Rings and Algebras
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Sunday 8 November
For each integer $n\ge5$, we construct a hereditary artinian ring of finite representation type with Coxeter diagram $I_2(n)$. Both of its diagonal division rings are two-generated over the field $\mathbb{Q}(\zeta)$, where $\zeta$ is a primitive root of unity of order $n-2$. The construction is based on a chain of quantized Weyl relations and a shift automorphism. We give one-sided bases for the corresponding division ring extensions and compute the right dimension sequence of the bimodule. The resulting ring has exactly $n$ isomorphism classes of indecomposable finitely generated right modules.
Cryptography and Security
Delegating Computation with Approximate Arithmetic: Interactive Proofs for Wizards
We study interactive proofs for arithmetic circuits over the complex numbers $\mathbb{C}$ in which both the prover and the verifier compute approximately (unlike muggles, wizards can use magic, i.e., floating-point numbers). The goal is to prove that a claimed output $y$ is close to the exact output $C(x)$: for thresholds $0<\Delta_0<\Delta$, the honest prover is accepted whenever $|y-C(x)|\le\Delta_0$, and every prover is rejected with high probability whenever $|y-C(x)|>\Delta$. Existing proof systems for real-valued data verify rounded finite-field computations. Bitan et al.~[EUROCRYPT~2026] gave an approximate sum-check protocol. We extend the protocol of Goldwasser, Kalai, and Rothblum (GKR) [STOC~2008] to public-coin interactive proofs for logspace-uniform layered circuits over $\mathbb{C}$ of size $S$, depth $L$, and $n$ inputs, under public bounds on gate values and wiring coefficients. At constant soundness error, the verifier runs in time $(n+1)\mathrm{poly}(L,\log S,b)$, the communication is $\mathrm{poly}(L,\log S,b)$, and the prover runs in time $\mathrm{poly}(S,b)$, where $b$ is the bit precision. To the best of our knowledge, these are the first doubly efficient interactive proofs in which both parties compute approximately. GKR reduces two claims about a layer to one on a random line, which leaves the unit circle, where our error bounds hold. We instead label gates by complex roots of unity and combine the two claims with a random root of unity as coefficient. We choose the error thresholds backward from the input layer, so that a large output error survives the verifier's rounding at every step. The honest prover needs only $O(S \log S)$ complex operations. We also obtain zero-knowledge proofs from one-way functions, one-round arguments from private information retrieval, and interactive PCPs whose proof length is polynomial in the witness length.
Read the full abstract
High Energy Astrophysical Phenomena
Unveiling a Concentration of Hard X-Ray Point Sources in the Nuclear Stellar Disk with XMM-Newton
The Galactic Center is home to the largest known concentration of X-ray sources ever detected in our Galaxy, including a sizable population of accreting compact objects. Utilizing data from the XMM-Newton Galactic Plane Heritage Survey of the central $20^{\circ}\times2^{\circ}$ of the Milky Way, we investigate the varying source populations in the inner regions of the Galaxy. We identify for the first time a concentration of extremely hard X-ray sources within $|l|<1.3^{\circ}$ and $|b|<0.3^{\circ}$, a region that overlaps the nuclear stellar disk (NSD). We determine that this hardness is intrinsic to the sources and not a manifestation of high line-of-sight column density ($N_{\rm H}$). Unlike the "soft" X-ray sources, this population does not trace the stellar mass distribution, indicating an overabundance compared to the underlying stellar population. We simulate spectra for a variety of source types to model their X-ray hardness and compare their distributions to those we observed in our survey data. We find that a significant subset of the Galactic Center X-ray population may comprise intermediate polars (IPs) with massive white dwarfs, possibly formed through dynamical processes. Our study provides novel observational insights into the distribution of stellar remnants and other X-ray sources in the vicinity of our supermassive black hole, Sgr A*.
