For a permutation $u\in S_n$, let $N\ast u\in S_{Nn}$ be the permutation with scaled Lehmer code. For given $u,v,w\in S_n$ and integer $N$, the stretched Schubert coefficients are defined as $f_{u,v,w}(N):=c_{N*u,N*v}^{N*w}$. Our main result is that the function $f_{u,v,w}(N)$ is eventually quasi-polynomial. This proves Kirillov's conjecture (2004), that the generating function for the sequence $\{f_{u,v,w}(N)\}$ is rational. For the proof, we use combinatorics of pipe dreams to show that Schubert coefficients are given as an alternating sum of the numbers of integer points in certain polytopes. These polytopes behave nicely under stretching, and we use Ehrhart theory to obtain the result. As a consequence of the proof, we also present new counterexamples to the saturation conjecture for Schubert coefficients, and give computational applications.

Molecular communication suffers from severe inter-symbol interference, which makes constrained coding essential for reliable transmission. Run-length-limited ISI-mitigation codes are attractive because they select low-weight constrained codebooks, reducing ISI while allowing more molecules to be assigned to each transmitted 1-symbol under the usual molecular-communication normalization. Previous results showed strong bit-error-rate performance for these codes, but their original realization required full codebook generation and storage. This exponential storage growth is unsuitable for resource-constrained molecular communication channels and also limits the exploration of larger information dimensions. This is particularly important for nano-scale molecular communication, where transmitter and receiver nodes are expected to operate under severe memory and computational constraints. This paper removes that realization bottleneck by replacing full codebook storage with an enumerative realization based on Cover's ranking framework, constant-weight run-length-limited counting, and cumulative weight-layer offsets. The resulting encoder and decoder preserve the selected RLIM codebooks and the original projection-based decoding behavior while storing only polynomial-size counting tables. Storage and runtime measurements confirm the resulting exponential-to-polynomial reduction, and diffusion-based molecular-communication simulations show that the newly accessible larger-dimensional RLIM regimes can improve the best attainable bit-error-rate performance in the tested settings.

Accurate segmentation and localization of left atrial (LA) ablation scars from Late gadolinium enhancement (LGE)-MRI is essential for assessing the lesion completeness and guiding ablation therapy. Incomplete or discontinuous lesions can increase the recurrence rate of the therapy and inaccurate localization can misguide treatment planning. However, reliable quantification and localization of scar in LGE-MRI is challenging. The severely class imbalanced scar voxels, thin structure of the LA wall, and weak tissue contrast often lead to unrealistic scar predictions. In this paper, we propose a two stage nnUNet based framework that takes LA anatomy into account to help with more precise scar localization and segmentation. In the first stage, an nnUNet model is trained to segment the LA cavity. In the second stage, patient specific cavity and wall signed distance maps (SDMs) are derived from the predicted anatomy to use as geometry aware inputs, and explicitly encode each voxel's signed spatial relationship to the atrial cavity and wall. This approach transforms scar segmentation from a solely intensity-based classification into anatomy-conditioned localization task, providing a continuous spatial prior that stabilizes learning for the thin atrial wall and suppresses topologically invalid predictions. To further address boundary ambiguity, we introduce a wall ROI-masked weighted loss combined with boundary uncertainty-aware supervision strategy that restricts learning to the atrial wall, while accounting for severe class imbalance. We evaluated our approach on the LAScarQS 2022 dataset and achieved a Dice of 61.1% and ASSD of 1.711mm. Our reliable and effective framework improves scar segmentation and localization accuracy by enforcing anatomical validity through geometry-aware supervision, and lowering the false positive detections far away from the atrial wall.

