High Energy Physics (Theory)
The s-Rule from Geometry
Sebastian Franco and Diego Rodriguez-Gomez
String Theory provides powerful tools for constructing and studying five-dimensional
Quantum Field Theories, notably through
geometric engineering and five-brane webs. Understanding how these two descriptions are related in the presence of 7-branes requires incorporating the s-rule into the geometry. To address this question, we focus on basic three-legged five-brane webs ending on 7-branes. Building on the local M-theory description of five-branes ending on 7-branes, we consider triangular toric diagrams with a single non-primitive edge. We show that the corresponding local complex-structure deformation is well defined on the overlaps with the neighboring toric patches, while the corresponding gluing map remains well behaved as a function of the holomorphic coordinates throughout the neighboring patches precisely when the s-rule is satisfied. At saturation, the geometry is a T-cone and the 7-brane deformation becomes its Q-Gorenstein smoothing. Our results thus identify the s-rule with the obstruction to globalizing the local 7-brane deformation and provide a direct
geometric link between T-cones and five-branes ending on 7-branes.
Read the full abstract ▾
High Energy Physics (Theory)
Localizing $\mathcal{N}=(2,0)$ $AdS_3$ solutions of M-theory
Carolina Matté Gregory, Catherine Notman and James Sparks
We apply equivariant localization to supersymmetric $AdS_3\times M_8$ solutions of eleven-dimensional supergravity, dual to two-dimensional $\mathcal N=(2,0)$ SCFTs. We construct equivariantly closed polyforms encoding the
magnetic flux and its dual, calibrated M2-branes and the warped volume, reducing protected observables to fixed point data of the $U(1)$ R-
symmetry action. We apply the framework to $S^4$ bundles over coassociative four-cycles and $S^2$ bundles over compact six-
manifolds, obtaining topological formulae for leading central charges and gravitational anomalies in terms of quantized fluxes and topological data. We recover the known results for $T^2\times KE_4$ solutions and construct a compact numerical solution with quantized flux that lies outside the known
polynomial branch.
Read the full abstract ▾
High Energy Astrophysical Phenomena
Uncovering the Physical Origins of Black Holes in the Milky Way for Astronomical Surveys
Ilia Kiato, Pierson Lipschultz, Vicky Kalogera et al. (+5 more)
The observational landscape of
stellar-mass black holes (BHs) in the Milky Way is expanding beyond X-ray binaries, with
Gaia DR4 and Roman expected to open new windows onto BH populations. We use the binary population synthesis framework POSYDON to predict the present-day
Galactic BH population and connect it to its evolutionary origins. We test the robustness of these predictions to uncertainties in binary
evolution and BH formation, including core-collapse mechanisms, natal kicks, common-envelope efficiency,
metallicity, and initial binary distributions. Across our models, we find that most
Galactic BHs are isolated, with
stellar mergers providing the largest contribution and binary disruptions the second largest. Among surviving binaries, detached BH-BH systems dominate. We identify a distinct, narrow feature at higher BH masses ($\simeq13\,\mathrm{M}_{\odot}$ in the fiducial model) produced by wind-driven
evolution of massive Wolf-Rayet progenitors. This feature is particularly prominent among detached BH-BH binaries and persists across variations in binary-
evolution and BH-formation physics, but depends strongly on
metallicity. At lower masses, these BH-BH systems predominantly originate from progenitors that undergo stable mass transfer before BH formation. Finally, for more direct observational applications, we embed the synthetic populations in a
Galactic model, evolve their kinematics, perform first-order microlensing calculations for BH-BH lenses, and examine BH systems with luminous companions in the context of the
Gaia BHs. In the latter, we uncover a robust stable-mass-transfer population largely absent in rapid population-synthesis predictions. These results provide an evolutionary framework for interpreting the emerging
Galactic BH population across complementary observational channels.
Read the full abstract ▾
All authors Ilia Kiato, Pierson Lipschultz, Vicky Kalogera, Jeff J. Andrews, Max M. Briel, Abhishek Chattaraj, Seth Gossage, Philipp M. Srivastava
Computational Physics
Constitutive-Set Mechanics: variational mechanics on an admissible set of constitutive laws
Altaf Ahmad Lone, R. T. Durai Prabhakaran, Tawqeer Nasir Tak et al. (+2 more)
A structural simulation needs a constitutive law, and experiments rarely determine one uniquely: several laws may fit the same data and satisfy the same physical constraints, yet differ where the structure is loaded in ways the tests never were. Constitutive-Set Mechanics (CSM) keeps all of those laws. It replaces the single law in the variational formulation by the admissible set, and lets the mechanics itself decide which members of the set affect the structural prediction. The framework rests on one observation: a finite element assembly consults the law only at the strains the structure reaches, so the incremental energy depends on the law through a weighted record of those strains, the occupation measure of the state. The energy of a state under the set is a support function on that measure, and the
equilibrium solve one performs anyway supplies the information needed to bound the worst case.
Equilibrium states become certificates on the robust response, rigorous on the upper side for any mechanics solver and two-sided under global minimisation; the certificate has at most one more state than the number of constitutive directions the structure interrogates; a coherence
theorem identifies when the pointwise-worst material is one no single material can be; and constitutive inference acts on the same support function. The method is demonstrated on linear elasticity, a data-constrained constitutive function, phase-field fracture, an assembled finite-strain composite shell, and a membrane characterised by one-mode tests alone. On the shell the structure interrogates four of 226 constitutive directions; on the membrane the certified worst-case energy exceeds the reference more than fourfold, and the experiment that most contracts the certified prediction is not the one where the constitutive uncertainty is largest.
