Computation and Language
Size Matters: Foundation Model for Czech HTML documents
Martin Dvořák, Vít Tlustoš, Artyom Voronin et al. (+4 more)
Creating universal, high-quality representations of web documents in high-traffic industrial environments requires models that are both performant and economic. Existing approaches, however, often depend on large models, overlook the structural information inherent in HTML, or are constrained by short context windows, limiting their ability to process real-world web pages. We present HTML-LM, a compact foundation model with 154 million parameters that addresses these limitations through HTML-aware training and a ModernBERT-based architecture. It was trained on 100 million web documents using multiple objectives, including masked
language modeling, bag-of-words prediction, and contrastive
distillation from large
language models. Consequently, HTML-LM sets a new state-of-the-art for classification and regression applications in the Czech Internet domain, surpassing both larger encoders and small-sized LLMs. The model is deployed in production, processing thousands of web documents per second, and released to the community under the CC BY-NC 4.0. https://huggingface.co/Seznam/html-lm.
Read the full abstract ▾
All authors Martin Dvořák, Vít Tlustoš, Artyom Voronin, Martin Habrovec, Kateřina Podlesná, Barbora Rišová, Josef Vonášek
Computer Vision
Semantic-ITC: A Frame-wise Indoor Mobile Laser Scanning Dataset and Benchmark for Semantic Segmentation
Haiyang Wu, Muhammad Affan, George Vosselman et al. (+1 more)
Semantic labels for indoor mobile
laser scanning (MLS) frames remain largely absent from current point cloud semantic
segmentation benchmarks, which mainly focus on reconstructed indoor scenes or outdoor
LiDAR perception. This paper introduces Semantic-ITC, to the best of our knowledge the first public
dataset and
benchmark for frame-wise indoor MLS semantic
segmentation. The
dataset contains 52 indoor sequences, 79,108 MLS frames, and 1.23 billion labeled points collected in classrooms, corridors, meeting rooms, offices, and study areas. Labels are attached directly to measured
LiDAR points in each frame using 16 semantic classes covering structural elements, furniture, room equipment, vegetation, and other indoor objects. Semantic-ITC preserves the sparse, non-uniform, and frame-wise sampling pattern of indoor MLS, making it distinct from scene-level reconstructed point clouds and mesh-based indoor
datasets. The annotations are produced by a hybrid workflow that combines predictions from a visual foundation model applied to synchronized RGB images, structural information from BIM, and manual refinement, with the final labels assigned to the original
LiDAR frames. A single-frame
benchmark is provided, and the best baseline reaches 79.27% mIoU. Remaining errors are concentrated around object boundaries and ambiguous indoor classes, indicating the challenges of indoor MLS
segmentation under sparse frame geometry and long-tailed class distributions. The
dataset provides a public
benchmark for evaluating semantic
segmentation directly on measured indoor MLS frames and supports future studies on frame-wise indoor MLS semantic
segmentation.
Read the full abstract ▾
All authors Haiyang Wu, Muhammad Affan, George Vosselman, Ville Lehtola
Geometric Topology
Families of knots that cannot be made Legendrian parametrically
Javier Martínez-Aguinaga
The fact that every smooth
knot type admits a Legendrian representative is a classical result in contact
topology. However, the analogous surjectivity question was open at the parametric level. In this work we address the $n>1$ case. We prove that for every $n\geq 3$, every
knot type $\mathcal K$, every Legendrian representative $\mathcal L$ and every formal Legendrian representative $\mathcal{FL}$, the associated group homomorphisms $\pi_n(\mathcal{L})\to\pi_n(\mathcal{K})$ and $\pi_n(\mathcal{FL})\to\pi_n(\mathcal{K})$ are never surjective. We then show that surjectivity at the $\pi_2$-level depends on the
knot type. This work thus proves the presence of
rigidity for parametric families at every higher homotopy level beyond $\pi_1$.