Understanding the role of galactic scale bars in disk galaxy evolution requires detailed measurements of bar properties across galaxies hosting bars at many redshifts. We present measurements of bar lengths and widths in a sample of 8230 disk galaxies from Hubble Space Telescope (HST) Legacy surveys. The highest-redshift barred galaxies in the sample have $z \sim 3$; most have $z \leq 1$. Using a mass-complete sample from the COSMOS field, we examine bar properties and evolution within $0.25 < z < 1$ in galaxies with stellar mass $\log(M_{\ast}/M_{\odot}) \geq 9.5$. The lowest-mass galaxies in our sample have similar star formation rate (SFR) distributions whether or not they host bars. For galaxies with $\log(M_{\ast}/M_{\odot}) \geq 10$, barred galaxies are more likely to be quiescent or quenched, consistent with bars mainly participating in slow quenching processes. The median physical bar length increases with increasing stellar mass. Relative bar lengths and widths (as a fraction of disk radius) peak at stellar mass $\log(M_{\ast}/M_{\odot}) \sim 10.25$, and change together with mass such that the median ratio, a proxy for bar strength, does not significantly change with stellar mass. Bars in our sample tend to be slightly ($\approx 13$%) weaker at higher redshift. Quiescent and quenching galaxies have longer and wider bars than those in galaxies on or above the star-forming sequence, especially at lower redshift and higher masses; at the low-mass end of our sample, starburst galaxies host relatively longer and stronger bars. Our findings are consistent with other results from studies at both higher and lower redshift, cementing the fundamental importance of bars in disk galaxy evolution.

The generalized composition graph is used by Cardoso and some researchers for factorization of the adjacency spectrum and Laplacian of a simple graph. Because the generalized composition graph is an example of a set-theoretic linear operad, this operation can be iterated at more than one level, where the complex language of partition refinement in the iteration is represented in terms of Schr"oder trees. This allows us to generalize the factorization of the adjacency spectrum and Laplacian of a simple graph presented by Cardoso in terms of Schr"oder trees and colorings over the edges of a graph. Cardoso's technique has been generalized by other authors for the universal adjacency matrix of a graph. This work also presents generalized factorizations in terms of Schr"oder trees and colorings on the edges of a graph for the universal adjacency spectrum, the characteristic polynomial of the universal adjacency matrix, and the generalized characteristic polynomial of a graph.

We discovered generalized structures, named astrocytes due to their shape, that constitute a defined region characterizing regular behavior within the parameter plane (PP) of dynamical systems (DSs). Morphologically, they are characterized by a branch and a soma with several vertices (arms) and sometimes with multiple periodicities. A bunch of infinite astrocytes emerge through their branches from a region, in general, of low periodicity. Astrocytes are embedded in a quasiperiodic-chaotic scenario. The soma complexity (number of vertices) determines a kind of hierarchy of the astrocytes; moreover, bunches of subsequent structures from the astrocyte have been emphasized, revealing a self-similarity property. We conducted a detailed analysis in a Zeeman laser model, but we also observed astrocytes in many other DSs. The multiperiodicity exhibited by the astrocytes in their soma gives rise to harlequin dress-like patterns and tri-, quad-, and quint-critical points, which indicate the coexistence of different higher-order periodicities. In the concave borders of the soma, a doubling cascade of quint-points emerges as a bifurcation in the PP, defining regions of ordered sequences of higher periodicity in the route to chaos.

The inclusive cross section for top quark pair ($\mathrm{t\bar{t}}$) production in lead-lead (PbPb) collisions is reported for the first time at a center-of-mass energy per nucleon pair of 5.36 TeV. The analysis uses data corresponding to an integrated luminosity of 1.58 nb$^{-1}$ collected by the CMS experiment at the CERN LHC in 2023. The $\mathrm{t\bar{t}}$ production cross section, $σ_{\mathrm{t\bar{t}}} $ = 3.42$^{+0.54}_{-0.51}$(stat)$^{+0.50}_{-0.43}$(syst) $μ$b, is measured in dilepton final states using a fit to a multivariate discriminator that combines the decay electron and muon kinematic properties with the multiplicity of bottom quark jets. The result is consistent with perturbative quantum chromodynamics calculations at next-to-next-to-leading order (NNLO) accuracy employing several nuclear parton distribution functions. In addition, the Drell$-$Yan production cross section ($σ_\text{DY}$) for dilepton masses above 10 GeV and the ratio of $\mathrm{t\bar{t}}$ to DY cross sections ($R_{\mathrm{t\bar{t}}/\mathrm{DY}}$) are found to be compatible with the NNLO predictions. The observables $σ_{\mathrm{t\bar{t}}}$, $σ_\text{DY}$, and $R_{\mathrm{t\bar{t}}/\mathrm{DY}}$ are measured separately for central and semicentral PbPb collisions to investigate for the first time the dependence of top quark production on the collision impact parameter.