Read the full abstract ▾
All authors Altaf Ahmad Lone, R. T. Durai Prabhakaran, Tawqeer Nasir Tak, Timon Rabczuk, Xiaoying Zhuang
High Energy Physics (Phenomenology)
Monopole Catalysis from Standard Model Anomalies
Nathaniel Craig and Dan Sehayek
We study the generalized
symmetries associated with Standard Model
flavor and their fate in the presence of
magnetic monopoles. This illuminates a striking phenomenological consequence: even when
baryon and
lepton number are classically conserved, the electroweak anomaly induces $\Delta B = \Delta L \in 3 \mathbb{Z}$ in monopole-
fermion scattering for any monopole activating the anomaly. We demonstrate this mechanism in Pati-Salam theory and trinification.
Read the full abstract ▾
Earth and Planetary Astrophysics
Detecting multiple planets in Gaia data using Bayesian model selection
Emmanuel J. Greenfield and James E. Owen
Gaia will soon provide a census of cold
giant planets. However, an astrometric signal contains the superimposed reflex motion of all companions, so planetary signals cannot be treated individually. We therefore investigate
Gaia's
planet-detection capabilities for 1-
planet and 2-
planet systems, with special
attention to systems near 2:1 mean-motion resonance.
Gaia can detect 2-
planet systems, although the lower-S/N
planet requires a higher S/N threshold, especially when the outer
planet is more massive and near 2:1 commensurability. When fitting a 1-
planet model to a circular 2-
planet resonant system, the resulting fit shows biases in eccentricity (up to e = 0.8), mass (up to 100%), and inclination (up to 50 deg). The bias is maximised when the
planets' astrometric signatures are similar, since the combined orbital signal deviates most from a single Keplerian orbit. Many of these systems can be identified as having 2
planets using
Bayesian model selection. Still, in some cases the second
planet is hidden and undetectable due to its S/N, leading to these biases for the outer
planet's orbital parameters. When the outer
planet is roughly over three times as massive as the inner
planet, orbital approximations allow us to recover the astrometric resoeccentric degeneracy from the bias. When the
planets are in dynamical resonance, apsidal precession observed over a long enough baseline (e.g. DR5) can contribute to breaking the degeneracy at short orbital periods. Overall, these biases might significantly affect constraints on
planet formation, migration, and
evolution models from
Gaia population-wide demographics.
Read the full abstract ▾
Solar and Stellar Astrophysics
Unlocking new information in stellar circular polarisation with second-order least-squares deconvolution
Oleg Kochukhov
Least-squares deconvolution (LSD) is the standard technique for measuring
stellar magnetic fields from circular polarisation spectra. It rests on the weak-field approximation and yields the longitudinal field alone. We aim to extend LSD beyond the weak-field limit and interpret departures from a linear field response in terms of the field modulus. The Stokes V absorption coefficient is expanded in odd wavelength derivatives of the intensity profile, with coefficients given by the Zeeman pattern moments. Truncation after two terms gives every line a second weight, formed from the third moment of its splitting pattern, and two mean Stokes V profiles follow from multiprofile LSD. This description was tested with polarised spectrum synthesis and applied to archival spectra of two Ap
stars and two active M
dwarfs. The second mean profile is proportional to the second derivative of the first, with a coefficient measuring the mean square field modulus weighted by the longitudinal flux. This observable separates field strength from the area it covers. Simulations at realistic signal-to-noise ratios show that the second component is reliably detected from about 500 G and constrains the field modulus up to several kG. Applying this methodology to the Ap
stars, we found the modulus consistent with their known field strengths, whereas for the M
dwarfs it was close to the Zeeman-broadening measurements and far above the large-scale fields from Zeeman-Doppler imaging. Second-order LSD turns the breakdown of the weak-field approximation from a source of error into a new
magnetic observable, revealing for the first time the field strength in the regions emitting circular polarisation. It applies to any
star with a detectable Stokes V signature and bridges the gap between polarimetry and Zeeman broadening, most promisingly for cool active
stars in the infrared.
Read the full abstract ▾
High Energy Physics (Theory)
Deriving the $\text{AdS}_3\times\text{S}^3\times \text{T}^4$ Quantum Spectral Curve II: T-system and QSC
Andrea Cavaglià, Davide Polvara, Nicolò Primi et al. (+2 more)
We complete the systematic derivation of the
Quantum Spectral Curve (QSC) for strings on $\text{AdS}_3\times\text{S}^3\times \text{T}^4$ with pure Ramond–Ramond (RR) flux, building on the extended Y-system obtained in our preceding paper. We establish the precise relation between the two formalisms of the Thermodynamic Bethe Ansatz and the QSC, derive the previously conjectured QSC, and, along the way, uncover additional relations satisfied by its solutions. Our analysis also clarifies how massless excitations, although not explicit among the basic QSC building blocks, are encoded in their analytic structure. The derivation culminates in the so-called $\mathbf{P}{\mu}$-system, a compact formulation of the QSC in terms of a small set of functions and their analytic continuations.
