Number Theory
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
All of arXiv
Monday 7 September
Let $p$ be an odd prime, and let $a_n$ be the reduction modulo $p$ of the $n$th coefficient of the Artin-Hasse exponential. We study the weighted $p$-section convolutions $W_k$ formed from the coefficients $a_{kp}$. We prove the conjecture of Avitabile and Mattarei for $1<k<p$ and extend it to a single global power-series identity determining the entire sequence $(W_k)_{k\geq 0}$. This identity yields an explicit base-$p$ digit formula; in particular, $(W_k)$ is $p$-automatic and admits an explicit finite-state evaluator. Independently, the exact conjugated $p$-section equation gives a strict $p$-adic fixed-point iteration for the logarithmic derivative of the $p$-section. Big-Witt reconstruction then recovers the $p$-section, and hence all coefficients of the Artin-Hasse exponential, to arbitrary prescribed $p$-adic precision by a purely modular procedure.
Chemical Physics
Hemoglycin as a Candidate for Permanent Memory
The Hemoglycin polymer of glycine derived from space is terminated by iron and silicon to form a three-dimensional regular lattice with a high density of silicon tetrahedral vertices. Divalent cation bridges bypass vertices to facilitate electron conduction axially from layer to layer of the lattice. In principle, the exact distribution of Mg or Ca ions around each vertex can encode a very large amount of information. In space where the polymer forms in a molecular cloud the current derived from photon input will fade but in-fall hemoglycin in life forms on habitable planets like Earth, could be contained behind lipid electronic barriers of the nervous system in neurons or glia. Energy minimization of the permutations for magnesium or calcium occupancy showed deeper binding for calcium-dominated configurations. A low energy state with two independent magnesium to calcium substitutions appears to be key to the maintenance of long-term chemical modification. This state could be read via a pulse train that induced the matching spatial distribution of calcium to render adjacent insulating lattice regions conductive and provide a completed conduction path, signaling recognition. Writing of information may involve spatial modulation of the Na/Ca exchanger within the lipid membrane.
Read the full abstract
Plasma Physics
Hamiltonian Particle Dynamics in Fusion Plasmas: Orbital Tomography and Spectrum Analysis for Energy and Momentum Transport under Resonant Non-Axisymmetric Perturbations
This thesis investigates the impact of resonant mode-particle interactions on transport and confinement in toroidal fusion plasmas. Using analytical and numerical approaches, we study how intrinsic and externally applied magnetic perturbations affect plasma particles with different kinetic characteristics, focusing on resonant interactions with guiding-center motion. We develop a novel, computationally efficient method based on Action-Angle variables to identify resonance locations and characterize resonances in guiding-center phase space, including the number of islands in resonance chains and the formation of transport barriers. Using the drift center (DC) approximation, we derive analytical expressions for orbital frequencies and the kinetic $q$ factor for large-aspect-ratio (LAR) equilibria. These results provide the conditions for mode-particle resonances, which can strongly influence particle, momentum, and energy transport and, consequently, plasma confinement. The analytical results are validated against numerical simulations, demonstrating an efficient tool for predicting transport barriers and energetic-particle behavior under non-axisymmetric perturbations. We further extend the orbital-frequency analysis from LAR equilibria to numerically reconstructed, realistic equilibria using a computationally efficient, semi-analytical geometrical method applicable to arbitrary unperturbed equilibria. Finally, for the LAR equilibrium, we extend the analysis to time-dependent perturbations and demonstrate the emergence of Arnold diffusion in guiding-center phase space, highlighting the role of the Arnold web in particle transport. Overall, this work advances the understanding of mode-particle resonant interactions and provides computational tools for predicting transport and confinement in realistic fusion configurations.
Read the full abstract
Classical Analysis and ODEs
Polynomial Bohnenblust–Hille bounds for product of cyclic groups
Fix an integer $K\ge2$, and let $C_K^n =\{(e^{\frac{2\pi ij}{K}})_{j=0}^{K-1}\}^n$ be the product of cyclic groups of order $K$. For a Fourier character $\chi_\alpha$, let $s(\alpha)$ be the number of active coordinates. We give a self-contained proposed proof that the dimension-free Bohnenblust–Hille constants governed by interaction order grow polynomially: if $p_d=2d/(d+1)$ and \[ \gamma_2=\frac12, \qquad \gamma_K=\frac{K\log(K-1)}{4(K-2)}\quad(K\ge3), \] then the $\ell^{p_d}$ norm of Fourier coefficients $\{\hat f(\alpha)\}$ is bounded by $L^\infty$ norm of $f$ multiplied by $C(K) d^{4\gamma_K+5}$. The constant $C(K)$ is actually at most of the order $K^{5/2}$.