Solar analogs, stars that closely match the fundamental properties of the Sun, provide key benchmarks for testing stellar structure and evolution across different ages and activity levels. Their detailed characterization helps place the Sun in context within the broader population of solar-like stars. This study presents the characterization of six seismic solar analogs observed by the NASA Kepler and K2 missions. Combining asteroseismic constraints from space-based photometry with high-resolution spectroscopy and \textit{Gaia} astrometry, we derived their fundamental parameters and assessed their resemblance to the Sun. Global seismic properties and individual oscillation modes were extracted from the photometric light curves, while atmospheric parameters were obtained from data collected by the HERMES spectrograph at the Mercator telescope. Stellar modeling using seven independent stellar evolution codes yielded consistent masses, radii, and ages. These stars have masses between 0.91 and 1.04~$\mathrm{M}_\odot$, radii between 0.95 and 1.08~$\mathrm{R}_\odot$, and ages from about 1.8 to 9.1~Gyr, with typical systematic uncertainties of $\pm$ 0.02~$\mathrm{M}_\odot$, $\pm$ 0.01~$\mathrm{R}_\odot$, and $\pm$ 0.7~Gyr, respectively. One star, EPIC~206064678, exhibits properties very similar to those of the Sun, with $M = 1.016 \pm 0.033\,\mathrm{M}_\odot$, $R = 0.990 \pm 0.011\,\mathrm{R}_\odot$, and an age of $5.40 \pm 0.12$\,Gyr. It can therefore be considered a close solar twin, although it is slightly older and more metal-rich ($0.25 \pm 0.07$\,dex). Four targets display binarity signatures and all exhibit very low chromospheric activity. This work broadens the sample of well-characterized seismic solar analogs towards a larger sample of metallicities and ages, providing new references for comparative stellar studies and future asteroseismic investigations.

Fluid equations are nonlinear, dissipative, and non-Hamiltonian, which makes their relation to Schrödinger evolution and quantum algorithms nontrivial. We derive an exact Eulerian Cole-Hopf-type reformulation of isothermal compressible Navier-Stokes (NS) flow in Schrödinger-type amplitude variables. To our knowledge, this gives the first exact Cole-Hopf-type Schrödinger-variable reformulation of compressible NS flow. In two dimensions, a Helmholtz decomposition separates the velocity into compressive and vortical potentials, whose logarithmic transforms yield two scalar imaginary-time Schrödinger-type equations with nonlinear self-consistent potentials. We show that the mixed density-compressive amplitude $Ψ_α=ρ^αΘ^{1-2α}$, where $ρ$ is the density, $Θ$ is the compressive amplitude, and $α\neq 0,\,1/2$, satisfies a nonlinear Schrödinger-type equation with a vector-potential-coupled Laplacian. The transformed system is exactly equivalent to compressible NS and is nonlocal only through Helmholtz and Poisson projections. In three dimensions, the density-carrying equation retains the same vector-potential-coupled structure, while the solenoidal sector admits a compressible analogue of Ohkitani's incompressible NS Cole-Hopf formulation. Unlike unitary hydrodynamic Schrödinger-flow representations, the present equations are imaginary-time heat or drift-diffusion equations with self-consistent potentials, but they remain an exact change of variables for compressible NS. A two-dimensional Kelvin-Helmholtz unstable shear-layer calculation verifies the transformed equations against a direct compressible NS simulation. The formulation exposes operator structures that may be useful for reduced flow descriptions, quantum algorithms for operator evolution, and quantum partial differential equation solvers.

The walled Brauer algebras $B_N(m,n)$ govern Schur--Weyl duality for unitary groups $U(N)$ acting on mixed tensor spaces $V_N^{\otimes m}\otimes \overline{V}_N^{\otimes n}$ and play an important role in applications ranging from AdS/CFT to quantum information theory. In the stable regime $N\ge m+n$ the algebra is semisimple and its representation theory is well understood. For $N<m+n$, however, $B_N(m,n)$ becomes non-semisimple. The representation of the algebra on tensor space has a non-trivial kernel and the corresponding quotient algebra is semisimple, with representation dimensions differing from those in the stable regime. We introduce \emph{restricted Bratteli diagrams}, obtained by modifying the standard Bratteli diagrams for $B_N(m,n)$. This construction provides a systematic way to use representation-theoretic data from the stable regime to compute the dimension modifications arising in the non-semisimple regime. In the regime $N=m+n-l$, with $l$ small compared to $m,n$, we show that the restricted diagrams exhibit a stability property and enable an efficient counting of the paths responsible for these dimension corrections. Remarkably, the resulting generating functions are governed by the partition function of an infinite tower of simple harmonic oscillators. We briefly discuss implications for the construction of orthogonal bases of matrix invariants in gauge theory and related applications in quantum information theory.