Read the full abstract ▾
All authors Andrea Cavaglià, Davide Polvara, Nicolò Primi, Alessandro Sfondrini, Roberto Tateo
Strongly Correlated Electrons
Heavy-fermion hybridization waves from Dirac spinons near a Kondo-breakdown quantum phase transition
Pedro M. Cônsoli, Matthias Vojta and Onur Erten
Motivated by the recent observation of spatially modulated screening in different paramagnetic heavy-
fermion materials, we propose a theoretical mechanism for the emergence of such hybridization waves in frustrated Kondo
lattices with no
magnetic order. By applying both a parton mean-field theory and a perturbative calculation to a minimal Kondo-Heisenberg
lattice model, we demonstrate that a heavy
Fermi liquid can naturally develop hybridization waves when it approaches a
quantum phase transition where Kondo screening breaks down and the local
magnetic moments form a Dirac
spin liquid. We relate the strength of this mechanism to a kinematic condition on the electronic and spinon band structures and show that the spatial periodicity of the ensuing hybridization wave is controlled by momenta $\vec{Q}$ that connect the Dirac nodes in the spinon dispersion to low-energy portions of the electronic band structure.
Read the full abstract ▾
High Energy Physics (Theory)
The Back-to-Back Plateau of the Energy-Energy Correlator from Weak to Strong Coupling
Hao Chen, Ian Moult, Bo Wang et al. (+1 more)
The back-to-back limit of the energy-energy correlator (EEC) exhibits a finite plateau, capturing the
evolution from perturbative soft-collinear radiation at weak coupling to string dynamics at strong coupling. At weak coupling, we resum its leading and next-to-leading-power logarithms through N$^3$LL in $\mathcal N=4$ SYM, including non-planar contributions. Combining this weak-coupling endpoint with bootstrap and strong-coupling data, we construct an interpolation for the plateau height from weak to strong coupling. The calculation uses a human-in-the-loop agentic workflow.
Read the full abstract ▾
All authors Hao Chen, Ian Moult, Bo Wang, Hua Xing Zhu
Quantum Physics
Decoupling Spectral Gaps from Entanglement Complexity in Quantum Annealing
Roopayan Ghosh and Sougato Bose
Exponentially small spectral gaps are the central obstacle to adiabatic
quantum annealing, yet whether such bottlenecks demand correspondingly complex
entanglement remains unclear. We show that gap scaling and
entanglement complexity need not be linked. Conventional first-order transitions combine exponentially small gaps with bounded $O(1)$ bipartite Schmidt rank, even though the critical states retain global multipartite correlations, while continuous transitions combine polynomially closing gaps with logarithmically growing
entanglement. We show that each pairing can also be inverted: in a known first-order transition in a staggered-field
antiferromagnetic Ising chain the gap closes polynomially despite bounded Schmidt rank, while we construct a mirror-twin
manifold model whose gap closes exponentially even as its
entanglement entropy grows logarithmically with system size. We derive bounds connecting Schmidt rank to Hamming-space geometry, and identify
manifold density and connectivity as the ingredients that determine which combination occurs. These results show that
entanglement complexity alone cannot characterize annealing difficulty, and point toward a broader search for
quantum resources that capture the geometry of state
evolution in
Hilbert space.
Read the full abstract ▾
Quantum Physics
Kardar-Parisi-Zhang superdiffusion in chaotic quantum circuits
Rustem Sharipov, Urban Duh, Sun Woo P. Kim et al. (+1 more)
We construct a family of chaotic brickwork
quantum circuits exhibiting superdiffusive transport with dynamical exponent $z=3/2$ in the full Kardar–Parisi–Zhang (KPZ) universality class. This is enabled by the
equilibrium current associated with a inhomogeneous conserved density having a nontrivial dependence on the chemical potential. Using nonlinear fluctuating hydrodynamics, we derive the propagation velocity, broadening scale, and universal scaling form of the charge correlation function from microscopic
equilibrium properties. Tensor-network simulations of a two-site
spin-$1$ (qutrit)
circuit reproduce these parameter-free predictions, including the full stationary KPZ profile. We further construct a three-site
spin-$\frac12$ (
qubit)
circuit, where transport varies from diffusive to superdiffusive with the chemical potential, while long-lived coherent quasiparticles delay the convergence of the profile to the KPZ scaling form. Our results establish that KPZ superdiffusion in
quantum systems is not restricted to fine-tuned integrable models.
Read the full abstract ▾
All authors Rustem Sharipov, Urban Duh, Sun Woo P. Kim, Friedrich Hübner
General Relativity and Quantum Cosmology
The quasinormal modes of Kerr-Newman black holes: a separated approach
Kyriakos Destounis, Vitor Cardoso and Peter Hintz
We compute Kerr-Newman quasinormal modes using a recently found exact separation of the gravitoelectric perturbation equations. For each
spin-label $\varsigma=\pm1$, a
matrix angular ordinary differential equation and a scalar second-order radial ordinary differential equation define a joint spectral problem for the complex frequency $\omega$ and separation constant $\lambda$. We solve this problem by imposing regularity at the angular poles, ingoing behavior at the event horizon, and outgoing behavior at infinity. We
benchmark the numerical code built to solve the separated equations, and the resulting frequencies, against recently published spectral calculations of the coupled system. Comparisons with those results yield absolute differences of order $10^{-10}$ to $10^{-5}$, depending on the BH parameters and overtone number. We also find numerical hints of
spin-label isospectrality in both the electromagnetic-led and gravitational-led sectors of fundamental quasinormal modes, even at near-extremal Kerr-Newman configurations. These results provide numerical evidence that the separated equations reproduce Kerr-Newman quasinormal frequencies and offer a practical route to exploring the ringdown of rotating, charged black holes.