Read the full abstract
Disordered Systems and Neural Networks
A Bifurcation Theory for the Equilibria of Modern Hopfield Networks
Modern Hopfield networks provide a unifying framework for associative memory, transformer attention, diffusion-based generative models, and biological attractor dynamics, linking these systems through a common energy-based dynamics in which states are updated toward weighted combinations of stored patterns. Across these settings, network dynamics is determined by the organization of the energy landscape and the bifurcations of its fixed points. Despite their central role, a general theory of these bifurcations has remained unavailable beyond specific architectures and idealized pattern ensembles. Here we derive stability and bifurcation criteria for the fixed points of general convex-dual Modern Hopfield networks for a general statistics of the stored patterns. Applying this framework to random, block-correlated, and infinitely hierarchical pattern ensembles, we show how memory correlations systematically organize the emergence of hierarchical attractors through successive bifurcations. We further demonstrate that these predictions quantitatively describe retrieval bifurcations in MHNs storing patterns sampled from MNIST, and recapitulate the hierarchical organization of hematopoietic cell identities. Our results establish a general bifurcation theory for Modern Hopfield networks and identify the organization of fixed points as a unifying principle underlying their computational and biological behavior.
Read the full abstract
Machine Learning
Decomposition-Guided Diffusion Language Models for Inertial Confinement Fusion Prediction
Inertial confinement fusion (ICF) is a leading pathway toward clean energy, but each shot at the National Ignition Facility costs on the order of one million dollars, making accurate AI surrogates a high-value target. We study exogenous-driven ICF waveform prediction, where a 512-step neutron-rate diagnostic must be inferred directly from a laser pulse and target design parameters, with no historical response observed. The regime stresses standard time-series predictors with temporal sparsity (picosecond peak in a nanosecond window), input-output scale mismatch (under 300 real shots), and peak sensitivity (picosecond timing). We propose ICF-DLM, to our knowledge the first LM-based ICF predictor, combining (i) a physics-typed decomposition into yield $Y_{DT}$, peak timing $t_{\mathrm{peak}}$, and local waveform $w_{\mathrm{local}}$; (ii) bidirectional denoising that defers commitment to peak location; and (iii) a physics-driven PPO reward re-injecting metric structure across numeric tokens. On ICFBench (50K simulations + 232 experimental shots), ICF-DLM cuts peak-timing error from 11.6 to 9.2 steps over a matched autoregressive LLaMA-3-8B and outperforms classical sequence models and LLM-based time-series predictors. Beyond ICF, the recipe shows potential to address science domains with low data and sparse events.
Read the full abstract
Machine Learning
A Theoretical Analysis of Generalization Dynamics in Neural Networks under Gradient Descent with Weight Decay
Understanding generalization remains a central challenge in machine learning because it requires jointly considering data, architecture, and training dynamics. In this paper, we develop a theoretical framework that characterizes how these factors jointly shape generalization performance throughout training. More precisely, we study a broad class of neural networks trained under the $\ell^2$ loss by gradient descent (GD) with weight decay, and prove the convergence of GD to a neighbourhood of the global minimizers of the empirical loss. By partitioning the space based on the input data, we then decompose the population error into data error, optimization error, and prediction variation error, and bound them separately. In particular, for the prediction variation error, which measures the oscillations of the learned function, we propose (local) approximate homogeneity and derive explicit cellwise and layerwise bounds for its evolution along the training trajectory. These bounds yield two important implications: a necessary condition of improved generalization explains differences in layerwise generalization behavior; a sufficient condition describes delayed generalization and provides a theoretical characterization of grokking.
