Peer Revenue Returns subject to peer review

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

All of arXiv

Tuesday 1 September

Cryptography and Security

HEAT: Faster Fully Homomorphic Inference via Approximations-Weights Co-Adaptation

Fully homomorphic encryption (FHE) allows a server to run a language model directly on encrypted user prompts, but current approaches remain prohibitively slow. Ciphertexts natively support only addition, multiplication, and rotation, and multiplications may be composed only to a bounded depth before a costly bootstrapping operation is required to continue. Every nonlinearity must therefore be approximated by an iterative method; each iteration increasing the number of multiplications. A higher iteration count buys precision but exhausts the available depth more frequently and thus triggers more bootstraps, which dominate latency. We introduce Homomorphic Encryption-Aware Training (HEAT), a fine-tuning method that makes the per-nonlinearity iteration counts learnable, enabling them and the model weights to co-adapt during training. HEAT optimizes iterations with respect to the task objective, allowing the model to adapt to approximation errors encountered during inference without architectural changes or retraining from scratch. We further relate iteration count to quantization bit width and bound, at fixed weights, the gap between our objective and quantization-aware training. On encrypted GPT-2 decoding, HEAT reduces iterations by $3.1\times$, bootstraps by $1.6\times$, and end-to-end latency by $1.4\times$, while improving decode agreement over the calibrated encrypted baseline.
Read the full abstract

All authors Alessandro Zirilli, Davide Marincione, Evgenios M. Kornaropoulos, Giuseppe Ateniese, Emanuele Rodolà

Machine Learning

RecKAN: Kolmogorov-Arnold Networks with a Learnable Recursive Polynomial Basis

Kolmogorov–Arnold Networks (KANs) replace the fixed scalar weights of a standard network with learnable univariate functions on each edge, but existing variants still fix the basis that those functions are built from: B-splines, Chebyshev polynomials, wavelets, or Jacobi polynomials, and learn only the combination weights over it. We introduce RecKAN, which instead defines the basis itself by a second order polynomial recurrence, $R_{n+1}(x) = (ax^2+bx+c)R_n(x) + (dx+e)R_{n-1}(x)$, whose five coefficients are learned jointly with the network. We show this recurrence recovers several classical polynomial families including both kinds of Chebyshev polynomials, Fibonacci, Pell, and Jacobsthal polynomials as special cases, and prove that its degree grows linearly in $n$ exactly on the sub-family containing all of them, giving a concrete sense in which the learned basis can move beyond any fixed classical choice. Across multiple benchmark datasets spanning image, text, biomedical time series classification, and time series forecasting, RecKAN outperforms three parameter-matched KAN baselines (Chebyshev, Jacobi, and spline based) on all classification tasks and achieves the lowest MSE on the ETTh1 forecasting benchmark. Additionally, when used as a classifier head with a convolutional backbone, RecKAN achieves higher accuracy than standard MLP heads on Fashion MNIST, CIFAR-10, and SVHN. On a synthetic function fitting benchmark it tracks a sharply oscillatory target that a parameter comparable MLP under fits. We further show that the learned recurrence coefficients are interpretable: on the task requiring the most local structure, training moves the basis away from the linear degree growth regime that contains every classical family we identify, consistent with our theoretical analysis of what that structural shift enables.
Read the full abstract

Materials Science

A Spatial Localizer for Constituent-Resolved Exciton Wannier Functions

Excitons are composite quasiparticles: beyond a center-of-mass position, each carries an internal electron-hole dipole governing its coupling to electric fields and other excitons. Exciton Wannier functions locally represent exciton bands, but resolving this dipole requires localizing the electron and hole simultaneously. We show that, in one dimension, the projected electron and hole position operators fail to commute when the covariant derivative of the quantum geometric dipole (QGD) matrix (the difference between the hole and electron non-Abelian Berry connections) is nonzero. This precludes a common eigenbasis and bounds the joint electron-hole spread from below. For one band, the internal dipole is gauge invariant and center-of-mass methods suffice; for multiple bands, no existing construction yields a gauge minimizing both position uncertainties. We introduce an “exciton spatial localizer,” a Hermitian operator embedding both projected positions in a Clifford-algebra structure. Its spectral minima locate the exciton's center-of-mass and dipole coordinates, while its eigenvectors yield exciton Wannier functions jointly localized in electron and hole coordinates without gauge fixing, an ansatz, or iterative optimization. In an interacting bilayer model, combined reflection–time-reversal symmetry or a nonsymmorphic particle–hole symmetry forces the QGD matrix to be traceless at every momentum while allowing it to remain nonzero. A two-band exciton subspace with zero net internal dipole then decomposes into a symmetry-related pair of exciton Wannier functions with opposite center-of-mass positions and internal dipoles. Adding the interlayer dipole as a Clifford component further separates intralayer and interlayer exciton Wannier functions in a six-band subspace.
Read the full abstract

High Energy Physics (Phenomenology)

Tremaine-Gunn Control: Evading Bounds on Light Fermion Dark Matter

Pauli exclusion is often regarded as imposing a model-independent lower bound on the mass of fermionic dark matter in galaxies. We show that dark-sector interactions can substantially weaken this conclusion. In particular, eV-scale fermions can form bound structures with the characteristic sizes and densities of dwarf galaxies, thereby circumventing the conventional Tremaine-Gunn mass constraint. Within these objects, the degeneracy pressure is balanced by an attractive finite-range scalar Yukawa force that is significantly stronger than gravity. The interaction acts on dwarf-galaxy scales while remaining screened on larger scales. We derive the evolution of the resulting dark-matter fluid and identify a viable cosmological history in which structures first form gravitationally, with the new interaction becoming active only at late times. As a proof of concept, we construct a model in which a single late-time phase transition generates the required dynamics. These results provide a new route to exceptionally light fermionic dark matter and demonstrate that dark-sector interactions can qualitatively alter conventional phase-space limits.
Read the full abstract