Read the full abstract ▾
Classical Analysis and ODEs
On Kolmogorov's rearrangement problem and Garsia's conjecture
Mark Lewko
We give negative answers to Kolmogorov's rearrangement problem and Garsia's
conjecture. We construct a complete uniformly bounded orthonormal system for which every rearrangement admits a square-summable series divergent almost everywhere. The finite construction uses two copies of the trigonometric system in different orderings. The main ingredient is a combinatorial lemma which guarantees a prescribed ordering along an arithmetic progression in at least one of two related permutations. Its proof uses Szemerédi's
theorem and a counting argument. A Walsh variant gives $N$-term $\{\pm1\}$-valued systems with $L^2$ maximal norm at least $c\log\log N$ in every ordering, matching Bourgain's upper bound.
Read the full abstract ▾
Computer Vision
Learning A Unified Template for Gait Recognition
Panjian Huang, Saihui Hou, Junzhou Huang et al. (+1 more)
"What I cannot create, I do not understand."Human wisdom reveals that creation is one of the highest forms of learning. For example,
Diffusion Models have demonstrated remarkable semantic structure and
memory in image generation, understanding, and restoration, which intuitively benefits representation learning. However, current gait networks rarely embrace this perspective, relying primarily on learning by contrasting gait samples under varying complex conditions, leading to semantic inconsistency and uniformity issues. To address these issues, we propose Origins with generative capabilities whose underlying philosophy is that different entities are generated from a unified template, inherently regularizing gait representations within a consistent and diverse semantic space to capture accurate gait differences. Admittedly, learning this unified template is exceedingly challenging, as it requires the comprehensiveness of the template to encompass gait representations with various conditions. Inspired by
Diffusion Models, Origins diffuses the unified template into timestep templates for gait generative learning, and meanwhile transfers the unified template for gait representation learning. Especially, gait generative and representation learning serve as a unified framework for end-to-end joint training. Extensive experiments on CASIA-B, CCPG,SUSTech1K, Gait3D, GREW and CCGR-MINI demonstrate that Origins performs unified generative and representation learning, achieving superior performance.
Read the full abstract ▾
All authors Panjian Huang, Saihui Hou, Junzhou Huang, Yongzhen Huang
Atmospheric and Oceanic Physics
Butterfly Effect and the Kinetic Energy Cascade in Probabilistic Machine Learning Weather Prediction Models
Jiakai Chen, Joel Oskarsson, Simon Driscoll et al. (+1 more)
This study analyses kinetic energy (KE) spectra, difference kinetic energy (DKE) spectra, and signatures of KE transfer across spatial scales in four state-of-the-art probabilistic machine learning weather prediction (MLWP) models: NeuralGCM-ENS, FourCastNet 3, AIFS-ENS, and GenCast. Results are compared with those from the physics-based numerical weather prediction model IFS-ENS. While NeuralGCM-ENS successfully reproduces the expected upscale transfer of KE, noise injection at its encoder stage underestimates mesoscale KE. Conversely, AIFS-ENS, GenCast, and FourCastNet 3 produce realistic KE spectral magnitudes but do not capture the expected upscale transfer of KE. In particular, AIFS-ENS and GenCast, which employ spatially uncorrelated
stochastic perturbations, exhibit enhanced accumulation of KE at high wavenumbers. All examined models exhibit upscale error growth, reflected by the progressive shift of the DKE spectral peak toward larger wavelengths over time. However, the MLWP models struggle to reproduce the rapid initial growth of ensemble spread at small spatial scales associated with the butterfly effect. The results show that MLWP models can misrepresent the known scale transfer of kinetic energy despite producing skilful weather forecasts.