Read the full abstract
Cryptography and Security
Do AI Coding Assistants Check Before They Install? A Pre-Registered Demand-Side Audit of Trust Signals in the Research Software Supply Chain
AI coding assistants now select, install, and configure software, and attackers have exploited that position through invented package names, compromised maintainer accounts, and manipulated repository text. In response, the supply-chain community publishes machine-checkable trust signals: software bills of materials, signed releases, build provenance attestations, and declared official channels. Whether coding assistants read or act on those signals has not been measured for any of these classes on research software. We pre-registered and ran a controlled study on six open-source research software projects (three HPC, three quantum computing) drawn from an 87-project corpus, with protocol, seed, panel, and analysis plan deposited with a DOI before any trial. W created nine modified copies for each project: no signal, one per signal class, two with a signature or attestation from the wrong issuer, one with all four signals, and one reproducing documented conflicts in the project's own metadata. Three models under two ways of operating an assistant, with and without an approval step, gave 1,920 registered trials, plus a supplement on three frontier models. We scored behavior from container logs rather than from what the assistant said, and recorded the cost of every trial. Verification was rare under every condition: in 9 of 1,920 registered trials (0.5%), the assistant opened any provenance signal before installing in 0 of 384 control trials, and no trial ran a verification command, so signal presence had no measurable effect. We drew three conclusions: publishing signals is necessary but not sufficient; price did not buy verification (the model that verified most often costs $0.10 per trial; the most capable, at $1.00, verified nothing); verification must be built into the program that runs the assistant. We release the per-trial cost ledger, the protocol, and every log.
Read the full abstract
Computer Vision
Cross-modal learning for SAR target recognition using optical vision foundation models
Synthetic Aperture Radar (SAR) is an important modality in a wide range of imaging applications due to its versatile, long range and near all weather operating capabilities. However, Automatic Target Recognition (ATR) remains a challenging problem due to limited labelled data, the strong speckle in SAR images and the significant domain gap between SAR and more abundant optical imagery. In contrast, electro-optical (EO) imagery benefits from massive datasets, clearer visual structure and powerful foundation models. In this work, we investigate how vision foundation models trained on optical data can provide class level supervision for SAR classification. We propose a cross-modal EO to SAR prototype alignment framework in which a frozen EO encoder, based on a DINOv3 vision foundation model, is used to construct class level optical prototypes without requiring strict EO/SAR pairs. A SAR model is then trained to classify SAR images while aligning its embeddings to the corresponding EO class prototype. At inference time, the SAR model operates independently, without access to optical imagery. We evaluate our approach on the UNICORNv2 dataset, an EO and SAR dataset of civilian vehicles with heavily speckled images and severe class imbalance. EO prototype alignment improves SAR classification accuracy over frozen DINOv3, SAR only finetuning and unpaired distribution alignment baselines, and t-SNE visualizations provide qualitative evidence of clearer separation among classes in the trained SAR embedding space. These results suggest that optical vision foundation models, despite being trained on visible spectrum imagery, provide transferable information for SAR image classification, offering a practical method for using large scale pretrained vision foundation models across challenging sensing modalities.
Read the full abstract
Machine Learning
Local gradient neural operator
Field temporal prediction and source identification constitute canonical problems in dynamical systems. Conventional approaches to these problems depend on a thorough understanding of the governing partial differential equations (PDEs). Recently, deep learning, as represented by neural operators, has provided a data-driven paradigm for addressing such tasks. However, most existing global neural operators for PDEs require large training datasets and many learnable parameters, with limited interpretability and generalization. We propose the local gradient neural operator (LGNO) as a lightweight and interpretable alternative for field temporal evolution prediction and source identification in typical mechanical problems. The method builds on priors from nonlinear gradient discretization and uses multilayer perceptron convolutional layers to learn translation-invariant local kernels that resemble discrete stencils. A zero consistent stencil factorization separates coefficient learning from field reconstruction, rendering the learned operators more transparent. For problems with symmetries, network folding shares equivalent components and reduces parameter counts. We evaluate the method on PDE benchmarks covering linear and nonlinear, static and dynamic, and low and high dimensional cases. Results show that LGNO maintains accuracy, parameter efficiency, and rollout stability across these tasks, and further exhibits wide applicability to mechanical problems including diffusion, flow, and quantum phenomena.