All authors Joel Barir, Diego Blas, Anubhav Mathur, Tomer Volansky

Quantum Physics

Non-Markovian interactions with pulses of quantum radiation

The interaction of a traveling pulse of quantum light with a localized quantum system shows non-Markovian features when scattering takes place in a structured environment. Here, we devise a non-Markovian input-output theory by coherently coupling the scatterer to a pseudomode, which decays into a Markovian reservoir. This incorporates memory effects of a structured environment without altering the cascaded nature of the Lindblad master equation, whose solution provides the quantum state of the output field of any desired mode. We apply our theory to the stimulated emission by a two-level atom, and to transmission of a Gaussian pulse through a cavity. We observe that the non-Markovian revival of coherence in the scatterer distorts the single-mode nature of the incoming pulse, thus resulting in a multimode output field.
Read the full abstract

Computer Vision

From Visual Cues to Spoken Narration: Rethinking Audio Description

Audio Description (AD) provides spoken narration of visual events during dialogue gaps, making movies accessible to visually impaired audiences. The problem requires determining both what (which visual event) and when (position for inserting the AD) to narrate, to achieve the best user experience. Prior work has largely reduced the problem to video captioning of pre-segmented video clips, i.e., what is largely predefined and when is ignored entirely. We propose Cue2Narrate, a two-stage pipeline that jointly predicts what and when to narrate in longer untrimmed movie clips. A dual-head audio-visual localizer predicts two temporally distinct windows per AD utterance: a visual cue window and a spoken narration window. A LoRA-adapted VLM then generates concise ADs from the predicted visual evidence, trained with a Description Ranking Loss that ranks captions (negative samples) of the same frames lower than the GT AD. To benchmark this new problem statement, we introduce the LongLSMDC benchmark with up to 8-min movie clips (~6.5min on average). On LongLSMDC, Cue2Narrate outperforms video-only and audio-only localization baselines by 5–12 points in avg. mAP. Under both predicted- and GT-window evaluation, Cue2Narrate improves AD generation over the corresponding fine-tuned base VLM. These results establish the first benchmark for multi-segment AD generation on long-form clips. Data & Code: https://github.com/multimodal-ai-lab/Cue2Narrate
Read the full abstract

All authors Akshita Gupta, Aditya Arora, Federico Tombari, Marcus Rohrbach, Anna Rohrbach

Quantum Physics

Bilinear correlations in Fluctuating Gaussian States with anti-unitary symmetries

We study bilinear order-parameter correlations in a general class of sign problem-free systems that are described by what we call Fluctuating Gaussian States (FGS) with anti-unitary symmetries, where the weight of each Gaussian measurement in the space-time path integral is positive-definite. Supported by Monte Carlo simulations on fluctuating Gaussian fermionic and bosonic examples, we argue that such systems generally exhibit two types of broken-symmetry phases: a saddle point phase and a non-local contraction phase, with the competition between the two determined by the stability of the fluctuation saddle point of FGS. Due to the Fermi liquid instability in fermionic FGS with anti-unitary symmetries, we also discuss the possibility of representing fermionic FGS with a particular Projected Entangled-Pair State (PEPS) ansatz, which we argue a certain construction provides an efficient description for the saddle point phase, but not for the non-local contraction phase. Finally, we discuss the potential implication of our result in non-equilibrium systems and stabilizing a target bilinear order.
Read the full abstract

Cryptography and Security

Hearing the Whispers: Black-Box Membership Inference Attacks on Finetuned TTS Models

Text-to-Speech (TTS) foundation models are increasingly fine-tuned on private datasets to synthesize highly personalized voices, introducing severe privacy risks by exposing both biometric identities and sensitive speech content. Existing black-box membership inference attacks (MIAs) follow a two-stage pipeline of query generation and representation engineering, both of which face unique challenges when adapted to TTS. For query generation, dual conditioning on synthesis text and reference speech creates a large and underexplored query design space with no established criterion for identifying an effective query. For representation engineering, the multi-level speech characteristics and temporal variability of speech make low-level representations and direct comparisons inadequate for capturing membership signals. To address these challenges, we present the first black-box MIA framework explicitly tailored to TTS models at both the speaker and record levels. For query generation, we characterize the feasible query space and establish two criteria, scorable extent and memorization elicitation, for evaluating five representative queries, identifying recitation as the strongest. For representation engineering, we obtain multi-level speech representations from embedding models and temporally align the generated and target audio for fine-grained comparison. Evaluations across three state-of-the-art TTS models (CosyVoice2, F5-TTS, and XTTS-v2) fine-tuned on two benchmark datasets (VCTK and British Dialect) reveal severe privacy leakage: speaker-level AUC remains above 0.80 and approaches 1.0 in the strongest settings, while record-level AUC ranges from 0.80 to 0.90 and remains effective even in challenging scenarios where both members and non-members are of the same speakers. We further identify speech characteristics associated with disproportionate vulnerability to memorization.
Read the full abstract

All authors Kunlin Cai, Kaiyuan Zhang, Zihang Xiang, Jinghuai Zhang, Abeer Alwan, Fnu Suya, Yuan Tian

Earth and Planetary Astrophysics

From orbit to ground: pre-impact meteorite strewn field predictions for imminent impactors and meteorite recovery