Read the full abstract ▾
All authors Jiakai Chen, Joel Oskarsson, Simon Driscoll, Sebastian Schemm
Computer Vision
Beyond Random Couplings: Contrastive Noise Alignment in Generative Flows
Lennart Wittke and Vinicius Azevedo
Diffusion and flow-matching models are typically trained by corrupting data through independently sampled Gaussian noise. While simple and scalable, this forward process induces arbitrary data-noise couplings, forcing the network to learn high-
curvature transports between unrelated endpoints. Existing optimal-transport methods reduce this burden by reassigning fixed noise samples to data, but the source noise distribution itself remains passive. To address this, we introduce Contrastive Noise
Alignment (CNA), a training-time method that creates dynamic, contrastive couplings by optimizing the noise representations directly. By modeling the noise batch as an interacting particle system, CNA employs a cross-modal InfoNCE objective to align noise particles with their paired data targets. To prevent spatial collapse, this
alignment is regularized using an angular
entropy term and a radial norm penalty. We show theoretically that this
equilibrium asymptotically preserves Gaussian structures, maintaining tractability during inference. Empirically, CNA improves the
alignment between noise and data, reduces flow
curvature, and provides better generation quality with fewer required sampling steps. For few-step, pixel-space generation (2-4 NFEs), CNA reduces FID by over 50% compared to standard rectified flow, and by at least 24% against Optimal Transport baselines.
Read the full abstract ▾
Robotics
ActionPiece: Rethinking Action Tokenization for Autoregressive Vision-Language-Action Models
Shijie Lian, Bin Yu, Zhaolong Shen et al. (+5 more)
Action tokenizers play a central role in autoregressive
vision-
language-action (VLA) models, determining both the targets for policy training and the executable commands recovered from predicted
tokens. Their fidelity is commonly evaluated using pointwise reconstruction metrics such as mean squared error (MSE), yet small individual errors do not fully characterize how faithfully action adjustments across demonstrations are preserved. After compression, similar actions may still cluster around a representative motion, while the adjustments needed for different contexts are diminished, distorted, or even reversed. We introduce physical rank consistency (PRC) to measure how well tokenization preserves local physical distance rankings after reconstruction. Evaluating decoded actions provides a common reference across
token vocabularies and decoder architectures, complementing pointwise accuracy with a measure of relational fidelity. We further present ActionPiece, which preserves physical action relationships through joint supervision of representation learning and quantization. Physical rank preservation supervises near-far ordering in encoder and quantized feature distances, while quantization regularization applies the same ordering to codeword assignment distributions. Both objectives augment reconstruction, producing discrete action
tokens for standard autoregressive policy learning and execution through a frozen decoder. Under the same Qwen3-VL-4B policy training setup, ActionPiece achieves 94.8% on LIBERO and 68.8% on unseen LIBERO-Plus, with additional evaluations reaching 71.9% on SimplerEnv and 51.5% across VLA-Arena L0-L2. Component ablations show that the two objectives jointly improve PRC and policy success, demonstrating the value of physical relationship supervision for action tokenization.
Read the full abstract ▾
All authors Shijie Lian, Bin Yu, Zhaolong Shen, Xiaopeng Lin, Yichao Du, Zhirui Zhang, Laurence T. Yang, Kai Chen
Logic in Computer Science
Towards a Proof-Theoretic Analysis of Incorrect/Incomplete Proofs
Matthias Baaz and Mariami Gamsakhurdia
We investigate the proof-theoretic structure of incorrect/incomplete proofs, that is, derivations containing syntactic errors or incomplete inferential steps that nonetheless preserve partial semantic validity. Building on
Hilbert's epsilon calculus, we formalize how such derivations can be corrected through semantic projection and weakest preconditions, leading to valid Herbrand disjunctions. We show that the epsilon calculus provides a natural framework for analyzing tolerance of falsity in proofs and for identifying conditions under which an incorrect proof can be semantically repaired. This approach extends
Hilbert's program beyond correctness, toward a logic of error and recovery. Moreover we show that the extended first epsilon
theorem is false-tolerant.