Read the full abstract ▾
High Energy Physics (Theory)
Checking the Schwinger-Dyson equation in the purely open matrix-tensor model
Daniel L. Jafferis, Liza Rozenberg, Diandian Wang et al. (+1 more)
We test the proposed
matrix-tensor description of three-dimensional
gravity using a Schwinger-Dyson relation in its purely open sector. We enumerate and classify the topological contributions to a cylinder amplitude. Under a proposed counting prescription motivated by Virasoro TQFT, we find agreement with the conjectured unit difference between the numbers of contributing
manifolds associated to the two interaction terms in the potential. By contrast, counting all inequivalent surface
embeddings separately would instead give a mismatch. This comparison highlights the importance of specifying the set of bulk
topologies to be included in the gravitational path integral.
Read the full abstract ▾
All authors Daniel L. Jafferis, Liza Rozenberg, Diandian Wang, Cynthia Yan
Cosmology and Nongalactic Astrophysics
And Then There WEre Baryons (ATWEB): The Baryon Budget of the Universe in Halos and the IGM over 13 Billion Years of Cosmic Time
Mohammadreza Ayromlou, G. Mark Voit, Thomas H. Reiprich et al. (+2 more)
Observations of the local Universe show that, except in the most massive
galaxy clusters, the
baryon fraction within
halos falls below the
cosmic mean. This deficit can be characterized using the previously established closure radius, the halocentric distance within which the enclosed
baryon fraction reaches the
cosmic value. Using the TNG, EAGLE and SIMBA simulations, we examine how, when, and on what scales the baryonic and dark matter distributions diverge. At z=0 less than 30% of
cosmic baryons reside within
galaxy-hosting
halos, compared with ~45% of the dark matter, leaving over 70% of all
baryons in the IGM. Building on the closure radius, we introduce three further characteristic scales: the compensation radius, closure envelope, and compensation envelope. The compensation radius marks where the enclosed total mass recovers its
gravity-only value, and today it typically lies inside the closure radius in all three models. The closure and compensation envelopes are the largest median radii across
halo mass.
Galaxy groups provide the largest contribution per dex of
halo mass to IGM enrichment with
halo-associated
baryons and redistribute
baryons over the largest distances. They set the closure envelope, which reaches ~2-6.5 Mpc at z=0, depending on the model. The closure radius usually increases with
cosmic time at fixed
halo mass. Tracking individual
halos, we find open
baryon cycles with substantial redistribution beyond
halo boundaries across all
halo masses, while even the progenitors of today's most massive clusters fall below the
cosmic baryon fraction. The compensation envelope grows toward low
redshift, reaching ~1.5-5.5 Mpc at z=0. Across
cosmic time, it coincides with the scale on which the matter power spectrum recovers its
gravity-only form, establishing a direct link between the spatial extent of
baryon redistribution and the impact of feedback on large-scale structure.
Read the full abstract ▾
All authors Mohammadreza Ayromlou, G. Mark Voit, Thomas H. Reiprich, Cristiano Porciani, Dominique Eckert
Quantum Physics
Protecting bosonic codes from ancilla-induced errors with continuous-variable flags
Thomas Décultot, Ronan Gautier and Mazyar Mirrahimi
Bosonic
quantum error correction represents a route towards significant hardware overhead reduction for fault-tolerant
quantum computation. However, controlling
quantum harmonic oscillators that encode the
quantum information comes with its own challenges. In many platforms such as superconducting
circuits and trapped ions, the bosonic degree of freedom is controlled by an ancillary
qubit whose decay during a
qubit-oscillator gate propagates to the oscillator as a continuous error. Such errors generally lie beyond the scope of errors that the bosonic code is capable of correcting. We introduce the continuous-variable flag: an auxiliary oscillator coupled to the ancilla whose phase-space position encodes the continuous error corrupting the storage mode. After the gate, a heterodyne measurement of the flag estimates this continuous-variable error and a subsequent feedback control undoes it up to a bosonic code stabilizer. The scheme requires no new ingredients, as the flag is driven by the same dispersive coupling, displacements and rotations that implement the gates it is supposed to protect. We construct sequences protecting conditional rotations to all orders in ancilla decay with one flag, and conditional displacements to first order with one flag and to all orders with two. Simulations with realistic
circuit-QED parameters predict more than three orders of magnitude fewer logical errors for the parity measurement of a four-component cat code, and a finite-energy GKP logical lifetime near its noiseless-ancilla bound. By discretizing continuous-variable errors, the flags introduced in this work are a useful tool for any hybrid oscillator-
qubit processor, and open the route towards fault-tolerant simulation and computation with bosonic modes.
