulated and tested within a common constrained-representation framework.
Read the full abstract ▾
Computers and Society
Anatomy of a Spreadsheet Failure, Analysing the EuSpRIG Horror Story Corpus
Simon Thorne and Angela Collins
Spreadsheet disasters have been documented for more than thirty years, but they have not gone away. This paper examines the EuSpRIG horror stories archive as evidence of how ordinary spreadsheets fail and the consequences that follow. After cleaning and deduplication, 121 incidents from 1995 to 2026 were classified into twelve failure categories and analysed by mechanism, context and consequence. Formula errors, data-entry slips and data-handling mistakes account for 71 cases and recur across the whole period, from Fidelity's missing minus sign in 1995 to a mistyped date that cost Norway's sovereign wealth fund about $92 million in 2024. A second pattern concerns information present in a workbook but not visible to the recipient: hidden rows, hidden sheets, pivot-table caches, embedded objects and retained source layers can cause harm even when no calculation is wrong. The clearest rise in reported cases is in this hidden or embedded data category, whose most serious example is the 2022 Ministry of Defence spreadsheet that exposed details of roughly 18,700 Afghan applicants through hidden rows. The paper argues that spreadsheet risk is no longer only a problem of wrong numbers. It also involves visibility, disclosure, governance and trust. Existing taxonomies remain useful for formula, entry and handling errors, but do not fully capture the public harm caused by spreadsheets used as uncontrolled operational infrastructure. The same mistakes have persisted for three decades, what has changed is the authority organisations give to the spreadsheets that contain them.
Read the full abstract ▾
Artificial Intelligence
DriveHierarchy: A Benchmark for Diagnosing VLM Driving Capabilities from Open-Loop Understanding to Closed-Loop Execution
Chengkai Xu, Jiaqi Liu, Yicheng Guo et al. (+2 more)
Evaluating VLM-based autonomous driving remains difficult because driving competence is composite, where a capable system must ground traffic participants and hazards, integrate context across views and time, reason about future
evolution, and act appropriately under closed-loop interaction. Existing
benchmarks usually assess either open-loop understanding or closed-loop driving but provide limited structure for explaining how these abilities are organized, how they relate, and how they may inform model diagnosis and improvement. We present DriveHierarchy, a hierarchical
benchmark that organizes VLM-based autonomous driving into four ranks, spanning perceptual grounding, contextual
memory, mental
reasoning, and closed-loop execution. To instantiate this hierarchy, we integrate multiple open-source autonomous-driving
datasets into a unified open-loop
benchmark with 76,798 question-answer pairs over 84,279 frames and develop a closed-loop simulation platform with interactive scenario construction on a real-world road network, from which 100 driving scenarios are curated for
embodied evaluation. Experiments on 15 VLMs show that DriveHierarchy captures structured but non-redundant capability variation, relates open-loop understanding to closed-loop driving, and provides a practical basis for diagnosis and
benchmark-guided optimization. DriveHierarchy therefore serves as a unified framework for evaluating and improving VLM-based autonomous driving systems. An anonymized project has been released on https://github.com/PerfectXu88/DriveHierarchy
Read the full abstract ▾
All authors Chengkai Xu, Jiaqi Liu, Yicheng Guo, Peng Hang, Jian Sun
Image and Video Processing
Awaken, Then Scale: Tiny Adaptation for MRI Reconstruction
Mohammed Wattad, Tamir Shor and Alexander M. Bronstein
Dormant Awakening (DA) specializes a
magnetic resonance imaging (MRI) reconstruction model by fitting a small set of zero weights to one labeled slice. Across fourteen U-shaped convolutional network (U-Net) and
vision transformer (ViT) sources, fitting 1,004 or 10,560 weights gives mean peak signal-to-noise ratio (PSNR) gains of .430 and .509 dB. We analyze how scaling the fitted output correction changes evaluation PSNR, then use the calibration correction size to cap changes on new images. The cap bounds changes in root mean squared error (RMSE) relative to the source without an evaluation reference. At 10% adaptation budget, we compare fixed halving and dynamic capping for DA, unrestricted sparse adaptation and low-rank adaptation (LoRA). The controls increase both measured calibration-evaluation correlations across all six tested architecture-adapter settings. Experiments concern one constructed fastMRI-to-M4Raw shift with fixed source checkpoints and participant panels.
Read the full abstract ▾
Image and Video Processing
Beyond Sparsity: Weight Location and Network Context in Pruned MRI Reconstruction
Mohammed Wattad, Tamir Shor and Alexander M. Bronstein
Pruning reduces the number of weights in
magnetic resonance imaging (MRI) reconstruction networks. Equal sparsity, however, can retain weights with different computational roles and different compatibility with the trained network. We study these effects across 120 convolutional U-Net and
vision transformer models using controlled edits evaluated before retraining. In U-Net, preserving high-resolution computation improves reconstruction at equal deletion counts across all tested sparsity levels. Equal-operation controls reveal additional sensitivity of the first convolution, which operation count alone cannot explain. In
transformers, retained pretrained weight energy correlates with quality within sparsity levels, yet the same masks reverse their reconstruction ranking when the surrounding trained weights change. At 90% sparsity, matching intermediate output scales largely removes a high-energy-mask penalty while retaining an advantage for the network's original mask. Thus sparsity alone does not characterize reconstruction quality: architecture-specific descriptors are informative, but mask quality can still depend on the surrounding trained network.