Read the full abstract ▾
Mesoscale and Nanoscale Physics
Zhanning Wang, Yongtao Li, Nelson E. Rivas et al. (+5 more)
Gate-defined hole
spins combine strong
spin-orbit coupling with exceptional strain sensitivity, making surface acoustic waves a natural route to remote, phase-coherent control. We show that counterpropagating surface acoustic waves can differ in coupling strength and in whether they predominantly drive
spin rotations or modulate the
qubit frequency. We call the transverse transition-strength imbalance
spin-acoustic chirality and show that it is tunable through gate-controlled reshaping of the dot confinement. Combining a multiband Luttinger-Kohn
Hamiltonian with the Bir-Pikus description of strain and 3D piezoelectric finite-element simulations, we obtain the acoustic g-
matrix modulation governing coherent SAW driving. Quantizing the same strain-mediated interaction yields the corresponding single-
phonon coupling, and a generalized anisotropic Rabi interaction with co-rotating, counter-rotating, and longitudinal components, linking classical chirality to
quantum anisotropy and directional one-
phonon emission.
Magnetic field orientation and dot shape tune the acoustic chirality and operator selectivity. We illustrate this framework through bichromatic coherence protection and
phonon-mediated heralded Bell-state initialization.
Read the full abstract ▾
All authors Zhanning Wang, Yongtao Li, Nelson E. Rivas, Gonzalo García, Irene Castro, Rubén Seoane Souto, Daniel Ramos, José C. Abadillo-Uriel
Computational Complexity
Rational Reductions and Regular Languages of Constant Circuit Complexity
Stefan Göller and Amaldev Manuel
We study the
circuit complexity of regular
languages in terms of unbounded fan-in Boolean
circuit families. We characterize the regular
languages of constant
circuit complexity in terms of the one-variable fragment of first-order logic with regular predicates, in terms of the pseudovariety of stamps $\mathbf{QEJ}_\mathbf{1}$, suitable word congruences and regular expressions. We analogously characterize the neutral letter regular
languages of constant
circuit complexity. Our lower bound result implies that the class of regular
languages of sublogarithmic
circuit complexity coincides with the one of constant
circuit complexity. In addition we show that deciding whether a regular
language, given as a nondeterministic finite automaton, has constant
circuit complexity is $\mathbf{PSPACE}$-complete. We introduce a strong notion of reduction, called rational truth-table reduction, that is tailored towards algebraically defined classes of
languages. We show that, for a class of functions we call mild, rational truth-table reductions preserve both upper and lower bounds on
circuit complexity. We show that the class of regular
languages, whose
circuit complexity is bounded by a mild function, is in fact a length-multiplying variety of
languages. Slightly extending the class of regular
languages of constant
circuit complexity, we analogously characterize the class of regular
languages that are in the pseudovariety $\mathbf{QEACom}$. For these we derive logarithmic
circuit complexity upper bounds.
Read the full abstract ▾
Human-Computer Interaction
EasyFashion: A Human-AI Co-Creation System for Personalized Fashion Design and Sewing Pattern Generation
Hong Qu, Zhaoxiang Xu, Jinbo Luo et al. (+3 more)
People often want garments that reflect their aesthetic preferences, fit their bodies, and meet their sizing needs, yet turning these requirements into physical garments remains difficult. Ready-to-wear options provide limited personalization, while custom tailoring is costly and time-consuming. Recent generative artificial intelligence (AI) systems can visualize garment ideas but often stop short of supporting downstream production. To address this gap, we present EasyFashion, a human-AI co-creation system that enables users to iteratively refine design intent for personalized garment style and size, evaluate designs through virtual try-on on reconstructed personal avatars, and generate sewing patterns for garment production. Using reference images, text descriptions, and body photos as input, EasyFashion translates user intent into structured garment specifications and try-on results. Technical experiments, user studies, and a real-world production case demonstrate the value of EasyFashion for
multimodal design expression, body-specific evaluation, and production-oriented outputs in personalized garment design. Our project page is available at https://zoequ.github.io/easyfashion.github.io/.