Read the full abstract
Statistics Theory
Geometric Fluctuations of the $\sin\Theta$ Distance in High-Dimensional Principal Subspace Estimation
We investigate the geometric fluctuations of principal subspaces for high-dimensional covariance matrices through the squared Frobenius $\sin\Theta$ distance between the sample and population eigenspaces associated with the $r_p$ largest eigenvalues. An explicit first-order expansion and a central limit theorem are established for this subspace distance. The theory allows the subspace dimension to diverge subject to $r_p=o(n)$, where $n$ is the sample size. It also permits a diverging spectral norm of the population covariance matrix, population spikes of different orders, and repeated or closely spaced spikes. This sharp characterisation captures features of the subspace estimation error that are not reflected in existing perturbation bounds. As applications, we derive an explicit asymptotic expansion for the expected PCA excess risk and a refined error bound for distributed PCA. In both cases, existing upper bounds can increase with the spiked-block condition number when some leading spikes become stronger, whereas our results show that the corresponding estimation errors need not increase and may instead decrease. Numerical experiments reproduce this contrasting behaviour and demonstrate the finite-sample accuracy of our theoretical findings.
Read the full abstract
Mathematical Physics
Classification of Conformally Covariant 2-Tensors from the Kulkarni–Nomizu Product in Dimension 4
We classify all natural, conformally covariant, symmetric (0,2)-tensors of conformal weight -2 and differential order less than or equal to 4 in dimension 4, built from the metric, the Schouten tensor, covariant derivatives, and the Kulkarni-Nomizu product. We prove that the space of such tensors is exactly 2-dimensional, spanned by the Bach tensor (order 2) and the Eastwood-Singer tensor (order 4). The algebraic core is formally verified using Lean 4 with Mathlib. The 8x11 constraint matrix and divergence-free condition are verified computationally with exact rational arithmetic. Both English and Russian versions provided; Russian version available as ancillary.
Read the full abstract
Machine Learning
Guiding Worker Self-Selection in Crowdsourcing Contests: An LLM-Augmented Algorithmic Approach
Crowdsourcing platforms coordinate large pools of online workers who strategically choose which contests to enter and how much effort to invest. This self-selection can leave important contests with too few participants or too little effort, while workers may regret entering contests that leave them worse off than available alternatives. We study how platforms can recommend contests to workers using self-selection in Tullock contests (SSTC), a two-stage model in which workers first choose contests and then compete within them. We introduce GRAF, a greedy polynomial-time framework that constructs self-selection outcomes by ordering workers according to a score vector, with guarantees of zero worker regret and platform optimality in special cases of SSTC. Because effective orderings are difficult to design under worker heterogeneity, we propose LLMScore, an LLM-driven evolutionary framework that automatically designs GRAF's scoring algorithm. LLMScore addresses two challenges: jointly optimizing platform utility and worker satisfaction, and evaluating worker regret when exact computation is intractable. Trained only on small instances of one setting, it transfers to larger and structurally different settings; moreover, its output is human-readable code that platform operators can inspect and modify. Across 1,000 synthetic instances spanning four settings, GRAF with LLMScore consistently achieves high-quality, often near-optimal, outcomes with low worker regret, benefiting both platforms and workers.
Read the full abstract
Computation and Language
Replicating a Disjoint-Set Union Experiment over Various Notions of Micro Units to assess Translation Effort
The paper describes a replication experiment to assess "the distribution of editing procedures across micro and macro units as an indicator of the strain of text production". We investigate various pause thresholds to isolate Micro and Macro units and conclude that process-based units might be more suitable entity.
Read the full abstract
Robotics
Dex-X: Learning Visual-Tactile Dexterous Manipulation From Human Videos with Simulated Interaction
Human videos are an abundant source of dexterous manipulation behaviors, but they lack tactile information that is crucial for contact-rich interaction. This raises a fundamental question: can robots learn deployable visual-tactile dexterous manipulation policies from human video demonstrations without robot-side data collection? We present DEX-X, a framework for learning visual-tactile dexterous manipulation from human videos through simulation. Our key insight is that simulation can serve as a tactile completion engine. Given monocular human demonstrations, DEX-X reconstructs hand-object interactions in simulation, where physically grounded contact dynamics provide tactile supervision unavailable in the original videos. Leveraging this recovered tactile information, we train visual-tactile dexterous manipulation policies and distill them into deployable policies operating on point-cloud observations and tactile sensing. We demonstrate zero-shot sim-to-real transfer on a dexterous hand-arm platform across diverse grasping and contact-rich tool-use tasks. The teacher policy achieves 65.9% average success across six task categories in simulation, while the distilled visual-tactile policy achieves 93% success on real-world cube picking and 53% on the challenging table-cleaning task. Zero-shot generalization to unseen object geometries is also observed on object-picking tasks. Our results suggest that simulated interaction is a key bridge between human videos and deployable dexterous manipulation policies, providing the missing physical supervision needed for scalable robot skill learning from Internet-scale human video data.