The flux of meteoroids reaching the Earth is continuous, ranging from microscopic grains to occasional metre and decametre scale bodies. The smallest ones fully ablate in the upper atmosphere, whereas sufficiently large or strong objects survive entry and deposit fragments on the ground as meteorites. Predicting where these fragments land, and reconstructing the atmospheric trajectory and fragmentation sequence that produced them, is central both to hazard assessment and to the recovery of freshly fallen material. The accuracy of such predictions, however, remains limited by poorly constrained fragmentation processes and by sparse, heterogeneous observational coverage of individual events. Traditional strewn field simulations rely on detailed fireball data and event-specific assumptions on fragment masses, aerodynamics, and breakup. These approaches are effective for well-instrumented events, but their applicability degrades rapidly when observations are sparse, often resulting in huge uncertainties. We present an ab initio framework predicting strewn fields of near-Earth asteroids directly from pre-impact orbital solutions. It propagates luminous trajectory and dark flight using a physics-based translational dynamics model and realistic atmospheric conditions, without requiring fireball triangulation or event-specific tuning. Validation against recent asteroid falls with recovered meteorites shows agreement with observations, with nominal solutions reproducing fall locations within 100-200 m. The new method has been integrated into the ESA Aegis pipeline, which now enables hours-ahead computation of impact locations, supporting recovery efforts, minimizing contamination, and, where warranted by object size and predicted ground hazard, civil-protection decision making.
Read the full abstract

All authors Anna Moscati, Marco Fenucci, Laura Faggioli, Marco Micheli, Francisco Ocaña, Juan Luis Cano

Earth and Planetary Astrophysics

Mass constraints for the K2-223 system planets: An ultra-short-period sub-Earth, a short-period super-Earth, and a tentative long-period giant planet

We present mass constraints of two short-period Earth-sized planets transiting K2-223 based on high-precision radial velocity measurements from HARPS-N and ESPRESSO, as well as a tentative indication of an outer Jupiter-like planet orbiting the G1V dwarf K2-223. With a radius of $R_\mathrm{b}=0.79\pm0.10$ $\mathrm{R_{\oplus}}$ and a 3$\sigma$ upper mass limit $M_\mathrm{b}<2.8$ $\mathrm{M_{\oplus}}$, K2-223 b belongs to the small group of known sub-Earth planets and is currently the smallest known ultra-short-period planet ($P_\mathrm{b}\approx$ 0.5 day) transiting a solar-type star. With a radius of $R_\mathrm{c}=1.41\pm0.15$ $\mathrm{R_{\oplus}}$, a mass of $M_\mathrm{c}=4.2\pm1.3$ $\mathrm{M_{\oplus}}$, and a density of $\rho_\mathrm{c}=8.3\pm3.7$ $\mathrm{g\,cm^{-3}}$, K2-223 c is a short-period ($P_\mathrm{c}\approx$ 4.5 days) super-Earth. Based on almost 6.5 years of radial velocity monitoring of K2-223 with the HARPS-N spectrograph, we identified a tentative giant planet with an orbital period of $P_\mathrm{d}=4.53^{+0.32}_{-0.34}$ $\mathrm{years}$ and a minimum mass of $1.29^{+0.23}_{-0.17}$ $\mathrm{M_{\rm Jup}}$. Two close-in small planets accompanied by a distant candidate Jupiter-like companion would make K2-223 a system with a rare architecture. This is valuable for testing scenarios of planetary formation and evolution. The extreme proximity of K2-223 b to the parent star means that relativistic and tidal perturbations to Newtonian gravity need to be considered. We discuss the timescales and amplitudes of these effects in the context of the radial velocity model and dynamical simulations of the K2-223 multiple-planet system. We also present a parametrisation of the planets with directly measured masses, radii, and bulk densities in terms of the orbital period normalised by the Roche period. This provides an alternative representation in the context of the Neptune desert.
Read the full abstract

All authors Dawid Jankowski, Grzegorz Nowak, Gaia Lacedelli, Enric Pallé, Krzysztof Goździewski, Thomas Masseron, Ilaria Carleo, Rafael Luque, Felipe Murgas, Davide Gandolfi, Artie P. Hatzes, William Cochran, Pedro Figueira, Rafael A. García, Samuel Geraldía-González, Judith Korth, Pierrot Lamontagne, John H. Livingston, Savita Mathur, Giuseppe Morello, Jaume Orell-Miquel, Dinil B. Palakkatharappil, Carina M. Persson, Seth Redfield, Atanas K. Stefanov, Vincent Van Eylen

General Relativity and Quantum Cosmology

How Much Geometry Does Horizon Entropy Determine? A Discourse on Entropy, Gravity, and Observable Quantities

Building on Jacobson's thermodynamics derivation of Einstein field equations, the "generalized entropy and varying-G" (GEVAG) framework was proposed to consistently incorporate generalized horizon entropy $S=f(A)/4G$ into both the thermodynamics and the geometry of spacetime. This leads to an effective gravitational coupling $G_\text{eff}=G/f'(A)$. A related "entropy-gravity correspondence" (EGC) instead is equivalent to promoting this horizon relation to a radially varying bulk coupling $G_\text{eff}(r)$. We clarify the similarities and fundamental distinctions between these approaches. Although they share the same horizon consistency relation, their global causal structures and asymptotic geometries can differ substantially. In addition, the radial prescription of EGC introduces an additional ambiguity for multi-center configurations. We further examine a near-horizon extension of GEVAG, which restores the equivalence between thermodynamic and ADM masses that is absent in the horizon-only formulation. We then discuss the implications of these different prescriptions for the Bekenstein bound. The Bekenstein bound provides a surprising link between $G_\text{eff}$ in near horizon GEVAG, EGC, and the one obtained in the recent $f(R)$-reconstruction approach of generalized entropy geometry via Wald entropy. Finally, we raise the most important question: what exactly is the $G$ measured in laboratories?
Read the full abstract