Read the full abstract ▾
Astrophysics of Galaxies
Cracking open the "dragon egg": ALMA 1.3mm observations of a compact, young high-mass core candidate at 30mas resolution
A. T. Barnes, J. D. Henshaw, Á. Sánchez-Monge et al. (+14 more)
Massive prestellar cores are a central assumption of core-
accretion models of high-mass
star formation, yet convincing observational examples remain rare. The "dragon egg" (C2c1a) core in the infrared-dark cloud G028.37+00.07 has been proposed as a promising compact high-mass core candidate. We present new ALMA Band 6 observations at $0.035''$-$0.078''$ ($\sim175$-400 au at 5 kpc), providing one of the most stringent high-resolution tests of such a candidate. The region resolves into multiple fragments, but the brightest component, C2c1a(a), remains a single compact continuum peak down to $\sim175$ au, with a deconvolved radius of $\sim87$ au and a nearest neighbouring fragment at a projected separation of $\sim2400$ au. The continuum peak reaches a brightness temperature of $\sim40$ K, exceeding the $\sim10$ K gas temperatures inferred from lower-resolution NH$_3$ observations and indicating a strong radial temperature gradient. Continuum-based mass estimates range from $\sim2$ to $\sim30$ $M_\odot$ (implying extreme densities of $n_\mathrm{H2} > 10^9$cm$^{-3}$), depending on temperature, while lower-resolution data indicate a much larger mass reservoir of $\sim110$ $M_\odot$ on $\sim10^4$ au scales. The spectrum shows no strong hot-core tracers and only weak, spatially extended H$_2$CO and CO emission; no high-velocity
outflow can be unambiguously associated with C2c1a(a). We conclude that C2c1a(a) is an exceptionally compact core embedded within a massive cold reservoir, with no additional compact fragments detected within the present resolution and sensitivity limits, potentially tracing the earliest formation of a central protostellar "seed" within a still-dominant reservoir. [abridged]
Read the full abstract ▾
All authors A. T. Barnes, J. D. Henshaw, Á. Sánchez-Monge, P. Caselli, I. Jiménez-Serra, J. E. Pineda, K. Immer, M. De Simone, R. J. Parker, F. Fontani, E. Redaelli, S. N. Longmore, L. Böhm, Q. Zhang, C. Y. Law, G. Cosentino, J. C. Tan
Instrumentation and Methods for Astrophysics
Optimal Photon Counting with Fast, Low Noise Astronomical Imagers
Jack T. Dinsmore, Christopher Layden, Roger W. Romani et al. (+9 more)
Measuring rapid astrophysical phenomena requires many short exposures, which suffer from repeated read noise. A variety of recent technologies are achieving sub-electron read noise, which enable rapid time-domain studies of faint sources. These data can be analyzed with traditional methods, but the ability to resolve individual electrons enables a new class of more precise,
photon-counting techniques. We present a
photon-counting method for sub-electron read noise detectors which delivers optimal statistical uncertainties. We focus on Complementary Metal-
Oxide-
Semiconductor (CMOS) detectors and demonstrate our method on optical data collected by CMOS instrument "proto-Lightspeed," recently commissioned at the Magellan Clay
telescope at Las Campanas Observatory. The method can be applied to non-CMOS detectors as well, such as skipper CCDs and linear-mode avalanche photodiodes. The method substantially improves analysis of faint sources, increasing signal to noise ratios by up to $\sim 130\%$, which would require 5.3$\times$ more observing time to accomplish with data alone. As an example we show detections of four normal rotation-powered
pulsars with confirmed optical pulses, achieved in under two hours of total exposure. We provide the lightspeedpy pipeline for reducing proto-Lightspeed data with both
photon counting and standard techniques.
Read the full abstract ▾
All authors Jack T. Dinsmore, Christopher Layden, Roger W. Romani, Vera L. Berger, Kevin B. Burdge, Geoffrey Mo, Deepto Chakrabarty, Emma T. Chickles, Gabor Furesz, Juliana García-Mejía, Dave Osip, Jack Piotrowski
High Energy Physics (Phenomenology)
Minimal cascade decay of the top quark through a new scalar
Bogdan A. Dobrescu, Max H. Fieg and Alessandro Russo
We explore a possible decay of the top
quark into a new scalar $\phi_t$ plus a charm
quark. The decay is mediated by a dimension-5 operator that may be generated by the exchange of a heavy vectorlike
quark, $\chi$, and can have a branching fraction close to a percent. In a minimal model, $\phi_t$ decays due to the same operator that allows the nonstandard top decay, leading to a final state involving a $c$ jet, a $b$ jet and a $W$
boson. Hence, the top
quark would undergo a nonstandard cascade decay through $\phi_t$ and then through an off-shell top
quark. The main nonstandard
collider signal is a $t \bar t \to c \, \bar c \, b \, \bar b \, W^+ W^-$ process. In addition, production of the vectorlike
quark at the LHC would lead to a signal of the type $\chi \phi_t \, j \to 3c + b \, \bar b \, W W j$.