Read the full abstract
Mesoscale and Nanoscale Physics
Coherent Control of High-Order Lightwave Currents in Gate-Tunable Bilayer Graphene
The sub-cycle acceleration of charge carriers in an electronic circuit by tailored terahertz (THz) electric-field waveforms, rather than conventional pulsed bias voltages, opens a transformative frontier for lightwave electronics operating far beyond the scaling limits of gigahertz electronic technologies. Yet gate-tunable coherent control of multi-order and non-perturbative currents has remained elusive. Here, we demonstrate lightwave control over the amplitude, phase flips, and zero-crossings of THz field-driven high-order currents by utilizing phase-locked pairs of single-cycle THz fields in a dual-gated bilayer graphene (BLG) device. By leveraging the in-situ tuning of both bandgap and charge doping, we reveal extreme THz photocurrent rectification and sub-THz-cycle current generation that are governed by a strong field-driven crossover from interband multi-photon coherence to intraband lightwave acceleration, whose competition provides an exceptional sensitivity to the THz-pulse pair drive. Our interferometric nonlinear THz photocurrent spectroscopy of sub-cycle electronic excitations accesses and controls non-perturbative, multi-quantum photocurrents, which manifest as an extreme-order THz frequency comb. Quantum-kinetic simulations map these lightwave-controlled electron trajectories, revealing that sub-THz-cycle ballistic transport underpins gate- and THz-field-driven functionality enabled by electrostatic doping and bandgap tuning within the atomically thin device architecture. The demonstrated on-demand coherent control of nonperturbative photocurrent transport regimes can be extended to a broad range of atomically thin device architectures, opening new avenues for gate- and field-programmable logic, memory, and quantum simulators in which electronic pathways are steered on sub-cycle, petahertz timescales and atomic length scales.
Read the full abstract
Analysis of PDEs
A Talenti-type result for the Pompeiu-Schiffer Conjecture
Schiffer's conjecture asserts that the ball is the only bounded simply connected domain carrying a Neumann eigenfunction of the Laplacian that is constant on the boundary; it is equivalent to the Pompeiu problem. In August 2026 it was disproved in the plane by computer-assisted constructions. We revisit the shape-optimization approach of~\cite{Se24} in the light of these counterexamples. We first show that the vanishing of the material derivative on the boundary, assumed in~\cite{Se24}, never holds, and we compute the correct optimality condition, involving an adjoint state, for the associated functional $J_2$. We prove that every planar Schiffer domain satisfies $\max|u| > \sqrt{2}|c|.$ Our main result is a Talenti-type theorem: if $u < c$ in $\Omega$ , then $kR^* \geq j_{1,1},$ where $R^*$ is the radius of the disc of equal area, and equality holds only for the disc. Finally, using Siegel's transcendence theorem, we prove the infinitesimal rigidity of the disc for the Berenstein problem.
Read the full abstract
Differential Geometry
Pluriclosed flow on Kähler manifolds
We prove the maximal existence conjecture for pluriclosed flow on Kähler backgrounds. This leads to a priori generalizations of the Aubin-Yau and Calabi-Yau theorems, and resolves the question of connectedness of generalized Kähler structures on Calabi-Yau backgrounds.
Machine Learning
Daily Emergency Call Volume Forecasting using Univariate and Multivariate TimeGPT
This study investigates the potential of TimeGPT, a pretrained foundation model for timeseries, for forecasting daily emergency call volumes. Real-world emergency call data, comprising 2,101 daily observations, are used to evaluate both univariate and multivariate forecasting settings under one-step-ahead and seven-day multi-step-ahead scenarios. The multivariate forecasting incorporates temporal, meteorological, demographic, and socioeconomic variables. A chronological training, validation, and test split is adopted to prevent information leakage, while hyperparameters are selected using the validation set through grid search and early stopping when applicable. Forecasting performance is evaluated using multiple error metrics. Its forecasting performance is compared with several established forecasting approaches, including Seasonal Autoregressive Integrated Moving Average with Exogenous Regressors (SARIMAX), Poisson regression (PR), Negative Binomial regression (NBR), Extreme Gradient Boosting (XGBoost), Multilayer Perceptron (MLP), and Long Short-Term Memory (LSTM). The results show that SARIMAX achieves the best overall performance in the univariate setting. In the multivariate setting, NBR and PR provide best performance. These findings indicate that the large-scale pretraining of TimeGPT do not necessarily translate into superior forecasting accuracy for emergency call volumes. Frugal methods trained solely on the available data remain better for this application.