Quantum Physics

Numerical simulation of D-Wave's quantum advantage experiment with time-dependent variational Monte Carlo

Programmable quantum annealers can realize real-time dynamics of frustrated transverse-field Ising models on large, nontrivial graphs. Recent work by King et al. argued that the classical simulation of such experiments would require exponential computational resources for classical methods such as tensor networks and neural quantum states. Here, we numerically simulate the D-Wave spin-glass annealing protocol with time-dependent variational Monte Carlo (t-VMC) using a correlator state tailored to spin-glass dynamics. For the two-dimensional cylinder, three-dimensional dimer, diamond, and biclique instances considered, at annealing times of 7 and 20 ns, we show that systematically increasing the variational ansatz size enables t-VMC to approximate the final two-spin correlation errors of the quantum processing unit (QPU). We also perform an accurate large-scale simulation of a challenging biclique instance for which no other variational method is known to work, and we find close agreement with the quantum annealer. Through ablation studies, we identify poor Markov-chain mixing, high-variance local-energy estimators, and stochastic Runge-Kutta error estimates as the principal numerical failure modes. We address these numerical issues by using parallel tempering, blurred sampling and an importance-weighted differential equation solver, thereby clarifying the numerical requirements for stable, large-scale t-VMC simulations. Our results extend the frontier of classical simulation while providing a realistic assessment of the computational costs of simulating quantum dynamics at this scale.
Read the full abstract

High Energy Physics (Phenomenology)

Electromagnetic tails of $B$-meson light-cone distribution amplitudes

We investigate electromagnetic corrections to the $B$-meson decay constant and light-cone distribution amplitudes (LCDAs) within the framework of heavy-quark effective theory. The inclusion of QED effects extends the support of the LCDAs to negative values of the light-cone momentum. We clarify the relation between this region and QED-generalized decay constants sensitive to electromagnetic final-state radiation. We derive model-independent expressions for the asymptotic behavior in the large negative-momentum region of the QED-generalized LCDAs $\Phi_\pm(\omega,\mu)$ in terms of integrals over the two- and three-particle LCDAs defined in pure QCD. We show that our result for $\Phi_+$ is consistent with the renormalization-group equation satisfied by this object and study its evolution using the known one-loop anomalous dimension. Finally, we propose a phenomenological model consistent with these constraints
Read the full abstract

High Energy Physics (Phenomenology)

Primordial Black Hole Assisted Dirac Leptogenesis

Dirac leptogenesis offers a qualitatively different route to the baryon asymmetry: total lepton number is conserved, but equal and opposite asymmetries are stored in the Standard Model and right handed neutrino sectors, with only the visible component processed by electroweak sphalerons. We present the first dedicated study of how evaporating primordial black holes (PBHs) modify this mechanism. In a minimal charged-scalar mediator setup, we solve the coupled Boltzmann system including thermal production, Hawking emission of the mediator $X$ and of right handed neutrinos, the PBH contribution to the expansion rate, and entropy injection. We find that Hawking emission can populate $X$ in regions where thermal production is inefficient, allowing its subsequent CP-violating decays to enhance the baryon asymmetry. For representative benchmarks, PBHs extend the successful region from mediator masses around $10^9\,{\rm GeV}$ down to about $10^6\,{\rm GeV}$ and can compensate for smaller effective CP asymmetries. This effect is not universal: where thermal Dirac leptogenesis is already efficient, PBHs can instead reduce the final asymmetry. Their evaporation also deposits energy in the right handed neutrino sector and contributes to $\Delta N_{\rm eff}$. In the PBH parameter scan, part of the region that reproduces the observed baryon asymmetry lies below the Planck reference level $\Delta N_{\rm eff}=0.284$ and above the projected CMB-S4 sensitivity $\Delta N_{\rm eff}=0.06$.
Read the full abstract

High Energy Physics (Phenomenology)

SN1987A constraints on the neutrino-dark-fermion interaction from resonant scattering with C$\nu$B

Neutrino self-interactions mediated by a light scalar offer a compelling resolution to cosmological tensions and may naturally arise in neutrino mass generation mechanisms. When the scalar also couples to a light dark-sector fermion, supernova neutrinos can resonantly annihilate with the cosmic neutrino background (C$\nu$B) into invisible dark radiation, depleting the flux en route to Earth. We use this depletion to constrain the neutrino-scalar coupling from the SN1987A data. Within a Bayesian framework we analyze the data for two supernova neutrino emission models, a parameterized model and a 2D hydrodynamic simulation, and for two coupling types, a mass-independent one and a mass-proportional one. Our results show that new-physics limits from SN1987A cannot be quoted independently of the heavy-flavor emission, the flavor conversion, or the coupling structure. Finally, we forecast that a future high-statistics burst recorded by Hyper-Kamiokande would restore a meaningful upper bound even when flavor conversion is included.
Read the full abstract

All authors Christina Gao, Ke Hu, Kun-Feng Lyu, Vincent Zou

High Energy Astrophysical Phenomena

Measuring Small-scale Turbulence in Supernova Remnants with Jitter Radiation

Hard X-ray emission from Cas A has been attributed to jitter radiation, implying magnetic turbulence on scales near or below the ion Larmor radius. We propose these fields arise from ion-scale micro-instabilities (e.g., mirror and firehose) in the shock downstream, and use high-resolution hybrid simulations to demonstrate the generation of such small-scale turbulence. Its spectral index, once numerical dissipation is accounted for, is consistent with that inferred for Cas A, suggesting that jitter radiation is a direct probe of ion-scale turbulence in supernova remnants.
Read the full abstract