A high-resolution ($\sim$45000), high signal-to-noise ($>$100) K-band spectral atlas of massive stars is presented. It includes 81 stars consisting of known optical standards, spanning spectral and luminosity subclasses from O2 to O9, and supergiant luminosity and spectral subclasses from O2-B1. The telluric-corrected reduced spectra are publicly available, and are discussed here.

Radio Access Network (RAN) configuration has traditionally required significant manual effort due to indirect causal dependencies between observable Key Performance Indicators (KPIs), and context-dependent characteristics, where the optimal configurations vary with network conditions. Although recent data-driven approaches improve parameter tuning, they remain limited in distinguishing causal direction from statistical correlation and in generalizing across diverse operating contexts. To address these challenges, we propose BLINC (Bayesian Large Language Model (LLM)-Driven Intelligent Network Configuration), an LLM-assisted Bayesian Network framework that integrates telecommunications domain knowledge into causal structure learning. Trained and validated on a private 5G deployment, our method achieves throughput improvement of 63.5% with 19.7% reduction on block error rate over data-only baselines through joint optimization of power control and link adaptation parameters. The framework provides interpretable causal structure, while also quantifying prediction uncertainty. We also demonstrate the ability of the Bayesian Network framework to adapt to different deployment scenarios and propose an incremental Conditional Probability Distribution (CPD) update mechanism with learning rate for continuous model adaptation as network conditions evolve.

Take a compact Sasakian threefold $M$ and consider the associated irreducible $\text{SL}(r,{\mathbb C})$-character variety ${\mathcal R} := \text{Hom}(π_1(M, x_0), \text{SL}(r, {\mathbb C}))^{ir}/ \text{SL}(r, {\mathbb C})$ of $M$, where $\text{Hom}(π_1(M, x_0), \text{SL}(r, {\mathbb C}))^{ir}$ is the space of irreducible homomorphisms. We first construct a natural algebraic $2$-form on $\mathcal R$. Then it is shown that this $2$--form is closed. Finally we show that the restriction of this $2$--form to $\text{Hom}(π_1(M, x_0), \text{SU}(r))^{ir}$ is symplectic.

Entangling operations are at the heart of all approaches to quantum information processing. Parametric gates, in particular, offer a versatile solution to strongly couple off-resonant superconducting qubits with suppressed parasitic crosstalk to spectator qubits due to frequency-selective activation. In this work, we demonstrate a parametric iSWAP gate between two transmon qubits using the recently developed double transmon coupler (DTC). The DTC supports robust internally-defined cancellation point (``off'' state) for static interactions, while simultaneously mediating a fast parametric coupling between data qubits that can be deployed for high-fidelity two-qubit operations. We use robust phase estimation to calibrate non-commuting error terms in the parametric iSWAP gate, and achieve a 99.827% gate fidelity in 40ns without any numerical optimization. The circuit architecture and calibration techniques developed here are extensible to other gate implementations and qubit modalities, paving the way towards resource-efficient quantum information processing.

We develop a calculus for counting pseudoholomorphic disks with boundary in tropical Lagrangians contained in almost toric manifolds, using our previous work with Venugopalan. The results are mostly in dimension four under monotonicity assumptions although in principle the same technique works in any dimension and without monotonicity. The calculus is given as a sum over tropical graphs that interact with the tropical graph of the Lagrangian, generalizing results of Mikhalkin and Nishinou-Siebert for holomorphic spheres in toric varieties, and our previous result with Venugopalan which dealt with disks bounding almost toric moment fibers. The main contribution of this paper is the calculation of several multiplicities of vertices corresponding to disks, such as the holomorphic pant (half of the holomorphic pair of pants) and various univalent vertices occuring at trivalent vertices of the graph of the Lagrangian; a key tool is a Lagrangian isotopy from the Lagrangian pair of pants in the del Pezzo of degree seven to the inverse image of a diagonal, which is a special case of a results of Hind and Evans. We show that every integer eigenvalue of non-maximal modulus for quantum multiplication by the first Chern class is realized by such a sphere.