Read the full abstract ▾
All authors Hong Qu, Zhaoxiang Xu, Jinbo Luo, Yujie Zhao, Jie Zhang, Yadie Yang
Robotics
Real-Time Bounded Catenary Solver for UAV Tether Modeling
Max Beffert and Andreas Zell
For non-stationary tethered multirotor
UAVs in real-world conditions, simulating the forces imposed on the drone by the aerodynamic drag of the tether becomes crucial, with online use cases placing a hard bound on the maximum solve time. In previous work, a quasi-analytical catenary tether model reached a mean solve time of 0.51 ms using a general-purpose root finder, but without any worst-case guarantees or proven convergence. In this work, we reformulate the inner solver by reducing the catenary boundary-value problem to a single transcendental equation in one well-conditioned unknown. We derive a closed-form bracket and prove monotonicity and convexity as well as existence and uniqueness of the root, which together guarantee convergence of the solver. We further propose a two-regime initial guess which approximates the true root within 3.4 % and reduces the mean iteration count by 69.2 % to 2.24 compared to the textbook initialization. Building on the hybrid root-finding method rtsafe (Newton-Raphson with bisection fallback giving bounded iteration counts), we implement a specialized variant that exploits the problem structure to omit unnecessary checks while retaining correctness, which gives up to 1.37 times speedup. With the proposed solver the full tether model achieves a nearly constant solve time of 4.0 us on average and 4.6 us at worst, a 68 times speedup over an optimized re-implementation of the previous method, while agreeing with it to a relative deviation of 8.7e-9. Because the reformulation leaves the underlying physical model untouched, the experimental validation of the previous work carries over unchanged. We further demonstrate its suitability for embedded, resource-constrained platforms with a Lua implementation running directly in ArduPilot on a drone's flight controller, where it stays well inside the scheduling budget with a mean solve time of 0.74 ms.
Read the full abstract ▾
Artificial Intelligence
Hyperbolic Graph Representation Learning for Differential Diagnosis on Biomedical Knowledge Graphs
Pietro Miotto, Lucia Mellini, Tommaso Marzi et al. (+5 more)
Biomedical knowledge graphs combine ontology-derived hierarchies with transversal associations among heterogeneous entities such as phenotypes, diseases, genes,
proteins, and patients. This hybrid structure raises the question of whether
hyperbolic embeddings, which naturally capture tree-like organization, remain useful beyond purely hierarchical graphs. We present a preliminary study of
hyperbolic graph representation learning for Mendelian-disease differential diagnosis on a patient-integrated biomedical graph. Experiments on isolated ontology subgraphs show that
hyperbolic models achieve strong performance in substantially lower dimensions than Euclidean baselines. We then evaluate the models on a link-prediction task that ranks candidate diseases for each patient. Results suggest that
hyperbolic embeddings can exploit biomedical hierarchical structure while supporting diagnostic
reasoning over heterogeneous patient-level graphs.
Read the full abstract ▾
All authors Pietro Miotto, Lucia Mellini, Tommaso Marzi, Cesare Alippi, Elena Casiraghi, Alberto Paccanaro, Giorgio Valentini, Mauricio Soto-Gomez
Number Theory
Cyclotomic Prime Extractors
Joseph M. Shunia
We develop explicit
prime-recovery formulas from the values $\Phi_n(2)$ of cyclotomic
polynomials. Binary divisibility patterns detect repeated
prime factors and identify the least
prime divisor of a squarefree index, while small corrections to $\log_2\Phi_n(2)$ allow successive recovery of the distinct
prime factors. Specializing the index gives identities for
prime products and the least
prime above a given integer. The same mechanism extends from finite factorizations to an infinite
prime sequence: a normalized limit of cyclotomic values along the odd primorials defines a real constant $\Omega=0.25061403238015047218\ldots$, from which every odd
prime can be recovered by a recursive rounding rule.