All authors Emily Simon, Damiano Caprioli, Niccolò Bucciantini, Emanuele Greco, Marco Miceli

High Energy Physics (Theory)

On-shell recursion and soft theorems for worldline quantum field theory

We derive a set of on-shell recursion relations for amplitudes in worldline quantum field theory (WQFT), an effective theory of gravity coupled to a massive worldline which is relevant for computing classical gravitational observables. We overcome the poor large-$z$ falloff of amplitudes under an all-line complex momentum deformation by deriving and exploiting a new soft-graviton theorem in the presence of a worldline. This theorem generalizes the leading (Weinberg) and subleading soft theorems with additional contributions due to soft emission from the worldline. Our main result is a $d$-dimensional recursion formula, expressing any $n$-graviton rational amplitude in terms of lower-point amplitudes from WQFT and pure gravity. Using the soft-worldline theorem, we can obtain amplitudes with any number of external worldline fluctuations from the corresponding all-graviton amplitudes, making the recursion sufficient to construct any rational amplitude of the theory. We verify our results by explicitly computing two-graviton and three-graviton amplitudes.
Read the full abstract

General Relativity and Quantum Cosmology

One-cycle radiation-reaction Magnusian for eccentric binaries through relative 1PN order

We derive the one-cycle first Magnusian for nonspinning compact binaries on planar, bound eccentric orbits through relative first post-Newtonian (1PN) order and linear order in radiation reaction, and construct the associated first-order periastron return map. Combining the conservative 1PN dynamics and its Jacobi transport with the 2.5PN and 3.5PN radiation-reaction forces, we obtain all four components in closed form for \(0<e<1\), without expanding in eccentricity. The dependence on all eight Iyer–Will radiation-reaction gauge parameters reduces to periodic coboundaries and therefore cancels over a radial cycle. The result reproduces the averaged energy and angular-momentum losses, their circular limit, and the secular evolution of the 1PN radial orbital elements. The two remaining components encode the dissipative angle-sector information. Projecting the resulting phase-space update onto the perturbed periastron return section yields the corrections to the return time and apsidal phase. We then compare the strict first-order map with a partially exponentiated map built from the same Magnusian. The comparison is observable dependent: at fixed periastron index, the strict map gives smaller residuals for a representative evolution over 40 radial cycles, whereas partial exponentiation reduces the accumulated periastron-timing residual by about an order of magnitude. Across a representative one-cycle grid at equal masses, partial exponentiation improves the radial-action and energy updates at every sampled point.
Read the full abstract

High Energy Physics (Phenomenology)

Axion Quality from Exact Proton Stability

A discrete, anomaly-free $\mathbb Z_9$ gauge symmetry may persist to the deep infrared, exactly forbidding proton decay. This $\mathbb Z_9$ can emerge from Higgsing a lepton-flavour non-universal $U(1)_X$ that generates realistic neutrino masses and mixings through a high-scale seesaw and predicts an additional light Goldstone. Adding an anomaly-free chiral heavy-quark sector turns this mode into the QCD axion. The same selection rules that ban proton decay also forbid PQ-violating operators below a suitably high dimension, $d_{\rm PQ}\gtrsim10$, including operators involving the heavy quarks that radiatively match onto the scalar potential. We construct explicit models viable when PQ breaking occurs before or after inflation; in the latter case, avoiding stable heavy relics and eliminating the string-wall network tightly constrain the heavy-quark spectrum. We identify restricted regions of parameter space where thermal leptogenesis remains viable and, for $d_{\rm PQ}\geq13$, can coexist with a high-quality axion constituting all of the dark matter.
Read the full abstract

Astrophysics of Galaxies

JWST Reveals a Candidate Supermassive Black Hole Binary at z=4.3 in the Brightest Sub-millimeter Galaxy in COSMOS-Web

We present JWST/NIRSpec PRISM and G395M grating spectroscopy for AzTEC-1, a massive sub-mm bright galaxy at $z=4.34$ in the COSMOS extragalactic field. The PRISM spectrum reveals strong H$\alpha$, a significant Balmer break, and no H$\beta$ detection, indicating a $100-400$ Myr-old stellar population and high dust attenuation. BPT line ratios indicate the presence of an Active Galactic Nucleus (AGN). Decomposing narrow and broad line components, we recover broad, blueshifted H$\alpha$ with a velocity offset of $1245{\,\rm km\,s^{-1}}$ from the systemic narrow line velocity and with FWHM$\,\sim2500\,{\rm km\,s^{-1}}$. AzTEC-1's smooth morphology and stellar age is suggestive of a past merger-induced starburst period that would have brought in a second supermassive black hole, raising the possibility for a binary supermassive black hole system. In this scenario, we assume that the lower mass black hole hosts a broad line region orbiting a quiescent primary. Evidence for an extended outflow is not found in the 2D spectrum, NIRCam imaging, resolved ALMA observations of dust continuum, or CO, [C II]$_{157\,\mu \rm m}$ and [N II]$_{\rm 205\,\mu m}$ kinematics. AzTEC-1's high central gas mass surface density and dynamically unstable gas disk indicates that massive gas clouds external to the candidate binary SMBH's orbit might have played a role in stalling infall from $\sim10$ Myr to $\sim100$ Myr through dynamical torques, which has been theorized to occur in the nuclei of massive galaxies like AzTEC-1. If the supermassive black hole binary is confirmed, AzTEC-1 would be an excellent laboratory into the astrophysics driving low-frequency gravitational wave detections.
Read the full abstract