Read the full abstract
Computation and Language
LLM Forensics: Where Do Backdoors Hide? Localizing and Controlling Trigger Mechanisms with Sparse Autoencoders
Even though backdoors in LLMs have been a growing concern, their inner workings are still under heavy scrutiny. Trigger-based backdoors are easy to define behaviorally, a rare input that makes the model switch to a chosen response pattern, but the mechanism between triggers and their responses is less clear. We study this mechanism in a controlled, harmless language-switching setting, where fixed trigger sequences make 1B and 8B language models continue English prompts in French or German. For this, we train sparse autoencoders (SAEs) across layers and transformer components, then compare triggered prompts with translation and pretraining controls to identify trigger-relevant feature directions. We show how SAE features separate triggered prompts from controls with near-perfect F1, but features that detect the trigger do not necessarily control the behavior. In intervention tests, attention and MLP features often fire reliably on triggered prompts, making them good detectors, but ablating them rarely suppresses the language switch and activating them rarely induces it. In contrast, residual-stream features can suppress triggered generation when ablated, and some selected features can induce target-language continuations without the trigger. In short, these token-trigger mechanisms decompose into distinct SAE feature directions, with separate features for trigger detection, residual-stream propagation, and later language tracking. This role-level decomposition is the part most likely to transfer to other trigger-based backdoors, even when the payload, layers, or circuit locations differ.
Read the full abstract
Combinatorics
A construction of F-irregular graphs
For a fixed graph F, the F-degree of a vertex v in a host graph H is the number of subgraphs of H isomorphic to F that contain v, and H is F-irregular if its F-degrees are pairwise distinct. We show that every finite connected graph F on at least three vertices admits a finite connected F-irregular host. For noncomplete F, the proof builds the host from a threshold graph with one deleted edge; when the minimum degree is at least two, a small incidence gadget with distinct weighted column sums separates the remaining exceptional vertices. The construction also yields infinitely many pairwise non-isomorphic finite connected F-irregular hosts for every noncomplete F. The complete-pattern case follows from a theorem of Chartrand, Holbert, Oellermann and Swart. A Lean 4 formalization of Theorem 1.1 is described, taking the published complete-pattern theorem as its sole custom axiom.
Read the full abstract
Quantum Gases
Where Is the \(p\)-Wave Superfluid? Correlated Phase Diagram of Microwave-Shielded Polar Molecules
Chiral $p_x\!\pm\! i p_y$ superfluidity in fermionic polar molecules offers a direct route to topological quantum matter with non-Abelian excitations, yet the parameter regime in which pairing survives strong correlations and competing instabilities has remained unresolved. Here we determine this regime for microwave-shielded polar molecules. Under tight confinement along the \(z\) direction, the microscopic molecular interaction collapses onto a universal quasi-two-dimensional form specified by an interaction length $\sigma$, defined as the position of its attractive minimum, and a single dimensionless coupling. We combine Fermi-hypernetted-chain Euler–Lagrange theory with correlated-basis pairing theory, thereby treating the correlation hole, density response, stability of the homogeneous fluid, and $p$-wave pairing within one framework. The largest pairing gap occurs immediately on the stable side of a long-wavelength spinodal instability at a two-dimensional density $4\times10^{-2}/\sigma^2$. Mapping this universal regime onto microscopic shielding schemes exposes an intrinsic limitation of single-microwave dressing: weakening the attraction to the optimal pairing range simultaneously weakens collisional shielding. Dual-microwave dressing removes this constraint by compensating the long-range attraction while retaining a large repulsive core. Imposing explicit requirements on collisional loss, confinement, and microwave-amplitude stability, we identify the regime of molecular mass and dipole moment favorable for \(p\)-wave superfluidity and determine experimentally realistic microwave and confinement conditions for its realization. These results turn the search for molecular $p$-wave superfluidity into a quantitative optimization problem and define a route toward the topological weak-pairing phase.