Read the full abstract ▾
Human-Computer Interaction
Verify, Offload, Extend & Recommend: Selective Complementarity in AI Support for Physical Activity Planning with Longitudinal Patient Data
Pavithren V S Pakianathan, Rania Islambouli, Diogo Branco et al. (+5 more)
Self-tracking technologies create longitudinal patient-generated health data, yet integrating these data into clinical decision-making can increase information-processing demands. Generative AI may support sensemaking, but its value depends on clinical context and expertise. We investigate AI augmentation of a clinical decision support system for physical-activity planning in cardiovascular disease. In a counterbalanced within-subjects study, 26 exercise physiologists developed plans for four real cardiovascular cases with and without AI support, followed by evaluation of an AI exercise-plan generator. AI did not significantly improve workload, usability, confidence, or plan quality overall; however, its effect on plan quality increased as visualization literacy decreased and its effect on workload increased as visualisation literacy increased. Interviews and 152 chatbot queries revealed three recurring uses: verifying, offloading, extend and generate. Our findings position AI support as a selective complement to professional expertise while highlighting validation challenges when clinicians seek support precisely where their own knowledge is limited.
Read the full abstract ▾
All authors Pavithren V S Pakianathan, Rania Islambouli, Diogo Branco, Gil Batista Rosa, Rita Pinto, Albrecht Schmidt, Tiago Guerreiro, Jan David Smeddinck
Optics
350-GHz-Band 4 by 4 RTD Monostatic Radar Array for Sequential Multidirectional Ranging
Li Yi, Ryoma Nakamura, Shota Ito et al. (+6 more)
Terahertz (THz) sensing offers significant potential for nondestructive evaluation, imaging, and metrology, but the cost and complexity of existing systems remain barriers to practical deployment. This letter presents a compact 4 by 4 monostatic radar array based on a single-RTD-per-pixel architecture, in which each resonant tunneling diode (RTD) functions as both a bias-tunable oscillator and a self-mixing detector. The RTD elements are sequentially addressed through a low-frequency switching network and share the same baseband control and readout electronics. A shared 3D-printed dielectric lens maps the RTD elements to spatially separated sensing directions. We experimentally verify the operation of all array elements and demonstrate sequential multidirectional ranging using four selected elements, followed by preliminary 16-pixel THz imaging. By combining self-mixing at each pixel with low-frequency selection and shared optics, the array avoids a separate THz receiver chain for each pixel and provides a compact architecture for electronically addressable THz sensing and imaging.
Read the full abstract ▾
All authors Li Yi, Ryoma Nakamura, Shota Ito, Yousuke Nishida, Koji Terumoto, Toshihisa Maeda, Bryce Chung, Yunbin Jiang, Daniel Headland
Cryptography and Security
A Global Readiness and Sovereignty Capability Model for Post-Quantum Cryptography Migration
Mohamed Aly Bouke
Cryptographic dependence predates the
quantum era, but the migration to post-
quantum cryptography (PQC) opens a rare window to reshape it, because the
algorithms, implementations, hardware, and standards adopted now can lock in dependence or sovereignty for decades. This paper introduces the Readiness-Sovereignty Capability Model (RSCM), a national measurement model that operationalizes PQC readiness together with cryptographic sovereignty, which current maturity models score only as readiness and the sovereignty literature defines without measuring. RSCM decomposes sovereignty into three distinct constructs, indigenous cryptographic capacity, indigenous post-
quantum control, and external dependency, and certifies a post-
quantum maker only through a gate requiring demonstrated, institutionally sustained creation in at least one core layer, whether design, implementation, or validation. Applying it to fifty-seven documented cryptographic actors coded from cited public evidence, and testing that coding with an independent second coder, a plausible-state bootstrap, and convergent-validity checks, we find that twenty countries clear the gate, fifteen as full-stack makers and five as research makers, eleven hold strong general capacity without post-
quantum control, one is a ready adopter, and twenty-five are dependent. The gate
cells show substantial weighted agreement, a quadratic-weighted kappa of 0.71, and the maker classification is stable in its core though uncertain at the threshold. Readiness tracks independent cyber indices at rank correlations up to 0.70, while post-
quantum creation shows no significant correlation with the commitment index, a rank correlation of only 0.22 that separates control from readiness. The paper contributes the framework, the evidence-graded assessment, and policy directions for building indigenous
quantum-safe capacity.