Read the full abstract
High Energy Physics (Phenomenology)
The $\phi$-$\eta$ Puzzle and the Five-Layer Effective-Temperature Hierarchy in $pp$ Collisions at 2.76 and 7 TeV
Sequential kinetic freeze-out in high-energy collisions probes the dynamics governing the final stages of strongly interacting matter. Inspired by the recently established quadruple freeze-out scenario, we present a thermodynamic analysis of transverse momentum spectra for a set of identified particles produced in proton-proton collisions at $\sqrt{s} = 2.76$ and $7$ TeV. Using a modified Tsallis-Pareto function, we extract the effective temperature and the non-extensivity parameter for each species. We observe a robust five-layer hierarchy in the effective temperature. At both energies, the layers are: non-strange hadrons, singly-strange hadrons, multi-strange hadrons, light nuclei, and the $\phi(1020)$ meson. At $7$~TeV, the $\eta$ meson joins the $\phi(1020)$ meson in the highest-temperature layer. At both energies, each pair of neighboring layers is separated by more than $ 3\sigma$, and the separation at the top exceeds $7\sigma$. The grouping persists at both energies, while the temperatures shift upward at 7 TeV. The observed pattern shows that the decoupling sequence is driven by an interplay of strangeness, cross-sections, and the OZI rule rather than by particle mass alone. Notably, the $\phi(1020)$ meson has a mass close to the proton, but a significantly higher temperature due to its OZI-suppressed coupling to the hadronic medium. The unexpected coincidence of $\eta$ and $\phi(1020)$ at 7 TeV presents a puzzle that may point to a common early production mechanism or in-medium effects, which are not yet understood. The five-layer hierarchy provides new constraints on theoretical models of hadronization. It suggests that the sequential decoupling of particles from the evolving medium is even more sensitive to their internal structure and interaction than previously assumed, with implications for the QGP-like matter formation in small systems.
Read the full abstract
Rings and Algebras
Lift Height and the Structure of Classification Failure
We introduce a framework for measuring the depth of classification failure relative to a chosen invariant. A shadow map records visible structure while suppressing finer information; lift packages recover this information in successive layers, and lift height measures the depth through which classification ambiguity persists. Canonical reduction removes redundant stages and yields an essential notion of lift height. For nilpotent groups, the lower central series provides a natural commutator lift tower, with lift height controlled by nilpotency class and decreasing under refinement of the shadow. The associated graded structure gives a Lie-theoretic interpretation of the successive commutator layers. The main examples show that classification depth is genuinely unbounded. Over any field of characteristic different from two, and for every positive integer, there exist nilpotent Lie algebras with the same fixed four-dimensional abelianisation whose commutator lift packages agree through the prescribed number of layers but differ at the next layer. Moreover, such separation can occur at arbitrarily large degree even when the complete underlying lower-central graded vector space is fixed. Thus neither a fixed shadow nor the dimensions, or even the underlying vector-space structure, of the lower-central graded object impose a uniform bound on the depth at which new commutator information may appear.
Read the full abstract
Geophysics
Estimating Permeability of Carbonate Rocks from NMR T2 Distribution and Formation Factor Using a Physics-Based Model
Nuclear magnetic resonance (NMR) logging provides a nondestructive method for characterizing pore structure and fluid distribution in reservoir rocks. Conventional permeability models, such as the Schlumberger-Doll Research (SDR) and Timur-Coates (TC) equations, have been widely used to estimate permeability from NMR T2 relaxation time distributions. However, these empirical models require calibration against core-measured permeability, and their fitted coefficients often vary substantially across lithologies and pore systems. In this study, we formulate and evaluate an NMR-informed implementation of a physics-based permeability relationship derived from critical path analysis (CPA) by linking the characteristic pore length scale to the dominant NMR relaxation time. We estimate permeability (k) using NMR T2 distributions and electrical formation factor (F) for 14 carbonate rock samples collected from the Morro do Chaves Formation in the Sergipe-Alagoas Basin, Brazil. We compare the predictive performance of this formulation with that of calibrated SDR and TC models. The proposed CPA-NMR formulation achieves a root mean squared logarithmic error (RMSLE) of 0.56 without calibration against measured permeability from the investigated formation. These findings indicate the potential of combining NMR relaxation measurements with electrical formation factor for physics-based permeability estimation. Further evaluation using independent datasets is needed to assess applicability across formations and lithologies.
Read the full abstract
Machine Learning
Building heterogeneous multistep workflows in automated microscopy: reward design and knowledge transfer for autonomous structure property discovery
Autonomous microscopy can turn fast acquisition into scientific progress when it decides not only where to measure, but also what information to acquire, at what scan rate, and when to recalibrate the instrument. We develop a budget constrained framework for such multistep structure-property workflows. It combines explicit scientific rewards, action costs, instrument observation models, transfer of pre experiment knowledge into the online campaign, and evaluation kept independent of the controller. The transferred knowledge includes predictor accuracy and computational cost, empirical learner maturity functions that weight model-derived rewards, and drift characteristics established before controller training. Using a real electron energy-loss spectroscopy dataset, we run two complementary campaigns. In the first, the controller chooses between imaging and spectroscopy, image size, and scan speed. In the second, the instrument drifts and the controller also decides when to recalibrate. The campaigns show that response discovery and structure-property learning favor different surrogate models and acquisition strategies. They also show that simple acquisition rules form strong, practical baselines. Independent error analysis reveals where model uncertainty is and is not a reliable learning signal. In the drift campaign, explicit registration greatly improves positional accuracy, and the framework measures what this costs in spectroscopy throughput. Together, these results provide a reproducible way to design, compare, and audit autonomous experiments, and they define what microscopes must provide for deployment: independently executable actions, characterized scan and detector responses, timestamped registration telemetry, and complete action records.