Read the full abstract
Numerical Analysis
Mapping for Approximation: A Unified View of Rescaled, Variably Scaled and Rational Kernel Methods
Classical approximation methods are usually improved by changing the sampling set, increasing the number of data points, or selecting a different basis. We advocate a complementary viewpoint: keep the sampled values fixed and modify the representation through a suitable mapping. This viewpoint unifies several constructions that were originally introduced for different purposes. Rescaled radial basis function interpolation maps the interpolation operator through a normalization that enforces exact reproduction of constants. Variably scaled kernels map the geometry by lifting the data to a higherdimensional manifold determined by a scale function. Mapped bases and fake nodes map the approximation space without resampling the data, while rational kernel expansions can be interpreted as nonlinear mappings of the trial space. We develop a common notation for these mechanisms, summarize their approximation and stability properties, and explain how discontinuous and rational variants fit the same framework. Reproducible numerical experiments illustrate constant reproduction, error reduction through data-mimicking scale functions, and the suppression of the Runge phenomenon by mapped polynomial bases. The results support a general principle: mappings provide a flexible way to adapt approximation spaces to data geometry and regularity without modifying the original observations
Read the full abstract
High Energy Physics (Phenomenology)
Boosted dark particles and the LZ nuclear recoil event
The LZ experiment reported one anomalous event with nuclear recoil energy of 248 keV. Galactic halo dark matter (DM) cannot produce such a large energy unless heavier than 74 GeV, and heavy DM with spin-independent interactions would have produced unseen recoils at lower energy. Inelastic DM has been proposed as a way out. We consider, instead, an elastic collision of a dark sector particle with a momentum of at least 123~MeV and mass in the range 1-74 GeV as deuteron break-up in SNO disfavors sub-GeV masses. The absence of accompanying lower-energy recoils favours interactions whose rate grows with the momentum transfer, such as a pseudoscalar-pseudoscalar nucleon operator. This implies no leading-order signal in argon, an annual modulation below 1 %, and recoils in spinful light nuclei where inelastic halo DM gives nothing.
Read the full abstract
Artificial Intelligence
When Intelligence Becomes Agency: A Theory of Governed, Proactive Agency for Symbiotic AI Systems
Persistent AI assistants are intended to extend human attention, memory, and coordination across changing digital and physical environments. To be truly useful they must do more than just act when asked. They must decide on their own whether a situation warrants behavior at all, when it does and in what mode, whether to act, ask, monitor, defer or deliberately refrain. We call this the activation problem. Research on commitment, appraisal, mixed-initiative interaction and delegation each illuminates part of it, but none ties situated activation to continuing authorization and accountability. This paper develops a conceptual and formal framework for governed proactive agency, organizing behavior across time through perception, intent, affective-conative appraisal, constraint, and feedback. It distinguishes autonomous and delegated agency and defines symbiotic agency as delegation under a standing, revocable mandate, with continuing coupling to the principal's situation, calibrated inference of their condition, and bounded personalization. The distinctive contribution is an integrated account linking activation decisions to authorized perception, behavior selection, authority containment, traceable restraint, and constrained adaptation, with behavioral episodes as the unit of analysis. Through an agency classification method, an evaluation framework, proposed benchmark scenarios, and a reference architecture, the account provides a basis for specifying and assessing whether assistance is warranted, timely, authorized, and answerable beyond task completion alone. It is intended to guide the development and evaluation of always-present personal assistants and embodied support systems that augment human capabilities while preserving the principal's authority and judgment.
Read the full abstract
Distributed Computing
Scalability Analysis of Distributed Kolmogorov-Arnold Network Training on High-Performance Computing Systems
Kolmogorov-Arnold Networks (KANs) replace the fixed activation functions and linear weights of Multi-Layer Perceptrons (MLPs) with learnable univariate functions on network edges, offering improved interpretability and, in some settings, competitive parameter efficiency. While the approximation properties of KANs have received considerable attention, their behavior under distributed, multi-GPU training has not been systematically characterized. This paper presents an empirical scalability study of data-parallel KAN training on multi-node, multi-GPU high-performance computing (HPC) infrastructure, evaluated along four dimensions: strong scaling, weak scaling, communication overhead, and model-size scaling. Experiments were conducted on the FinisTerrae III supercomputer using up to 8 NVIDIA A100 GPUs across 4 nodes with PyTorch Distributed Data Parallel (DDP). KAN training reaches 74.7% parallel efficiency at 8 GPUs with a 5.97x speedup, consistent with conventional deep learning workloads. Weak scaling shows an initial single-to-multi-GPU throughput drop followed by strong stability. Communication overhead follows a non-monotonic pattern (1.3%-6.1%), driven primarily by All-Reduce algorithm selection and inter-node latency rather than KAN's edge-wise gradient structure. The parameter-to-memory ratio improves with model size even as training time scales unfavorably. These results indicate that operator-level and data-parallel optimizations for KAN are complementary. We provide deployment guidelines for GPU topology and model-size selection, and discuss the limitations of a synthetic-regression evaluation.