Read the full abstract ▾
High Energy Physics (Theory)
A New Entropy Puzzle in ABJM
Alexandre Belin, Seyed Morteza Hosseini, Palash Singh et al. (+2 more)
We point out a puzzle about BPS black holes in AdS$_4\times S^7/\mathbb{Z}_k$: the
entropy grows when $k$ increases although the number of gauge-invariant operators in the dual conformal field theory that can contribute to the
entropy grows as $k$ decreases, due to the presence of monopole operators. Analyzing the superconformal index of ABJM, we show that the same puzzle exists for finite $N$ in the large charge limit. We propose a resolution based on an increasing number of trace relations as $k$ decreases, which dominates over the other effect. We argue that this mostly affects fortuitous operators and we provide an explicit example of a fortuitous operator that disappears at small $k$.
Read the full abstract ▾
All authors Alexandre Belin, Seyed Morteza Hosseini, Palash Singh, Rita Vadala, Alberto Zaffaroni
High Energy Physics (Theory)
All-Orders Classical Soft Graviton Theorems in the Newtonian Limit and Beyond
Debanjan Karan, Babli Khatun, Raghu Mahajan et al. (+2 more)
The classical soft graviton
theorem aims to determine certain terms in the early- and late-time tails of the gravitational waveform produced in a scattering process, using only the asymptotic momenta of the scattered objects. In this paper we find an infinite series of leading-logarithmic tail terms in the Newtonian limit, where the incoming and outgoing particles move slowly, but the scattering itself can be violent. The all-orders
conjecture of Alessio, Di Vecchia and Heissenberg agrees with our results for two-particle scattering, but does not hold in general when the initial or final state contains three or more particles. We also propose an all-orders
conjecture for these tails beyond the Newtonian limit, but do not prove it. Nevertheless, the
conjecture passes several nontrivial consistency checks.
Read the full abstract ▾
All authors Debanjan Karan, Babli Khatun, Raghu Mahajan, Biswajit Sahoo, Ashoke Sen
Earth and Planetary Astrophysics
First Direct Identification of a Multi-Star Microlens System Hosting a Planet
Sean K. Terry, David P. Bennett, Jean-Philippe Beaulieu et al. (+11 more)
Nearly 25 years have passed since gravitational microlensing delivered its first
exoplanet discoveries. The relative proper motion between lens and source systems allows them to be resolved as they separate on the sky over many years. This enables scientists to study light coming directly from the lenses and place tight constraints on the physical properties of these planetary systems. In this paper we present an analysis of high angular resolution images of the microlensing target OGLE-2006-BLG-284 using Keck adaptive optics and HST data taken more than 15 years after the event. We clearly detect the lens system and measure its motion across the Keck and HST epochs. We measure a lens-source relative proper motion of $5.80 \pm 0.15$ mas yr$^{-1}$. A combination of the light curve parameters and the host
star(s) brightnesses in $V,\, I$, and $K$ passbands gives primary and secondary host
star masses of $M_{\rm 1} = 0.43 \pm 0.05\,M_{\odot}$ and $M_{\rm 2} = 0.12 \pm 0.02\,M_{\odot}$, and a planetary companion mass of $m_p = 165^{+40}_{-33}\,M_{\oplus}$. Although we are unable to determine if the
planet orbits one or both
stars in the system, our analysis motivates next-generation high-precision RV measurements that can distinguish these two possibilities. This is the first direct detection of flux coming from a two-
star microlensing system which hosts a
planet.
Read the full abstract ▾
All authors Sean K. Terry, David P. Bennett, Jean-Philippe Beaulieu, T. Dex Bhadra, Aparna Bhattacharya, Joshua W. Blackman, Ian A. Bond, Andrew A. Cole, Przemek Mróz, Kansuke Nunota, Clément Ranc, Natalia E. Rektsini, Daisuke Suzuki, Aikaterini Vandorou
Computation and Language
CroissantMiner: Automated Extraction and Validation of Croissant Metadata for ML Datasets
Berke Arda, Ahmetcan Yavuz, Paul Gerry et al. (+7 more)
Croissant has emerged as a standard for machine-readable
dataset metadata, yet populating its fields remains labor-intensive and requires careful reading of accompanying
dataset documentation. We present the first
benchmark enabling end-to-end evaluation of metadata extraction aligned with a community-standard schema. The
benchmark comprises 602 papers, including 102 with human-validated gold annotations and 500 with LLM-generated silver annotations, covering the full Croissant schema with both core and Responsible AI (RAI) fields. Using this
benchmark, we evaluate a range of extraction systems spanning frontier models, open-weight models, and agentic architectures, under a two-tier evaluation framework that combines rule-based scoring with an LLM judge selected via human audit. We find that single-pass extraction consistently outperforms the four agentic architectures we evaluate: across backbones, these decomposed variants achieve lower accuracy than a single full-context pass. The largest gap appears on long-form RAI fields, which require synthesizing and interpreting information scattered across a paper rather than copying it from a single location, a setting where current systems remain far from reliable. We release the
benchmark, evaluation code, judge audit, a live demo, and a leaderboard open to new systems.