All authors Jed McKinney, Ansh Gupta, Julian B. Munoz, John Chisholm, Caitlin M. Casey, Stephanie M. Urbano Stawinski, Olivia Cooper, Erini Lambrides, Hollis Akins, Maximilien Franco, Archana Aravindan, Seiji Fujimoto, Kohei Inayoshi, Andreas L. Faisst, Jeyhan S. Kartaltepe, Michael Boylan-Kolchin

Quantum Physics

Purification of photonic graph states

Graph states constitute the main building block for quantum computing with photons. Quantum emitters with a hosted spin can deterministically generate photonic graph states, strongly lowering the overhead of multiplexing highly probabilistic linear-optics graph state generation. However, they generally suffer from various noise sources, resulting in reduced fidelities of the produced states. To mitigate this issue, we develop purification schemes for entangled photonic states. We first develop purification schemes to purify arbitrary photonic GHZ and other CSS states, which we generalize to all photonic graph states and stabilizer states. The proposed purification schemes have a high success probability of up to $1/2$ and require only linear optics and photon detectors. We optimize cascaded purification schemes for various graph states, taking into account phenomenological Pauli errors or physical noise in time-bin-encoded graph state generation with quantum emitters.
Read the full abstract

All authors Matthias C. Löbl, Aliki A. Capatos, Ming Lai Chan, Peter Lodahl, Anders Søndberg Sørensen, Stefano Paesani

Astrophysics of Galaxies

The Comprehensive Investigations of the NGC 7419 Open Cluster and the New Perspective on Mass Populations

We present a structural, kinematic, and photometric analysis of the young, red supergiant- and Be star-rich open cluster NGC 7419 using Gaia DR3 and 2MASS data. Using the HDBSCAN algorithm and conservative membership criteria, we identify 734 high-probability cluster members, separating the cluster from the dense Perseus Arm background. Isochrone fitting to the Gaia DR3 color-magnitude diagram, treating classical Be stars separately to account for their circumstellar excess, yields an age of 15 +/- 2 Myr, a distance of 3161 +/- 261 pc, a distance modulus of 12.50 +/- 0.08 mag, and a color excess E(GBP-GRP) = 2.58 +/- 0.06 mag. An independent 2MASS analysis gives a consistent distance modulus of 12.53 mag and E(J-Ks) = 0.99 mag. The Be stars show a prominent infrared excess in J-Ks, consistent with free-free and bound-free emission from circumstellar disks. We derive mean proper motions of -2.738 +/- 0.115 and -1.604 +/- 0.149 mas yr^-1 and a mean parallax of 0.288 +/- 0.062 mas, corresponding to a trigonometric distance of 3316 +/- 317 pc. The photometric distance is adopted as the primary estimate. A generalized King profile gives a core radius of 1.53 +/- 0.08 arcmin and a limiting radius of 8.16 +/- 1.1 arcmin. The stellar mass distribution is best described by a three-component Gaussian model, which we interpret as three distinct mass populations. Their coexistence and incomplete mixing are consistent with the young dynamical state of NGC 7419. Hierarchical formation provides a plausible framework for their origin.
Read the full abstract

Astrophysics of Galaxies

The HERON Coma Project: Exploring the low-surface-brightness frontiers in the Coma cluster

The technical complexity of data processing and residual instrumental systematics remain among the main barriers to reaching the faintest low-surface-brightness (LSB) regimes in the local Universe. Here, we present the HERON Coma Project, a deep $g$- and $r$-band imaging campaign on the Coma cluster carried out with the LSB-optimized 71-cm Jeanne Rich Telescope. The observing and processing strategy combines accurate flat-fielding, sky-background modeling and extended point spread function (PSF) characterization. As representative scientific applications, we explore the complex associated with the giant elliptical galaxies NGC 4839 and NGC 4816, in the southwestern region of Coma. Spatially coherent emission is recovered in direct imaging to $\mu_{\overline{g+r}}\simeq30~\mathrm{mag\,arcsec^{-2}}$, consistent with the nominal sensitivity inferred from the scaling of the background fluctuations. The PSF-deconvolved photometric profiles of NGC 4839 and NGC 4816 are traced to generalized isophotal radii of 705 and 329 kpc, reaching $\mu_{\overline{g+r}}=32.2$ and $31.6~\mathrm{mag\,arcsec^{-2}}$, with data-derived signal-to-noise ratios of 4.1 and 7.0, respectively. The agreement between direct detection and the predictions from background fluctuations provides no clear evidence of a systematics-dominated detection floor over the surface brightness range and spatial scales examined. We also assess the observational challenges involved in extending reliable measurements to still fainter surface brightness levels. The HERON Coma Project opens a new window onto the faint stellar record of the complex and ongoing assembly of Coma, while demonstrating the potential of modest ground-based telescopes with suppressed instrumental reflections to provide highly competitive measurements of extended diffuse structures. We make the final calibrated images publicly available to the community.
Read the full abstract

All authors Javier Román, R. Michael Rich, Pablo M. Sánchez-Alarcón, Yolanda Jiménez-Teja, Chester Li, Mónica Esquide

Astrophysics of Galaxies

Linking neutral gas inflows and outflows to offsets in the star-forming main sequence and mass-metallicity relation