Read the full abstract ▾
Combinatorics
Finite and infinite barrycades
Michał Dębski, Jarosław Grytczuk, Paweł Naroski et al. (+3 more)
An $n$-barrycade of height $h$ is a set of $h$ permutations of $[n]$ such that all the prefix sums are different. This notion was described in 2020 by Richard Guy, together with the central problem to determine for which values of $n$ there exists a break-free $n$-barrycade, that is, one in which every possible prefix sum from $1$ to $\frac{n(n+1)}{2}-1$ is covered. The height of such barrycade would be $\frac{n+2}{2}$. We prove that barrycades of linear height exist for infinitely many sizes: there is a constant $c>0$ such that for infinitely many positive integers $n$ there exists an $n$-barrycade of height at least $cn$. We also explore an infinite variant of the problem and propose three constructions – greedy, grasshopper and precise grasshopper – that give increasingly more satisfying results. Along the way we encounter new integer sequences and formulate
conjectures concerning omitted elements, missing partial sums, and word representations of infinite barrycades, which are firmly supported by computational data.
Read the full abstract ▾
All authors Michał Dębski, Jarosław Grytczuk, Paweł Naroski, Bartłomiej Pawlik, Jakub Przybyło, Małgorzata Śleszyńska-Nowak
General Relativity and Quantum Cosmology
Second-order slow rotation of wormholes in Einstein-scalar-Gauss-Bonnet gravity
Sardor Murodov, Bekzod Rahmatov, Javlon Rayimbaev et al. (+1 more)
We investigate slowly rotating traversable wormholes in Einstein–scalar–Gauss–Bonnet
gravity beyond the linear rotation regime. The static backgrounds are obtained without expanding in the Gauss–Bonnet coupling, while rotational effects are treated perturbatively up to second order. Both monopolar and quadrupolar deformations are included, allowing us to determine the
spin-induced changes in the asymptotic charges, throat geometry, energy-condition behavior, and mass quadrupole. The quadrupolar equations develop an internal regular singular point, which is handled by selecting the regular local solution through a Frobenius analysis. The rotation-induced response of the throat source is incorporated consistently through the junction conditions. Independent shooting and collocation calculations of the quadrupolar sector give mutually consistent solutions. The rotating throat remains oblate throughout the fixed-coupling family, while at $\chi=0.15$ the smooth-bulk near-throat radial null-energy-condition violation persists but its radial extent becomes latitude dependent. These results provide a self-consistent second-order description of rotating Einstein–scalar–Gauss–Bonnet wormholes.
Read the full abstract ▾
All authors Sardor Murodov, Bekzod Rahmatov, Javlon Rayimbaev, Bobomurat Ahmedov
Combinatorics
A somewhat sure note on an un-Schur problem
Swaroop Hegde, Hitesh Kumar and Pratibha
Parczyk and Spiegel initiated the study of an anti-Ramsey multiplicity variant of Schur's
theorem and proved that the maximum fraction of Schur triples that can be rainbow in a $3$-coloring of $\{ 1, \dots ,n \}$ is bounded asymptotically between $0.4$ and $0.66364$. Furthermore, they conjectured that their lower bound is optimal. We disprove this
conjecture and prove new bounds. In particular, we show that the maximum fraction of rainbow Schur triples that can be rainbow in a $3$-coloring of $\{ 1, \dots ,n \}$ lies between $9/22$ and $8/15$ asymptotically. Moreover, we study the problem in the general $k$-color setting and establish new non-trivial bounds.