Read the full abstract ▾
All authors Berke Arda, Ahmetcan Yavuz, Paul Gerry, Sebastian Lobentanzer, Nobin Sarwar, Joan Giner-Miguelez, Kongtao Chen, Luyao Zhang, Mrinmaya Sachan, Mubashara Akhtar
Machine Learning
The Implicit Bias of Hyperbolic Representation Learning for Multiclass Data: A Busemann Risk Perspective
Xingrun Li, Sho Kuno, Yusuke Mukuta et al. (+2 more)
We study the implicit bias of Riemannian gradient flow for
hyperbolic multiclass classification with fixed class prototypes in
hyperbolic space $\mathbb H^n$. Our framework accommodates general permutation invariant relative margin (PERM) losses, a class that includes cross
entropy and other standard multiclass losses. Our analysis is based on a decomposition: at large radius, the distance to each prototype splits into a radial term and a direction dependent term described by the Busemann function. This yields two main results. First, we prove a radial dichotomy: the sign of a drift coefficient $\mu$ determines whether the radius is pushed toward the ideal boundary or back toward the interior; if the positive drift persists, then $r(t)=\tfrac12\log t+O(1)$, while persistent negative drift returns the trajectory to the large radius threshold in finite time. Second, we show that large-radius directions can approach critical points of the Busemann risk on $\partial\mathbb H^n$, even during radial return. These results provide a rigorous large-radius perspective on two phenomena we refer to as boundary saturation and near-boundary clustering in
hyperbolic representation learning.
Read the full abstract ▾
All authors Xingrun Li, Sho Kuno, Yusuke Mukuta, Xin Yang, Tatsuya Harada
Artificial Intelligence
Is this machine playing?
Nathan Cloos, Antonio Norelli, Daniel Durbin et al. (+3 more)
We placed a modern AI coding assistant in an unintended role: as the mind of a body on an unknown digital island. With only a minimal instruction mentioning no specific task, reward, or activity, the machine started animating its virtual body. Across thirty-hour runs, the
embodied AI
agent climbed hills, stacked blocks into towers, drew mandalas, reinterpreted sports, ran experiments on the physics of its world, and learned techniques that later expanded what it could accomplish. These activities recurred across thirteen
agents but diverged into distinct histories. We examine whether this behavior satisfies classical criteria for play and ask whether play can become a mode of machine development.
Read the full abstract ▾
All authors Nathan Cloos, Antonio Norelli, Daniel Durbin, Jacob Andreas, Daniela Rus, Phillip Isola
Physics and Society
The search process of professional detection dog teams in realistic environments
Stephan P. Kaufhold, Jack Terwilliger, Samantha Kefer et al. (+1 more)
Scent detection work is a form of interspecies cooperation. Dogs sample odors that handlers cannot perceive, and handlers decide where the team searches and which of the dogs' responses counts as a find. However, detection performance is predominantly measured by outcomes and dog characteristics. To better observe the search process itself, we collected data at two judged detection events (109 searches of 37 professional narcotics teams). All searches were
video-recorded and included an outdoor area, a vehicle, and a building search, each with two targets and food and currency distractors. We synchronized computer-
vision tracking of dogs and handlers with continuous behavioral coding. We observed variation across teams and in total 42% of targets were found. Finds were rarely missed because dogs never came close to a target or because handlers did not call an indication. Rather, most were missed at the encounter itself, as dogs came within 1~m of nearly every target but indicated on only about one in six of these passes. Handlers did not behave differently near targets than near distractors before their dogs responded. Finding targets was moderately consistent within teams, whereas false calls depended on the individual search and on specific items. Observing the search process suggests that success is a result of the dog's response, the handler's decision, and their coordination.
Read the full abstract ▾
All authors Stephan P. Kaufhold, Jack Terwilliger, Samantha Kefer, Federico Rossano
Social and Information Networks
Aggregating User Preferences while Ensuring Equity, Diversity, and Inclusion using Graph Summarization
Adji Marieme Sita Cissé and Malek Mouhoub
Aggregating the preferences of diverse user groups into a collective outcome raises fundamental challenges of equity, diversity, and inclusion (EDI): classical aggregation rules such as Borda and Condorcet have no mechanism to prevent results from systematically favoring majority groups, collapsing onto homogeneous items, or under-representing minorities. We address this problem through EDI-constrained graph summarization. User preferences are modeled as a weighted attributed bipartite graph, and a greedy coarsening
algorithm iteratively merges user nodes while enforcing three structural EDI criteria: an equity gap constraint ($\Delta E$), an intra-list diversity constraint (ILD), and a group inclusion constraint. Rather than correcting
fairness after aggregation, our method embeds EDI preservation directly into the graph structure. We evaluate across five
datasets spanning four domains: MovieLens 100k and 1M, libimseti.cz, Rate My Professors, and OpenAlex (2018-2023). Our method, AURORA, achieves the largest and most consistent diversity gains over classical voting rules, and on MovieLens 100k at $k = 20$ it simultaneously improves all three EDI criteria over both Borda and Condorcet. On Rate My Professors, it combines high diversity (ILD = 0.808) with the highest female item representation (60%), at a moderate equity cost, and it achieves the lowest equity gap ($\Delta E = 0.031$) on OpenAlex, where Borda-based methods recommend zero female authors. On libimseti.cz, the only
dataset where the sensitive attribute is present on both sides of the bipartite graph, our method does not reduce the equity gap, a limitation we connect to prior findings that demographic parity is not always an appropriate target. These results demonstrate that
embedding EDI constraints into aggregation structure yields more robust
fairness-diversity trade-offs than post-hoc approaches.