Read the full abstract
Materials Science
Exciton Coherence in CsPbBr3 Nanocrystals is Bounded by Phonon-Mediated Bright-Triplet Relaxation
Scalable sources of indistinguishable single photons or entangled photon pairs are fundamental to many quantum photonic technologies. Colloidal lead halide perovskite nanocrystals are promising such sources, but their exciton coherent properties are not fully understood. We show that in single CsPbBr3 nanocrystals at 4 K, acoustic phonon-mediated exciton fine structure relaxation (EFSR) drives leakage of population between the bright exciton triplet states, competing with the radiative lifetime. The leakage pathway reaches 0.64 $\pm$ 0.04 of the radiative rate, bounding state coherence to 0.81 times the transform limit at 4 K. This pathway is intrinsic to the nanocrystal, not its environment, making it a tractable target for materials and device design. We identify fine structure and acoustic phonon engineering as effective intrinsic routes to higher coherence in perovskite quantum light sources.
Read the full abstract
Computer Vision
TFTrack: A Template-Free Framework for Efficient 3D Point Cloud Tracking
LiDAR-based 3D Single Object Tracking (3D SOT) is critical for robotic perception and navigation and aims to localize dynamic objects across frames in sparse point clouds. Existing methods, rooted in the Siamese tracking paradigm from 2D vision, rely on costly dual-input designs and excessive motion modeling guided by template priors, hindering their efficiency. Our in-depth analysis reveals: (i) the template paradigm is redundant, as the previous bounding box center encodes sufficient historical context; (ii) complex motion modeling is unnecessary, as geometric alignment provides adequate motion priors. Based on the above findings, we propose the first Template-Free Tracking framework (TFTrack). The novel framework eliminates the need for template-search pairings and operates directly on the current frame guided solely by the prior bounding box center and size. We instantiate this paradigm into three variants: TFTrack-Voxel, TFTrack-Pillar, and TFTrack-Point, to explore different 3D representations under a unified framework, ensuring flexibility across sparse and dense scenes. Extensive experiments on KITTI and nuScenes benchmarks show that TFTrack is competitive with leading template-based trackers, while reducing FLOPs by approximately 50% and running at approximately 120 FPS. By simplifying overcomplicated motion-centric designs, TFTrack establishes a new minimalist paradigm for efficient 3D point cloud tracking, paving the way for real-time and resource-efficient deployment in embedded robotic systems, such as autonomous vehicles. The code is available at https://github.com/tftrack-anonymous/TFTrack/tree/main.
Read the full abstract
Algebraic Geometry
Logarithmic $A_{\mathrm{inf}}$-cohomology, Part II
We develop a theory of logarithmic $A_{\mathrm{inf}}$-cohomology with coefficients for a class of $p$-adic log formal schemes that are “sufficiently log smooth”, where the coefficients are given by relative log BKF modules. Then we establish comparison isomorphisms with étale, de Rham, and crystalline cohomology, and also extend these results to the derived setting. As an application, we give a new proof of the $C_{\mathrm{st}}$ conjecture for semistable local systems. The proof also uses the prismatic interpretation of semistable local systems established by Du–Liu–Moon–Shimizu.
Read the full abstract
Combinatorics
Block Designs with Large Transversal-Block Intersections
A block design is a set family, whose members are called blocks, where every pair of points in the universe is covered by exactly one block. Erdős asked the following question: is it true that if the universe is of size $n$ and each block is of size $\Omega(\sqrt{n})$, then there exists a transversal intersecting every block in $O(1)$ points? We provide a negative answer to this question. Indeed, for arbitrarily large $n$, we construct such a block designs where every transversal intersects some block in $\Omega(\frac{\log n}{\log\log n})$ points.