Gas inflows and outflows regulate galaxy growth, but direct observational links between measured gas flows and galaxy scaling relations remain limited. Using ~6,000 star-forming galaxies with down-the-barrel Na I D absorption from DESI DR2, we examine how systems with detected neutral-gas inflows and outflows populate the star-forming main sequence (SFMS) and mass-metallicity relation (MZR). Inflow and outflow hosts are compared with stellar-mass- and redshift-matched controls, and with SFMS and MZR fits derived from galaxies without detected gas flows. Outflow hosts (v_flow $\leq$ -50 km s$^{-1}$) show enhanced sSFRs by 0.25-0.40 dex and elevated central metallicities by 0.04-0.06 dex in the lower-redshift sample. Slow inflow hosts (0 $<$ v_flow $<$ 100 km s$^{-1}$) show similarly enhanced sSFRs of 0.20-0.30 dex, but no significant metallicity offset, while fast inflow hosts (v_flow $\geq$ 100 km s$^{-1}$) show weaker SFR enhancement and modestly lower metallicities. Together, these trends support a regulator picture in which neutral gas flows trace different phases of the baryon cycle. Slow inflow hosts lie above the SFMS, consistent with accretion sustaining enhanced star formation without strong central metallicity dilution. This may indicate that inflowing gas is already metal-enriched or has mixed or enriched over extended timescales. By contrast, outflow hosts lie near the upper 1{$\sigma$} SFMS envelope, consistent with feedback regulating subsequent growth. Gas-flow hosts also show small but systematic offsets in the narrow 4000 Å break strength ($D_n$4000) relative to controls matched in redshift, stellar mass and SFR. Our results show that neutral gas flows are associated with population-level offsets from the SFMS and MZR, consistent with a baryon-cycle contribution to scaling-relation scatter.
Read the full abstract

All authors S. Weng, M. Pieri, A. Saintonge, D. Scholte, D. Muñoz Santos, T. Hu, J. Aguilar, S. Ahlen, F. Beutler, D. Bianchi, D. Brooks, A. Carnero Rosell, F. J. Castander, T. Claybaugh, A. de la Macorra, B. Dey, P. Doel, V. A. Fawcett, J. E. Forero-Romero, E. Gaztañaga, S. Gontcho A Gontcho, G. Gutierrez, A. Kremin, M. Landriau, L. Le Guillou, A. Meisner, R. Miquel, J. Moustakas, S. Nadathur, W. J. Percival, F. Prada, I. Pérez-Ràfols, C. Ravoux, G. Rossi, R. Ruggeri, E. Sanchez, C. Saulder, D. Schlegel, J. Silber, M. Siudek, G. Tarlé, B. A. Weaver, H. Zou

High Energy Physics (Theory)

Holographic algebras at null infinity

We construct an algebra of observables associated to a cut of null infinity in asymptotically flat spacetimes. The Bondi mass associated with a sharp cut of future null infinity is not a well-defined quantum operator: its fluctuations diverge even after smearing in retarded time. We instead introduce a finite-radius, time-smeared version of the Bondi mass and adjoin this operator to the matter and graviton observables in an arbitrarily small asymptotic neighborhood of the cut. We separately analyze spacetimes with and without black holes and find, in both cases, Type III$_1$ von Neumann algebras that satisfy non-trivial nesting relations. In Minkowski spacetime, the resulting algebra reconstructs the spacelike wedge associated with the cut. In a stationary black hole spacetime, it reconstructs the region bounded by the cut and the black hole bifurcation surface, providing an asymptotically flat analogue of an entanglement wedge. In the limit as the cut is moved to past infinity, we recover a Type I$_{\infty}$ algebra for spacetimes without black holes and a Type II$_{\infty}$ algebra for black hole spacetimes. For algebras at finite cuts of retarded time, we construct a Type II$_\infty$ regularization of the black hole algebra whose renormalized von Neumann entropy agrees with the generalized entropy. In appendices, we prove two technical results about quantum fields, in a Schwarzschild spacetime, that asymptote to the Minkowski vacuum at infinity: a split property for unbounded, spacelike separated regions and the construction of a faithful, normal, and semifinite Hartle-Hawking weight whose modular flow is Schwarzschild time evolution.
Read the full abstract

All authors Chang-Han Chen, Geoff Penington, Gautam Satishchandran, Elisa Tabor

Machine Learning

Generative Diffusion Surrogates with Analytical Variance Schedule

Stochastic transport describes physical systems in which an initially structured distribution spreads under unresolved forcing, scattering, or heterogeneous media. Useful surrogates for such systems should be probabilistic, time-resolved, and able to represent non-Gaussian distributional structure. Generative diffusion models, which corrupt data with Gaussian noise and learn a reverse flow back to structured states, have these properties. Their noise schedules, however, are usually chosen heuristically: image and audio generation—the canonical use cases—provide no physical clock. In transport, by contrast, the variance, or mean-square displacement, is often known from macroscopic theory or empirical scaling even when the full distribution is not. Here we prescribe the forward noising rate as the time derivative of this variance, turning generative time into a calibrated transport clock. The variance path is enforced by construction, while the learned score field represents how non-Gaussian structure inherited from entrance data is smoothed along that path, requiring no intermediate-time physical transport data. For ballistic-to-diffusive transport in turbulent plasmas, the surrogate matches test-particle distributions, reproduces the laboratory-measured variance scale, and tracks the simulated kurtosis evolution without schedule tuning, enabling calibrated emulation and likelihood-based inference.
Read the full abstract

All authors Patrick Reichherzer, Gianluca Gregori, David N. Hosking, Subir Sarkar

General Relativity and Quantum Cosmology

Geodesics and thermodynamics of a $\kappa$-deformed anti-de Sitter black hole surrounded by a quintessence field