Read the full abstract ▾
Computer Vision
PlaySuite: A Large-Scale Benchmark for Interactive Visual Intelligence
Dheeraj Varghese, Anna Vettoruzzo, Walter Simoncini et al. (+5 more)
Recent advances in
multimodal foundation models yield strong performance on static perception and
reasoning benchmarks, yet such evaluations largely overlook a central aspect of intelligence: acting competently in dynamic environments over extended time horizons. We introduce PlaySuite, a large-scale
benchmark for evaluating interactive visual intelligence across more than 5K open-source
video games curated from PyWeek and itch.io. Spanning diverse genres and engines, including Pygame, HTML5, Godot, and Unity, these independent games are largely out-of-distribution for current models, reducing the likelihood that success can be achieved by retrieving memorized walkthroughs or web-scale training artifacts. To enable scalable evaluation across heterogeneous titles, we develop a unified closed-loop interaction framework optimized for HPC clusters alongside a
Video-LLM-as-a-judge protocol that maps observable gameplay milestones to standardized progress levels. We evaluate fourteen recent open models spanning
vision-
language models, computer-use
agents, and
vision-
language-action models. Our results yield strong evidence of a perception-action gap: despite strong
reasoning capabilities, current models struggle to make sustained progress and exhibit recurring failures in spatial grounding, action execution, and self-correction. PlaySuite provides a reproducible and extensible testbed for measuring progress from visual perception to goal-directed interaction, and a foundation for developing models that can act, adapt, and generalize in dynamic visual environments.
Read the full abstract ▾
All authors Dheeraj Varghese, Anna Vettoruzzo, Walter Simoncini, Michelle Lorena Acevedo Callejas, Mohammad Mahdi Derakhshani, Kristof Meding, Joaquin Vanschoren, Cees G. M. Snoek
Number Theory
CUDA-MPQS: A GPU-Resident Self-Initializing Quadratic Sieve, and the Factorization of RSA-155
Fabian Januszewski and Christoph Heinrichs
The quadratic sieve (QS) is an irregular, cache-hostile, branch-heavy integer factorization
algorithm; prior GPU work accelerates individual stages of it. We present \swname{}, a self-initializing quadratic sieve in which every stage —
polynomial initialization, sieving, relation post-processing, large-
prime matching, \GFtwo{}
matrix construction, Block Wiedemann linear
algebra and the square root — executes on the GPU, the CPU restricted to orchestration, setup and I/O, the sieve window 99.9% GPU-busy with no device-wide or event synchronization and no synchronous copy. With it we factored the 512-bit RSA-155 challenge modulus end to end on GPUs: a 64$\times$H100 sieve feeding a single-H100 \GFtwo{} solve, 700.6~GPU-h and 242~kWh in total, 98.4% of it sieving. To our knowledge this is the largest integer factored by the quadratic sieve, exceeding the prior QS record RSA-150 (Buhrow, YAFU, 2025); all comparable general factorizations since 1996, RSA-155's own in 1999 included, used the number field sieve. On a single device we factor RSA-100 in 29.2~s on an H100 and 51.0~s on a consumer RTX~5070~Ti; in a controlled same-modulus measurement, with energy measured on both sides, the H100 is faster than the fastest CPU quadratic sieve, YAFU, on 96 cores of an EPYC~9655 (106–122~s), and than CADO-NFS's SIQS (292~s) and GNFS (263~s). Across four discrete GPUs (a 3.7$\times$ range in CUDA cores) RSA-100 throughput is close to proportional to core count,
memory bandwidth being a weaker predictor, and measured cost over seven
benchmark factorizations from RSA-100 to RSA-155 grows as $\LN{1/2}{1}^{1.04}$ ($R^2=0.993$). On RSA-150, the identical modulus on which the prior QS record was set, we measure 302.9~GPU-h against that run's 11{,}664 core-hours.
Read the full abstract ▾
Cryptography and Security
Jailbreaking Open-Weight LLMs via Random Embedding Perturbations
Abhinav Sudhakar Dubey, Scott Sirri, Vaggos Chatziafratis et al. (+1 more)
While open-weight models have enjoyed steady progress in capabilities and wide adoption across multiple domains, their safety remains an important concern. One key feature is the ability to refuse or deflect harmful, malicious, or insensitive
prompts. In this paper, we expose safety vulnerabilities across six common open-weight LLMs of various sizes that consistently lead to harmful or unsafe responses on the JailbreakBench
benchmark dataset. Our proposed attack, Perturbed
Embedding Vector (PEV), is a simple and fast "jailbreaking" technique that is cheaper than prior approaches, which typically require gradient computations, per-
prompt optimizations, or altering internal weights of the models. PEV just adds independent Gaussian noise in the
embedding vector representations of the
prompt, with no need for further manipulations. To generate unsafe responses, we repeatedly sample additive noise from this distribution. In our experiments, we observe that the average compute cost to get the first successful attack is up to an order of magnitude less than previous attacks. The first successful jailbreak on a new
prompt typically arrives within one minute on every tested model, and PEV generates unsafe responses across all models for all
prompts in JailbreakBench. No other tested method achieves such results, despite them taking longer to run. More broadly, we believe that understanding the behavior of LLMs under perturbations in the
embedding vectors is an important research direction: while perturbations constitute a major security risk, they can also serve as a valuable tool for exploring the dynamical behavior of such models.
Read the full abstract ▾
All authors Abhinav Sudhakar Dubey, Scott Sirri, Vaggos Chatziafratis, C. Seshadhri