Read the full abstract
Optimization and Control
The Canonical Family of Reduced Row Pivot Echelon Forms and the Induced Lexicographic Canonical Algorithm for Linear Programming
A complete canonical family of uniquely normalized, row-equivalent representatives of an $m\times n$ matrix is introduced and termed the family of Reduced Row Pivot Echelon Forms (RRPEFs). All RRPEFs represent the same row space, with the standard reduced row echelon form, renamed the Upper RREF (URREF), and the Lower RREF (LRREF) as its extreme members. A reformulation of Noether's Complementary Subspace Basis Theorem yields normalized bases of the row space that coincide with the nonzero rows of the corresponding RRPEFs. Constructively, each RRPEF is obtained by multi-index Gauss–Jordan reduction of a nonsingular transition submatrix. Factoring the admissible pivot-index sets by equality of their associated RRPEFs yields a quotient set in canonical bijection with the set of distinct RRPEFs. This bijection transfers the strict lexicographic order to the distinct RRPEFs and induces a finite chain from the URREF to the LRREF. The resulting structure gives rise to the Induced Lexicographic Canonical Algorithm (ILCA). Lexicographic confinement and singular-transition skipping restrict the candidate domain, whereas RRPEF reuse and scalar-equivalent transition bypassing avoid redundant Gauss–Jordan reductions while retaining the corresponding basic-solution tableaux. Strict lexicographic progression prevents cycling, while the traversal proceeds independently of objective improvement under degeneracy-induced stalling. ILCA terminates when a feasible tableau satisfies the standard optimality conditions and otherwise completes the finite traversal, including for Tucker- and Cunningham-type totally degenerate configurations.
Read the full abstract
Computer Vision
Rethinking Geometric Depth in Monocular 3D Object Detection: A Projection-Consistent Reformulation
Monocular 3D detection (Mono3D) hinges on recovering depth from a single image, for which the geometric formulation $z \approx f \cdot H / h_{2D}$ is widely adopted due to its simplicity and direct grounding in projective geometry. We show that this formulation harbors a systematic, projection-induced bias: substituting the 2D bounding-box height $h_{2D}$ recovers the depth of an object's visible front surface rather than its center, causing size-dependent underestimation that persists even under ground-truth inputs. We address this front-surface bias with MonoPCA, a unified framework built on a projection-consistent reformulation. At its core, PCDepth retains the measurable $h_{2D}$ and closes the gap to center depth through an orientation- and size-aware correction derived from perspective geometry, eliminating the bias without introducing ill-posed regression targets. Because PCDepth couples depth to predicted object size and orientation, attribute perception becomes essential. We therefore introduce GeoAlign, a teacher-student framework that transfers geometric representations from Vision Foundation Models into a lightweight student, strengthening attribute perception at no inference cost. Across four challenging benchmarks, MonoPCA achieves SoTA performance, including a +1.44% AP$_{3D}$ gain on the KITTI leaderboard and +2.80% on Omni3D$_{OUT}$. Code is available at https://github.com/alanzhangcs/MonoPCA.
Read the full abstract
High Energy Physics (Phenomenology)
A Poisson-gas approach to medium-induced gluon radiation
The precision era of jet-quenching phenomenology requires a quantitative control of medium-induced gluon radiation. We consider the BDMPS-Z spectrum in the soft-gluon limit for a Yukawa (Gyulassy-Wang) potential. Writing the Yukawa dipole cross section as a Laplace superposition of Gaussians, we express the fully resummed spectrum as an average over Poisson-distributed configurations of Gaussian scattering centers, each of which is solved in closed form. This yields a master formula for the spectrum both differential in the transverse momentum $k_\perp$ and integrated with a cut on $k_\perp$, valid for static and expanding media with arbitrary profiles. Its evaluation by Monte Carlo sampling of the configurations is efficient and trivially parallelizable. The method reproduces the first order in opacity and published results without kinematic cuts, while with a cut we find sizeable relative deviations from previous results. We argue that the same construction extends to the more complex building blocks of an in-medium parton cascade, such as finite energy fraction and emissions off color antennas, and provides a natural starting point for analytical approximations. The code is publicly available.