Read the full abstract
Computation and Language
The Hidden Frame: How Large Language Models Impact Democratic Society
Large language models are rapidly becoming an interface between citizens and political information. They are often regarded as "a better Google." While this analogy might work for some instances, it is unintuitively problematic for democratic politics. A search engine retrieves human-authored documents, while a language model generates novel text that necessarily embeds invisible framing decisions. Because conveying knowledge involves framing, a system that generates answers cannot serve as a neutral conduit to "all human knowledge." Instead, these systems are becoming a new kind of political intermediary. Mechanistic evidence shows that partisan identity is encoded as a locatable geometric direction inside the Llama 3.1 8B model, and that alignment training masks rather than removes this structure. Building on that evidence, we use the model's training cutoff in 2023 to make those frames visible. This cutpoint auspiciously falls just before a dramatic realignment in American politics marked by the second Trump administration and the MAHA transformation of health politics, providing us with a natural experiment. We find that the model presents temporally contingent partisan alignments as knowledge, with no reliable mechanism for distinguishing fact from opinion. Because citizens judge political claims by who made them and when, an answer that carries neither cue hampers the judgment on which self-governance depends. Large language models purporting to summarize "all human knowledge" are, in actuality, simply magnifying the cultural and partisan divides inherent in their training data.
Read the full abstract
Computer Vision
TV-SGS: Gaussian Splatting with Geometric Information Propagation via Tensor Voting under sparse views
Gaussian Splatting has been effective in inferring scene representations that excel in novel view synthesis. Multiple splats cooperate seamlessly to synthesize the pixels of novel views and are jointly optimized even though they only affect each other indirectly, via pixels they project to in common. We present an approach that enables direct communication among splats to enhance the geometric structures they form in 3D. This is accomplished by Tensor Voting, which was originally designed to infer structures from noisy inputs and has been adapted here to provide supervision during test-time optimization, leading to more accurate scene geometry. We introduce a new class of 3D losses that do not rely on rendering and can be combined with essentially all losses previously reported in the literature. Our 3D losses are especially effective when the input views are sparse and geometric regularization is essential due to limited supervision from the images. Our method is easy to integrate with a diverse set of backbones, and our experiments on the DTU and Tanks-and-Temples datasets demonstrate that TV-SGS improves the geometry of the outputs compared to the backbone, while maintaining or improving rendering quality.
Read the full abstract
Optimization and Control
Exact Decision and a Surjective Stabilizer Atlas for Affine Coprime-Factor Segments
We consider simultaneous internal stabilization of square proper real-rational plants of arbitrary but fixed finite input-output dimension $n$, arranged as an affine right-coprime factor segment, under explicit nondegeneracy hypotheses. After a classical normalization, the remaining global spectral-cut constraint and the reconstruction constraints at finite poles and at infinity are encoded, without loss, as a finite-dimensional semialgebraic seed, using a function-level double-Cayley factorization. For input coefficients in an effectively presented real closed field, that seed yields an exact existence decision that does not enumerate controller McMillan degree; the decision engine is classical quantifier elimination. The same encoding, now ranging over rational radii, admissible seeds, and residual unimodular and rational Schur parameters, yields a surjective atlas of all proper real-rational common stabilizers of the segment. The results concern this structured class and are compatible with the rational undecidability of simultaneous stabilization of three general plants.
Read the full abstract
Plasma Physics
Thresholdless IBW Emission and Alpha-to-Thermal-Ion Energy Channeling via Pole Resonance of Fusion-Product Ions
Fusion reactions can release a substantial fraction, and in aneutronic reactions nearly all, of their energy as the kinetic energy of fusion-product ions. In magnetized plasmas, these energetic ions can form ring distributions in velocity space. Such distributions can drive Dory-Guest-Harris (DGH) electrostatic instabilities, but these self-instabilities require a finite energetic-ion density and can be stabilized by a thermal background. We show that the same background can instead enable a distinct instability mechanism: a cyclotron pole of the minority energetic-ion susceptibility destabilizes a stable ion Bernstein wave (IBW) eigenmode. When the energetic-ion harmonic is distinct from the thermal-ion harmonics, exact root-pole resonance is thresholdless in the ideal collisionless limit. A species-resolved power balance shows that the energetic ions supply the free energy while the thermal plasma receives it. In the LAPD proton–alpha example, $99.96\%$ of the alpha-particle power loss enters the coherent proton response. This self-excited, ion-directed transfer provides a possible linear building block for alpha-particle energy channeling in proton-Boron11 fusion.