Optimization on Hadamard manifolds -- the natural Riemannian setting for globally geodesically convex problems -- relies on exponential maps to retract tangent vectors and parallel transport to connect tangent spaces across the manifold. These primitives are often computationally expensive, leading software packages to rely on approximations: first-order retractions and vector transports. However, existing results for optimization on Hadamard manifolds either require exact primitives or lack non-asymptotic rates. We bridge this gap by introducing a proximal bundle method for nonsmooth geodesically convex optimization and establishing the first oracle-complexity bounds that rely only on subgradients and inexact primitives. We obtain sublinear rates for general objectives and optimal linear convergence under sharp function growth.

Superconductors can transport spin without Joule dissipation, yet their coherent coupling to short-wavelength magnons in insulating magnets remains largely unexplored. Here we demonstrate experimentally and theoretically that a conventional diffusive superconductor can enable the conversion of the uniform ferromagnetic-resonance (FMR) mode into perpendicular standing spin waves (PSSWs) in an adjacent ferrimagnetic insulator. In Bi-substituted iron-garnet/Nb bilayers, the microwave transmission develops an additional resonance feature that appears only below the Nb transition temperature and lies close to the uniform FMR peak. A microscopic theory that self-consistently couples the quasiclassical Keldysh--Usadel description of the superconducting condensate to the Landau--Lifshitz--Gilbert dynamics shows that the conversion requires two ingredients: (i) an interfacial spin-transfer torque mediated by spin-polarized triplet Cooper pairs and (ii) a depth-dependent effective field produced by Abrikosov vortices (electromagnetic proximity). The resulting susceptibility reproduces the measured lineshapes and establishes superconductivity as an active control knob for exchange standing-wave modes in magnetic insulators.

Due to hardware-software co-development in embedded systems, continuous integration (CI) builds frequently fail because of complex cross-compilation, board configurations, and toolchain constraints. Although CI build logs contain valuable diagnostic information, they are short-lived and difficult to reuse due to heterogeneous runners, toolchains, and log formats. To address these challenges, we present PhantomRun, a unified abstraction layer and publicly reusable dataset that standardizes the retrieval, storage, and reproduction of CI build logs and metadata. Across 4628 failing CI runs, we reconstructed 91.8% of builds and preserved execution outcomes in 98% of evaluated cases. PhantomRun provides two core capabilities: retrieving the build log of any commit and faithfully re-executing the corresponding build in a controlled environment. By exposing all build artifacts and metadata in a uniform, machine-readable format, PhantomRun enables reproducible and longitudinal studies of CI failures. An empirical evaluation shows that reproduced builds closely match their originals, typically differing only in timestamps or minor nondeterministic reordering, demonstrating the feasibility of large-scale historical CI reconstruction.

We develop a framework to describe how quantum coherences between distinct charge sectors evolve under generic charge-conserving dynamics. Our framework captures the nonperturbative interactions between quantum coherences and hydrodynamic large deviations -- i.e., rare ``voids'' of low charge entropy. Conditional on surviving, the quantum coherence and its surrounding void form a collective polaron-like object. In one dimension, even at infinite temperature, we show that the lifetime of coherences is parametrically enhanced because they bind to voids. We use our framework to address two fundamental questions about generic quantum dynamics with a conserved charge. First, we argue that gapped Ruelle-Pollicott resonances are absent in the weak-noise limit, even in sectors of operator space that contain no hydrodynamic slow modes: instead, the spectral gap in all sectors vanishes nonperturbatively in the noise strength. Second, we compute the spacetime asymptotics of the dynamical single-particle Green's function, both in the weak-noise regime and in the absence of noise. In the noiseless case, we find that the void-coherence polaron undergoes subdiffusion, with an exponent we calculate. We support our general arguments with a microscopic derivation for random charge-conserving circuits, as well as numerical evidence from tensor-network simulations.