Read the full abstract
Robotics
ScanDreamer: A Map-Grounded World Model for Quadrotor Navigation
Quadrotors inspecting buildings and tunnels must navigate unknown clutter with a forward-facing depth camera whose limited field of view constrains scene understanding and collision-avoiding navigation. Recent world-model agents learn navigation through imagined experience (e.g., AirDreamer), but remain data-intensive and often lack explicit geometric memory and awareness of what the camera can observe. We propose ScanDreamer, a map-grounded world model that constructs an online 2D scan-memory map from depth images and vehicle pose. Our world model, actor, and critic are conditioned on this map, which also defines a geodesic progress reward. Moreover, a forward-cone penalty discourages horizontal motion outside the camera's viewing direction. During imagination, ScanDreamer propagates the quadrotor pose over a frozen map and derives progress and wall-contact signals from the mapped geometry. We train ScanDreamer for 0.8 million environment steps (5.5 h on one GPU) and evaluate it in NVIDIA Isaac Sim over 1560 episodes spanning unseen layouts and parameter ranges. Across three seeds (mean $\pm$ SD), ScanDreamer achieves 96.3 $\pm$ 0.3% success with 3.6 $\pm$ 0.3% collision, compared with 59.6 $\pm$ 3.0% success and 28.8 $\pm$ 2.8% collision for AirDreamer at the same online-interaction budget. Removing the forward-cone penalty increases collisions to 9.9 $\pm$ 3.1%. These results show that combining explicit map memory, map-grounded imagination, and perception-aware motion substantially improves navigation performance and data efficiency. The source code is available at https://github.com/WANGSHUAIzzzz/ScanDreamer.
Read the full abstract
Artificial Intelligence
Weight-Space Programming of Policies and Their Verification via Model Checking
Large language models (LLMs) can write the weights of small neural networks directly, without gradient descent. This weight-space programming creates policies for sequential decision-making from a task description alone, but nothing guarantees that a written policy behaves as intended. We combine weight-space programming with probabilistic model checking. Given an environment modeled as a Markov decision process and a requirement in probabilistic computation tree logic (PCTL), an LLM writes all weights of a policy network, optionally repairs them on a dataset of optimal actions, and receives model-checking feedback until the requirement holds or a round limit is reached. The model checker Storm computes, rather than estimates from simulations, the probability that the policy satisfies the requirement, so every returned policy is verified. With two LLMs on three environments, the created policy of each of the 18 runs attains the optimal probability, before any dataset or model-checking feedback. Four of six policies programmed for two transporter jobs also complete three jobs with probability 1, so they generalize to a task that the LLM never saw. The pipeline also programs a policy that satisfies a complex safety requirement expressed with the PCTL until operator in all six runs. In one of them, dataset repair raised the accuracy to 98.4% but broke the requirement, and one round of model-checking feedback restored it.
Read the full abstract
Mesoscale and Nanoscale Physics
Ballistic transport through 3D flat bands in rhombohedral graphene in an in-plane magnetic field
We explore how an in-plane magnetic field reshapes the transmitting channels of a finite slab of rhombohedral multilayer graphene and redistributes the current across its thickness. The two-terminal conductance counts these channels, while the voltage at each surface reflects the relative spectral weights of right- and left-moving states on the contacted layer. In sufficiently thick slabs or at sufficiently strong fields, the nearly flat bulk zeroth Landau level gives rise to a three-dimensional Hall effect with a plateau in the four-terminal Hall resistance.
Read the full abstract
Computer Vision
Martine: Benchmarking Multi-View 3D Surface Reconstruction Across Viewpoint Coverage, Resolution, and Lighting
Controlled acquisition setups are widely used for high-accuracy multi-view 3D surface reconstruction, as they allow sensor configuration, image resolution, viewpoint distribution, and illumination to be specified. Yet benchmarking remains difficult, since reconstruction accuracy depends jointly on the reconstruction paradigm, object properties, and acquisition regime. Existing benchmarks often restrict viewpoint coverage, resolution, or illumination, which can bias comparisons towards methods suited to the chosen setting. Others lack accurate 3D ground truth or quantified image-to-geometry alignment, limiting their use for metric surface evaluation. We introduce Martine, a benchmark for controlled 3D object surface reconstruction. Martine comprises 28 objects, including transparent materials and fine structures, captured at 9568 x 6376 pixels from 84 views over a spherical cap under both single-light and multi-light regimes. It provides object-independent camera calibration, accurate 3D ground-truth scans, calibration and image-to-geometry accuracy quantified on test images, and 3D property masks for analysing errors by material properties, visibility, curvature and fine structures. Experiments across photogrammetry, classical multi-view stereo, neural implicit surfaces, Gaussian-splatting-based reconstruction, and multi-view photometric stereo show that Martine enables regime-aware evaluation with explicit control over reference-data accuracy. The benchmark resolves sub-millimetric reconstruction differences, with a best observed Chamfer distance of 0.058 mm over the evaluated visible surface.