Read the full abstract ▾
Combinatorics
The stacking number of a tree
John Fairfax-Ball
Csernák and Soukup introduced the stacking number of a graph: the least integer $t\ge2$ such that every configuration of $t$ pebbles can be transformed by pebbling moves into a configuration supported on one vertex. They conjectured an explicit formula for the stacking number of a tree in terms of the degrees of its vertices and their distances from a chosen root. We prove this
conjecture for every finite tree with at least two vertices. We also show that whether a configuration on a tree can be stacked at a given vertex is decided exactly by the sign of an integer score built from recursive branch messages; consequently, stackability can be tested with a linear number of arithmetic operations. The proof combines this criterion with an explicit zero-score obstruction and a weighted cancellation argument for arbitrary nonstackable configurations. The main
theorem and the stackability criterion are formalized in Lean 4, and the main
theorem has passed Palomar mechanical verification; it is publicly registered as PALOMAR-2026-09-25-000010.
Read the full abstract ▾
Sound
Video-to-Music Generation for Gameplay Videos
Felipe Marra and Lucas N. Ferreira
Video-to-music models have advanced considerably in the last few years, particularly in film and music
video applications. In this paper, we investigate this problem in the
video game domain, which introduces new challenges for these models:
video frames are rendered graphics, music is mostly synthetic audio, and soundtracks loop across entire levels rather than following on-screen events. We introduce a new
dataset of 217.6 hours of Super Nintendo (SNES) gameplay
video paired with 485 hours of clean soundtracks, free of sound effects and voice-overs, matched to gameplay audio via audio fingerprinting. With this
dataset, we train a simple encoder-decoder
transformer that passes
video features directly to a MusicGen decoder, comparing different encoding strategies: textual descriptions (T5), independent frames (ViT), or spatiotemporal patches (ViViT). Each encoder is tested both frozen and fine-tuned, while the decoder is always fine-tuned. Frozen encoders match or outperform their fine-tuned counterparts on every metric, and the frozen ViViT achieves the best overall results. We compare this model with state-of-the-art baselines using both objective metrics and a listening study (N = 96). Despite having up to 18% fewer parameters, our model outperforms all baselines on objective metrics, surpasses GVMGen in the listening study, and performs comparably to OSSL.
Read the full abstract ▾
History and Philosophy of Physics
Does a Second-Class Primary Constraint Generate a Gauge Transformation? Electromagnetisms and Gravities, Massless and Massive
J. Brian Pitts
In constrained
Hamiltonian dynamics there are two views regarding how first-class constraints generate gauge transformations: individually or only in a certain combination, the Rosenfeld-Anderson-Bergmann-Castellani gauge generator. This gauge generator $G$ preserves Hamilton's equations and changes the canonical action at most by a boundary term;
Hamiltonian's equations are the Euler-Lagrange equations for the canonical action. Hence the canonical formalism is equivalent to the Lagrangian formalism, and indeed subsumed within it (in important examples) with many canonical momenta serving as auxiliary fields. $G$ generates transformations basically equivalent to the usual 4-dimensional Lagrangian expressions, such as a 4-gradient in electromagnetism or a space-time coordinate transformation (on-shell) in General Relativity. It has been shown recently that separate first-class constraints lead to inequivalent observables between Proca non-gauge and Stueckelberg gauge massive electromagnetism. There is, however, widespread agreement that second-class constraints do not generate gauge transformations. Here it is shown that in such a sense as the first-class primary constraint in Maxwell's theory generates a gauge transformation, the second-class primary constraint in Proca's massive electromagnetism also generates a gauge transformation. Likewise the second-class primary constraints in various massive
spin $2$ relatives of General Relativity generate as much of a gauge transformation as do the corresponding first-class primary constraints in GR. Hence the view that first-class constraints typically generate gauge transformations _individually_ faces a puzzle not faced by the gauge generator view.
Read the full abstract ▾
Probability
The principal spectral gap for birth–death processes with a strong Allee effect
Dun Zhou
For birth–death processes with a strong Allee effect, long survival near a positive stable
equilibrium does not by itself determine the principal spectral gap. We identify its large-population limit when the birth and death rates at population size $n$ are $n\widetilde\lambda(n/K)$ and $n\widetilde\mu(n/K)$, where $K$ is the population scale and $\widetilde\lambda,\widetilde\mu$ are the per-capita rates. Write $V(x)=x(\widetilde\lambda(x)-\widetilde\mu(x))$ for the deterministic drift, with unstable threshold $x_1$ and positive stable
equilibrium $x_2$. For the two smallest
eigenvalues $\rho_1(K)<\rho_2(K)$ of the negative generator killed at extinction, we prove \[ \lim_{K\to\infty}\bigl(\rho_2(K)-\rho_1(K)\bigr) =\min\{\widetilde\mu(0)-\widetilde\lambda(0),V'(x_1),-V'(x_2)\}, \] so the limit depends on the linearization rates at all three
equilibria. Each term can be the unique minimum even when $x_1$ and $x_2$ are fixed. The proof combines a frozen boundary model with two local oscillator limits through discrete Ismagilov–Morgan–Simon (IMS) localization. A uniform comparison of the global and stable local ground states controls the orthogonality constraint in the variational lower bound; projected local trial vectors give the matching upper bound. The result determines the limiting $L^2$ spectral gap and optimal Poincaré constant of the $Q$-process, which describes conditioning on indefinite survival. Combined with principal-
eigenvalue asymptotics from a companion work, it separates an exponentially growing quasi-stationary mean extinction time from a spectral relaxation time with a finite positive limit.
Read the full abstract ▾