Coded caching is a technique that leverages locally cached contents at the end users to reduce the network's peak-time communication load. Coded caching has been shown to achieve significant performance gains with a centralized placement orchestrated by the server and is thus considered a promising technique to boost performance in future networks by effectively trading off bandwidth for storage. To tackle issues caused by the synchronized placement, previous works focused on decentralized placement and found the exact worst-case load with uncoded placement. In this paper, we focus on a decentralized coded caching system with random linear coding placement, and investigate the fundamental limits of a linear coding placement where each user independently and uniformly caches random linear coding symbols of a single file. We propose achievable and converse bounds on the worst-case load, which are shown to meet under certain conditions.

We study dissipative phase transitions in a system of two coupled fully-connected quantum Ising models interacting with an environment. The dynamics is governed by a Lindblad master equation combining coherent unitary evolution and incoherent dissipative processes, where the unitary part is described within a self-consistent mean-field framework effectively acting on the local Hilbert space of two coupled spins at each site. We analyze two fundamentally different classes of dissipators. In the first case, the jump operators are defined in the instantaneous eigenbasis of the mean-field Hamiltonian and satisfy a detailed-balance condition. In this setting, the relaxation dynamics depends strongly on the quench protocol: a parametric quench of the Hamiltonian leads to conventional relaxation, whereas a temperature quench gives rise to a dynamical phase transition characterized by nonanalytic behavior in time. Yet, in both cases, the system relaxes toward a steady state determined solely by the post-quench parameters and the bath temperature, which closely resembles a thermal Gibbs state of the mean-field Hamiltonian. As a result, the dissipative phase transition occurs at a critical point consistent with the corresponding equilibrium transition. In contrast, when the dissipators are realized via local spin raising and lowering operators, the steady state is genuinely nonequilibrium, leading to a significantly richer phase diagram. In particular, for sufficiently strong system-bath coupling, we observe a reentrant phase featuring a symmetry-broken region bounded by two continuous dissipative phase transitions. Our results evidence how the structure of dissipative processes controls the emergence of equilibrium-like versus genuinely nonequilibrium critical behavior in open quantum systems.

Measuring the properties of cold molecular gas available for intense star formation in galaxy protoclusters at $z>2$ is a crucial step in understanding large scale structure formation. We present ALMA observations of CO(3$-$2) in five dusty star-forming galaxies within $\sim0.5-4$ cMpc of the core of the Spiderweb protocluster at $z=2.16$ to measure the molecular gas mass and kinematics in the most starbursting members of the protocluster. All five galaxies exhibit evidence for disturbed kinematics including non-Gaussian CO line profiles, irregular spatial morphology, and strong residuals when fitting the galaxies with a classical disk model. This could be indicative of an elevated merger rate in the outskirts of the mature Spiderweb protocluster, as all of the galaxies in our sample have multiple companions detected in H$α$. Both the gas fractions and the gas depletion timescales of the galaxies are similar to field relations at cosmic noon, indicative of the fact that their prodigious star formation rates are compensated by similarly high gas masses. The most massive galaxies, as well as all of the galaxies identified as X-ray AGN in previous works, have gas fractions $<30$%, compared to the sample average of 49%, indicating declining availability of gas for star formation. Finally, we find that the gas fractions and specific star formation rates decline with distance from the Spiderweb Galaxy, supporting the reversal of the SFR density--radius relation in high-redshift protoclusters.

Graph neural network architectures are advantageous for predicting core-electron binding energies which depend on local bond environment effects, as the number of message passing layers defines the topological (bond) radius of the model's receptive field. This provides an interpretable connection between the model's architecture and the definition of locality in the considered environment. Here we present a graph neural network model for predicting carbon 1s core-electron binding energies in organic molecules. The model is trained with multiconfiguration pair-density functional theory on 8637 carbon atoms in 2116 molecules with 4-16 atoms and evaluated against 570 experimental values in 113 different molecules containing 3-45 atoms. Previous work benchmarked a mean absolute error of 0.27 eV to experiment for the training data level of theory [J. Phys. Chem. A 2025, 129, 36, 8419-8431] and the present model demonstrates an experimental evaluation error of 0.33 eV with good size transferability to larger systems. By examining the effect of the number of message passing layers on the performance, we show that two chemically informed node features, the atomic binding energy and environment electronegativity, encode molecule-specific information when normalized across the graph and capture beyond nearest-neighbor environment effects outside the receptive field. A case study on the 45 atom avobenzone tautomers demonstrates the model's ability for instant and precise analysis of complex molecules. Finally, the model's E(3)-equivariance is shown to out-perform an invariant model on non-equilibrium geometries from a methanol C-O bond stretch. The software and data are provided by the open-source AugerNet package at https://doi.org/10.5281/zenodo.19689244.