Read the full abstract
Robotics
Kinora: A Versioned Workflow for Text-to-Motion Generation, Robot Retargeting, and Interactive Validation
Motion generation tools produce candidate trajectories, but interactive robot authoring also requires recoverable execution, traceable edits, explicit geometry, and model-specific validation. We present Kinora, a system that connects text-to-motion generation, a motion library, local agent tools, robot retargeting, and trajectory refinement through versioned artifacts. The current release uses an independently specified 52-joint authoring rig and procedural skeleton previews, separating the distributed human geometry from externally installed generation models and robot assets. Human-motion edits preserve frames outside a selected interval, while robot edits create immutable candidates whose reports govern acceptance and export. We evaluate server 0.5.1 and Gadget 0.15.2 with 156 core regression test executions, 49 Gadget executions, four CUDA generation/editing cases, a 30-frame G1 retargeting check, and a 3,600-frame synthetic G1 fixture paired with a suspended negative control. All four CUDA cases passed their numerical gates; the turning splice nevertheless retained a 14.3 cm support-foot drift proxy warning. The G1 positive fixture completed bounded simulation replay, whereas the suspended control was quarantined before replay. These results support selected integration and preservation contracts. They do not establish semantic-quality improvements, population-level feasibility, or hardware safety.
Read the full abstract
High Energy Physics (Phenomenology)
Threshold resummation for top-quark pair production in association with a jet at the LHC
We study threshold resummation for the invariant-mass distribution of a top-quark pair and a jet at the LHC. We perform the resummation at next-to-next-to-leading logarithmic accuracy in soft-collinear effective theory, with the logarithms of the jet radius treated at fixed order, and match it to the exact next-to-leading-order (NLO) cross section. We correct the one-loop soft anomalous-dimension matrices in the literature and validate the hard functions. We compare the leading-power kernel with exact real emission, and validate this comparison on Drell-Yan production. At invariant masses up to about 1 TeV with a fixed jet threshold the method does not apply. At 1-4 TeV with a jet threshold of 15% of the invariant mass, the kernel describes the exact real-emission density close to threshold but fails once the emitted energy exceeds about 100-200 GeV. We trace part of this failure to initial states absent from the kernel, which depend on $\tau=M^2/S$, with $M$ the invariant mass and $\sqrt{S}$ the hadronic centre-of-mass energy. In addition, at fixed $\tau$ the difference between the kernel and the exact result grows with the ratio $M/m_t$ of the invariant mass to the top-quark mass. As a consequence, the resummed correction beyond NLO, for a scale equal to the invariant mass, ranges from 1-2% to 17-23% for different treatments of this region. The upper end becomes 13-16% for a jet threshold of 20% of the invariant mass and 23-33% for 10%. The range is not a bound, and the correction is not under perturbative control. We therefore keep NLO as the central prediction. We compare the possible effect of resummation with the NLO scale uncertainty, with the known next-to-next-to-leading-order corrections, and with NLO matched to a parton shower.
Read the full abstract
Optimization and Control
Receding horizon control with sensor-induced positive semidefinite state penalization
This paper develops and analyzes a receding horizon control (RHC) framework for the stabilization of infinite-dimensional nonautonomous continuous-time systems with an infinite-horizon control objective. The considered stage costs penalize only a finite-dimensional component of the state and are therefore positive semidefinite with respect to the full state. Without imposing terminal costs or terminal constraints, we establish exponential stability and derive suboptimality estimates under suitable stabilizability and detectability-type assumptions, together with an online selection of the sampling times. In the linear case, we extend the framework to output-feedback RHC by combining it with a Luenberger-type state estimator based on finitely many sensors, and establish closed-loop stability using an augmented Lyapunov functional. The abstract state-feedback results are verified for a nonlinear Schlögl parabolic equation around an unstable steady state and, in the output-feedback setting, for nonautonomous linear parabolic equations. The control and observation realizations involve finite-dimensional actuation and sensing, including localized boundary controls and spatially localized state-penalization operators. Numerical experiments support the theoretical findings and illustrate the performance of the proposed RHC schemes.