We investigate the geodesic structure and thermodynamics of a $\kappa$-deformed Schwarzschild-anti-de Sitter black hole in a Kiselev quintessence field, where dynamics are governed by a mass-dependent lapse deformation. This term produces an inner Cauchy horizon without requiring charge or spin, yet it fails to resolve the central singularity; instead, the Kretschmann scalar diverges more strongly as $r^{-10}$. Because the deformation amplitude couples to the mass, the standard Bekenstein-Hawking area law violates the first law of thermodynamics. We resolve this by either deriving a modified entropy or rescaling the mass, both of which restore the first law and yield a consistent Smarr relation. Optically, an increasing deformation parameter shrinks both the photon sphere and the critical impact parameter. We also derive a closed-form Joule-Thomson coefficient, revealing that the $\kappa$-deformation triggers an inversion curve a feature absent in standard neutral SAdS. Notably, this occurs without generating a van der Waals critical point, demonstrating that Joule-Thomson inversion and van der Waals criticality can be cleanly decoupled. Finally, the deformation suppresses peak Hawking emission by a factor of roughly four and increases flux sparsity, an effect driven almost entirely by a drop in temperature rather than changes to the black hole shadow.
Read the full abstract

All authors Faizuddin Ahmed, Ahmad Al-Badawi, Izzet Sakalli, Erdem Sucu

Optics

Multidimensional Double Refraction Microscopy

Multidimensional optical microscopy - extracting 3D location, orientation, and spectral information from nanoscale emitters - is key to modern nanoscience. However, this multidimensional capability requires complex instrumental setups, demanding exceptional stability and specialized expertise in optics. Here, we overcome these barriers by introducing Double Refraction Microscopy (DRM), a passive imaging method providing a straightforward pathway to multidimensional super-resolution imaging. Based on bifocal imaging via a birefringent microscope slide, DRM splits nanoprobe light into a characteristic double image encoding 3D location, orientation, and spectral signature, while requiring zero hardware modifications to standard wide-field microscopes. To showcase its robustness, we implement DRM across three distinct slides of varying materials and thicknesses, calibrating 3D localization with 450-nm nanoparticles and demonstrating high orientation sensitivity using polarization-controlled emitters. Furthermore, we confirm spectral readout using a pinhole array acting as a grid of point emitters back-illuminated at various wavelengths to generate distinct spectral signatures. Finally, we showcase practical utility by performing a two-colour 3D localization measurement within a $10\,\mu\text{m}$ gap, utilizing the single optical pathway of a standard epi-fluorescence microscope. Ultimately, DRM provides the missing optical hardware simplicity to complement established computational tools, promising to transfer advanced super-resolution capabilities from specialized facilities to every laboratory bench.
Read the full abstract

Software Engineering

RosettaBitcoin: An Artifact-Backed Experience Report on Verification Infrastructure for Agent-Assisted Consensus Validators

Agent-assisted software projects are often reported through demonstrations or aggregate benchmarks that conceal how correctness claims were admitted. This experience report studies RosettaBitcoin, a single-developer project that built twelve separately implemented Bitcoin testnet4 consensus validators, through its immutable 17 June 2026 software snapshot (DOI 10.5281/zenodo.20738249). We analyze the snapshot's tracked SQLite evidence database, curated artifact index, conformance fixtures, validation scripts, blocker records, and version history. At the snapshot, all twelve ports had port-owned 45/45 script-corpus proofs and strict 5,000-block baselines. Nine had canonical clean 50,000-block, 100,000-block, and post-100,000 validation lanes. Java had one 19.86-second near-tip maintenance artifact. No port had an empty-state-to-tip proof, and no port satisfied the project's binary full-node gate; Docker and live-node capability gaps remained. The artifact history also records a 3 h 17 min 57 s Zig scaffold-to-50,000 span, but the observed intervals describe non-equivalent tasks and cannot estimate effort, productivity, or causality. A separate diagnostic supplement preserves evidence that a pure-Mojo cryptographic backend validated fresh state to height 100,000 and resumed to 140,234, agreed on a 45-case shadow comparison, rejected six crafted invalid classes, and was killed by three targeted mutations. That evidence is noncanonical, noncomparable, and class-bounded. The case suggests that explicit failure records, fixtures, port-owned proofs, and validating imports can make agent-assisted systems more auditable. Controlled ablations and external replications are needed to test whether such infrastructure causally improves development outcomes.
Read the full abstract

Fluid Dynamics

A reconfigurable multi-axis cyber-physical framework for multi-regime fluid–structure interaction experiments

Fluid–structure interaction (FSI) experiments are typically built around mechanical dynamics and constraints imposed by the physical apparatus, so changing mass, stiffness, damping, or allowable motion often requires hardware reconfiguration. Here we present a reconfigurable cyber-physical framework in which these properties are instead assigned through software-defined dynamics. The system provides three translational and one rotational degree of freedom, each independently configurable as prescribed, load-responsive, or locked, with operating roles that can also be reassigned during a running experiment. Measured forces and torques are incorporated into real-time virtual dynamic models, while a common supervisory architecture coordinates multi-axis motion, mode switching, synchronized data acquisition, and diagnostic positioning. The prescribed-motion pathway is validated using a pitching hydrofoil by comparison with published thrust and power scaling trends, while the load-responsive pathway is evaluated using an active-heave/passive-pitch benchmark that reproduces the expected frequency-dependent resonant response over the tested conditions. The same platform is then reconfigured for intra-cycle active–passive pitching, coordinated vertical-axis turbine-surrogate motion, force-driven passive surge, and automated multilayer stereoscopic particle image velocimetry. These results demonstrate that distinct FSI boundary conditions and measurement requirements can be implemented within a common motion, sensing, and control architecture. By treating mechanical roles and constraints as software-defined experimental variables, the framework provides a reusable basis for reconfigurable FSI experiments without redesigning the underlying platform for each application.
Read the full abstract

All authors Zihan Zhang, Alex Sorensen, Qimin Feng, Orion Roberts, Linhao Jin, Qiang Zhong