Effective suppression of surface-related multiples is essential to prevent imaging artifacts and erroneous structural interpretations. While conventional approaches rely on accurate priors or subsurface model knowledge, and supervised learning methods require labeled data that are impractical to obtain for real seismic data. To overcome these limitations, a recently proposed self-supervised learning (SSL) framework integrates multi-dimensional convolution (MDC) for multiple generation with a two-stage training strategy, eliminating the need for both prior knowledge and labeled data. However, their approach requires manual selection of a scaling factor to match the amplitudes between the MDC-generated multiples and the true multiples, thus introducing subjectivity and limiting its practical applicability. In this study, we propose an adaptive SSL method that treats the scaling factor as a learnable parameter, jointly optimized with the network weights in a unified single-stage training pipeline. This dynamic scaling implicitly introduces amplitude diversity into the training data, acting as an implicit regularizer that improves the network's robustness to amplitude variations of surface-related multiples. We further design a composite loss function with homoscedastic uncertainty-based adaptive weighting, which automatically balances the contributions of multiple loss terms without manual tuning. Synthetic and field data examples demonstrate that our method robustly and effectively suppresses surface-related multiples while preserving primary reflections, with migration results confirming improved subsurface imaging quality.

We provide a holographic bulk realization of Carrollian free-field structures arising in three-dimensional asymptotically flat (higher-spin) gravity. We construct a class of boundary conditions that generalizes the diagonal gauge of Anti-de Sitter to flat spacetimes. We show that the associated asymptotic symmetries decompose into genuine physical transformations and pure gauge redundancies, the latter being generated by Carrollian screening charges. This structure leads to a bulk-born realization of Carrollian Miura transformations, expressing physical observables in terms of celestial free scalars. Our results establish a concrete link between flat space (higher-spin) gravity and a Carrollian Coulomb gas description, thereby providing a promising route toward the quantization of flat holography.

We find the perturbative potential on the 3d $\mathcal{N}\!=\!2$ Coulomb branch arising from a chiral 4d $\mathcal{N}\!=\!1$ gauge theory on $\mathbb{R}^3 \times S^1$, zeta-regularizing the D-term couplings generated by the Kaluza-Klein modes. This fills a significant gap in the literature on circle-compactified SUSY gauge theories. Unlike earlier indirect approaches to the circle reduction of chiral theories, our formula provides a global view of the Coulomb branch, necessary for capturing holonomy saddles and for systematic implementation. The zero locus of the potential identifies perturbative SUSY vacua, and we show how data-analysis techniques (such as RANSAC hyperplane detection) numerically extract the structure of the moduli space when this locus is extended. Our formula yields new results even in abelian theories, and offers a new perspective on several earlier observations in the context of the Cardy limit of the superconformal index. In particular, circle reductions (of interest in the SCFT/VOA correspondence) found earlier from limits of the index can now be reproduced on $\mathbb{R}^3 \times S^1$. An appendix shows how our 3d $\mathcal{N}\!=\!2$ potential is related to a function arising in the Cardy limit of the index analogously to how the 4d $\mathcal{N}\!=\!2$ prepotential arises in a limit of the Nekrasov partition function.

We present a detailed study of radio-detected dwarf galaxies (with stellar masses less than 3 billion solar masses) to characterize extreme star formation and search for (variable) radio AGNs. Our sample comes from Reines et al. (2020) (arXiv:1909.04670), who used the Karl G. Jansky Very Large Array (VLA) with 0.25 arcsecond resolution to observe 111 dwarf galaxies with lower-resolution (5 arcsecond) detections in the Faint Images of the Radio Sky at Twenty Centimeters (FIRST) survey. While that work identified and focused on 13 compact radio AGN candidates in dwarf galaxies, here we focus on 16 compact radio sources consistent with star formation in dwarf galaxies. We find that these objects are dominated by thermal HII regions with ages less than 10 Myr, and the most extreme sources have ionizing luminosities requiring the equivalent of around 10,000 to 100,000 O-type stars. We also investigate the dwarf galaxies detected in FIRST but not detected in the high-resolution follow-up observations. Using the infrared-radio correlation parameter, we identify eight sources consistent with radio-excess AGNs. Five of these objects plus another 15 dwarf galaxies have no corresponding detections in the VLA Sky Survey (VLASS) indicating variability between the FIRST and VLASS observations. The FIRST radio luminosities of these sources are significantly higher than expected for supernova-related emission, suggesting the radio variability is likely associated with AGNs. Together, these results provide new context for the presence of compact star formation and massive black holes in dwarf galaxies, and highlight the utility of radio variability and multi-resolution data for identifying the dominant power sources in low-mass systems.