Read the full abstract
Computation and Language
Perturbing Inputs Improves Pretraining Distillation
Knowledge distillation trains a student language model to match a teacher's predictions on the training text. We ask whether learning the teacher's behavior beyond the training text can further improve the student. We investigate this using three input perturbations: additive token noise, blending two documents, and blending neighboring tokens. Specifically, for an early part of pretraining we perturb the inputs and train the student to match the teacher on them, and we spend the rest of the same token budget on ordinary distillation. When pretraining from scratch, the best method improves on standard distillation at both scales we tried, by 0.025 nats at 135M parameters and by 0.020 nats at 0.5B (mean of three seeds). At 135M this is 1.45 times the gain that distillation itself provides over next-token training. We also observe promising downstream transfer: at 0.5B, the best variant improves average zero-shot accuracy by about one percentage point across seven tasks. It also retains its validation-loss advantage after low-bit weight quantization. We also test a data-limited setting in which training data must be reused, and find that the gains persist. Finally, ablations suggest that, for additive noise, perturbing the teacher's inputs contributes most of the gain, while perturbing only the student's inputs yields much less.
Read the full abstract
Computation and Language
Out of Parameters, Saved by Syntax: UD-PrMPT for Parser Augmented Low Resource Machine Translation
Recent work advancing machine translation quality using LLMs for very low resource languages has relied on the injection of lexicon-based prompt components to expose vocabulary to models. However, these approaches run into a ceiling when models must guess sentence meaning purely from plausible combinations of possible word senses mapped to strings. In this paper we test UD-PrMPT, an approach using syntactic analyses in the Universal Dependencies formalism as prompt components to improve MT in five diverse low resource languages: Beja, Chintang, Chukchi, Coptic, and Xibe. Our results show additive improvements which stack on top of the lexicon based approach which are stronger for higher quality manual analyses but consistent for predicted parses as well.
Read the full abstract
Nuclear Theory
Symmetry-preserving higher-derivative formulation of non-relativistic baryon chiral effective field theory
A symmetry-preserving higher-derivative formulation of the non-relativistic baryon chiral effective field theory is introduced. Integral equations for the two-nucleon scattering and the deuteron structure functions are derived and numerical solutions at leading order are obtained.
Read the full abstract
Quantum Physics
A geometrical approach for obtaining exact analytical solutions for a scalar field in a one-dimensional dynamical cavity
The dynamical Casimir effect was first predicted by G. T. Moore in 1969, who considered a real scalar field inside a cavity with a moving mirror in ($1+1$)D, satisfying Dirichlet boundary conditions. Moore wrote the general solution for the field in terms of a function that satisfies a functional equation nowadays known as the Moore equation.In the literature, exact analytical solutions to the Moore equation form a rather small set, restricted to specific classes of laws of motion. Cole and Schieve (1995) developed a geometrical approach that yielded a general formula for the solution of the Moore equation for cavities initially at rest. However, its application has always been done through numerical calculations, and their approach has therefore long been regarded solely as a numerical method. In this work, we show that their method can also be employed as a powerful method for deriving exact analytical solutions, and show the applicability of this approach to deriving analytical solutions to the Moore equation. We obtain original exact analytical solutions for an expanding cavity with a uniformly moving mirror, a contracting cavity with a mirror that approaches the static one asymptotically, and, in addition, we recover the first analytical solution to the Moore equation for a resonant mirror motion, originally obtained by Law in 1994. Moreover, we extend the Cole-Schieve approach, showing that exact solutions can be constructed from any spacetime region where the Moore function has already been determined. We show that this extended geometrical approach allows us to expand the set of known analytical solutions to the Moore equation from the few classes of motion reported in the literature to an infinite set, as it enables the treatment of new classes of laws of motion obtained by combining those presented here with each other as well as with previously known classes of trajectories.
Read the full abstract
High Energy Physics (Theory)
Liberation in the Heavy Limit
We study CFT four-point functions of identical scalar primaries as their dimension $\Delta$ grows for fixed $0<z, \bar z<1$, and call them liberated if they grow at the same exponential rate as those of a generalized free field (GFF), as is the case for heavy free particles propagating along geodesics in AdS. We show that the correlator is liberated at the crossing symmetric point if and only if the light spectrum is sparse, and that liberation then holds wherever $z$ and $\bar z$ lie on the same side of $1/2$. If the low twist states are also sparse, liberation holds everywhere except in an explicit band around $z+\bar z=1$. In particular, if the primary twist gap lies above $\sqrt{2}\Delta$, then the correlator is liberated everywhere. For twist cutoffs below $\Delta$, composites of two spinning generalized free fields fail to be liberated in a region that fills the band near the double lightcone limit. In two dimensions, Virasoro primaries cannot be liberated unless the central charge grows faster than $\Delta$. The same framework sharpens the universality of large-$c$ torus partition functions of two-dimensional CFTs.
Read the full abstract
Combinatorics
Kronecker coefficients of large degree and Littlewood-Richardson inequalities
We prove that Kronecker coefficients are uniformly asymptotic to $f^\lambda f^\mu f^\nu/n!$ when all three irreducible dimensions are at least $\sqrt{n!}\,e^{-A\sqrt n}$, for any fixed $A>0$. This strengthens a conjecture of Pak, Panova and Yeliussizov. For triple staircase and triple square partitions, we obtain relative asymptotics with respective limiting factors $e^{-1/24}$ and $e^{-1/3}$. For Littlewood-Richardson (LR) coefficients, we bound the sum of the three dimension deficits by the LR coefficient deficit times $O(\log n)$, proving their Conjecture 5.5 and extending it to arbitrary fixed size proportions. We also construct vanishing LR coefficients with three Plancherel dimension sequences and determine the exact stabilization threshold of the largest LR coefficient with one input size fixed.
Read the full abstract
Sound
LAION-Audio-XL: A Web-Scale Corpus and Scaling Studies for Contrastive Language-Audio Pretraining
Recent progress on contrastive audio-language models can be attributed mainly to the scale and quality of their training data. Larger training datasets expose models to a broader range of sounds and corresponding audio captions. To exploit this, we collect 1.48 billion audio samples from academic datasets, internet datasets, web collections, Common Crawl, and YouTube, and process them with a multi-stage curation pipeline. Furthermore, we generate short and long captions for samples in our corpus that lack them. We use our dataset to study the scaling behavior of contrastive audio-language models based on the NaFlex architecture for the audio encoder. Our experiments identify trade-offs between model size and training data for a given compute budget. Moreover, our resulting models, CLAP2, outperform prior audio-text models of comparable size on the MAEB and HEAR benchmarks. We release our dataset, captions, model checkpoints, and code.
Read the full abstract
Probability
Discrete Distributions and Their Limit Theorems in the Surreal Numbers
We extend the discrete uniform, hypergeometric and binomial distributions to the surreal numbers, allowing infinitely many outcomes, chips and trials. A uniform distribution on $\mathbb{N}$ gives each outcome probability $1/\omega$; on the eventually periodic sets we define the needed sizes explicitly, obtaining a probability that is regular and strictly monotone. For an urn with infinitely many chips and finitely many draws, the hypergeometric distribution is within an infinitesimal of the binomial. With infinitely many trials, and under a hypothesis on factorials of omnific integers that we show can be satisfied, we prove a surreal de~Moivre–Laplace theorem, and we show that the hypergeometric distribution with infinitely many draws is locally normal with a finite-population correction. Near its center, it is relatively close to the binomial distribution if and only if the sampling fraction is infinitesimal, an exact form of a familiar rule of thumb. For rare successes we prove, without any hypothesis on factorials, a surreal Poisson limit theorem and describe urns with finitely many red chips, where the binomial approximation fails beyond the first success.
Read the full abstract
Optics
Comments on: "Optimized thermal control of a dual-wavelength-resonant nonlinear cavity"
This comment addresses the publication [Appl. Opt. 65, 7090 (2026)] which targets an alleged “optimization” of a patented invention conceived by the author. The invention had introduced a multi-temperature management scheme for a dual-wavelength-resonant nonlinear cavity that optimized the crystal temperature around the beam waist for phase-matching, while utilizing independently adjustable temperatures in other regions of the same crystal to simultaneously satisfy the double-resonance condition. A foundational form of this invention was initially utilized by co-author H. Vahlbruch without the author's prior knowledge during the construction of the squeezed-light laser for the Virgo gravitational wave detector. In their publication, the three authors now propose a variant that reduces the maximum achievable temperature differential. Here, I demonstrate that this approach does not constitute an optimization but, on the contrary, represents a regression that dismantles the invention. In general, a high-quality implementation of this invention fundamentally depends on maximizing the achievable temperature differential. Only then can the problematic phase differences—introduced by dielectric coatings, Gouy phase shifts, and suboptimal poling lengths at the crystal facets—be effectively compensated for, particularly in short crystals. If desired, smaller temperature differences can always be selected dynamically anyway.