We report the confirmation and characterization of four transiting giant planets orbiting early-M dwarfs discovered by the Searching for Giant Exoplanets around M-dwarf Stars (GEMS) survey: TOI-7189 b, TOI-7265B b, TOI-7393 b, and TOI-7394B b. Joint modeling of TESS and ground-based photometry with precision radial velocities from the Habitable-zone Planet Finder and NEID spectrographs yields self-consistent orbital and physical parameters for all systems. The planets have short orbital periods ($P = 1.25-4.17$ days), masses spanning from $0.5\,M_{\rm J}$ to $2.1\,M_{\rm J}$, and radii comparable to Jupiter ($0.95\,R_{\rm J} < R_p < 1.02\,R_{\rm J}$). TOI-7189 b ($0.50\,M_{\rm J}$), TOI-7265B b ($0.71\,M_{\rm J}$), and TOI-7393 b ($0.61\,M_{\rm J}$) are Saturn-like in mass and density, whereas TOI-7394B b is a dense super-Jupiter ($2.10\,M_{\rm J}$, $ρ_p \approx 2.4$ g cm$^{-3}$) on a 1.25-day orbit. All hosts are early-M dwarfs with a narrow range of stellar properties, enabling a controlled comparison of giant-planet outcomes around low-mass stars. Three systems orbit super-solar metallicity stars, while TOI-7393 ($\mathrm{[Fe/H]} = -0.35 \pm 0.16$) is the most metal-poor GEMS host identified to date, and exhibits kinematics consistent with the thin/thick-disk transition, suggestive of an older stellar population. Together, these systems reveal substantial diversity in the masses and bulk properties of short-period giant planets orbiting early-M dwarfs, demonstrating that markedly different planetary outcomes can arise around stars with otherwise similar fundamental properties.

We prove a Positivstellensatz for operator-valued noncommutative polynomials that are positive on matrix convex sets. Specifically, let $p$ be an operator-valued polynomial in $B(H)\otimes C<x>$ of degree at most $2d+1$, where $H$ is separable and infinite-dimensional. Let $L(x)=I+\sum_{j=1}^{g} A_j x_j$ be a monic linear operator pencil, and let $D_L=\{X: L(X) \geq 0\}$ be the associated matrix convex set. We show that $p$ is positive on $D_L$ if and only if $p=r^*r+q^*π(L)q$, where $q$ and $r$ have degree at most $d$, and $π$ is a unital completely positive map on the operator system generated by the coefficients of $L$. The proof combines a Hahn--Banach separation argument with a tailored GNS construction. The main challenge is that the separation occurs in the product ultraweak topology, so boundedness of the resulting GNS operators is not automatic. We first handle bounded matrix convex sets, using closedness of the cone of weighted squares in the product ultraweak topology as the key technical input, and then pass to the general unbounded case by an approximation argument. Finally, we apply this convex Positivstellensatz to prove an operator-valued noncommutative Fejer--Riesz theorem on free products of finite abelian groups. The key additional ingredients are the universal $*$-algebra povm(n) associated with POVMs, a perfect Positivstellensatz for povm(n), and Boca's theorem on free products of completely positive maps. As a consequence, every positive operator-valued trigonometric polynomial on a free product of finite abelian groups admits a sum-of-squares factorization with explicit complexity bounds.

We study a generalisation of the set-theoretic Yang-Baxter equation and investigate the connection between its solutions and matrix refactorisation problems. We refer to such solutions as scalene Yang-Baxter maps. Moreover, we construct scalene Yang-Baxter maps associated with integrable equations of KdV and NLS type.