Read the full abstract ▾
Computer Vision
CADSplat: Sparse-View 3D Gaussian Splatting Aided by CAD Models for Robust, Photorealistic Digital-Twin Reconstruction
Kristof Overdulve, Lode Jorissen and Nick Michiels
We present CADSplat, a framework that reconstructs photorealistic, geometrically accurate digital twins from sparse ($<15$ views), wide-baseline posed images of an object by regularizing 3D Gaussian Splatting (3DGS) with an explicit CAD shape prior. Using such a prior requires finding a CAD model whose shape resembles the object depicted in the images and determining the pose of each camera relative to the object. We obtain both by matching segmented object silhouettes against silhouettes rendered from a CAD library and keeping the camera-to-object poses of the best-matching model. We then anchor 3D Gaussian primitives to the surface of the retrieved model and jointly optimize the 3DGS parameters, the camera-to-object registration, and a non-rigid deformation field to account for shape differences between the physical object and the CAD model. Across two real-world
datasets, CADSplat outperforms unconstrained, few-shot, and mesh-texturing baselines and degrades gracefully to as few as 3 views. Our experiments show that most of the gain in rendering quality comes from how the splats are constrained—a fixed set of splats tied to a surface and moved by a single smooth deformation field—rather than from the CAD shape itself. The CAD model adds shape knowledge where views are scarcest, in the sparsest captures and on strongly self-occluded objects, and it places every camera in the object's own frame. This enables applications beyond novel-view synthesis, such as markerless augmented reality registration, per-image object pose estimation, physical simulations, and the transfer of part labels from the design to the reconstruction.
Read the full abstract ▾
Metric Geometry
Uncentered Blaschke-Santaló inequalities for the Gaussian measure
S. Artstein-Avidan, M. Fradelizi and K. Wyczesany
We study the maximizers of the generalized volume product \[ \gamma_\sigma^n(A)\,\gamma_\sigma^n(A^\circ) \] among all measurable subsets $A\subset\mathbb{R}^n$, where $A^\circ$ denotes the polar set of $A$, and where $\gamma_\sigma^n$ denotes the centered Gaussian probability measure on $\mathbb{R}^n$ with covariance $\sigma^2 I_n$, $\sigma>0$. It turns out that the maximizers depend on $\sigma$. We prove that they exist and are convex bodies. In dimension $n=1$, we find the exact form of the maximizers. In dimension $n\ge 2$, we show that they are smooth bodies of revolution whose support function satisfies a certain differential equation. Moreover, for $\sigma^2 \le \frac{1}{{n}}$ we show that the Euclidean unit ball is the unique maximizer, while this is no longer the case for $\sigma^2\ge {\frac{2}{n+1}}$. In dimension $n=2$, we close the gap by showing that the Euclidean unit ball is the unique maximizer for $\sigma^2 \le \frac{2}{3}$.
Read the full abstract ▾
Artificial Intelligence
First Token Matters: Understanding Safety Collapse in Large Reasoning Models
Yizheng Yang, Haining Yu, Yuechen Wang et al. (+4 more)
Large
Reasoning Models (LRMs) exhibit strong problem-solving abilities, yet their safety
alignment often degrades when handling harmful queries. Existing approaches to improving safety largely rely on additional training or preference optimization, while offering limited understanding of the internal mechanisms behind safety failures. In this work, we investigate this failure through a
token-level positional analysis of refusal dynamics and identify a localized vulnerability at the onset of
reasoning, which we term Onset Refusal Collapse (ORC). We find that the refusal-related signal of LRMs drops sharply at the first generated
token under harmful queries, which is associated with unsafe response generation. Motivated by this finding, we propose SafeToken, a lightweight inference-time intervention that injects a learned continuous safety anchor precisely at
reasoning onset. Despite updating only a single
token embedding, SafeToken effectively mitigates ORC, improves safety on harmful-query
benchmarks, and largely preserves
reasoning utility. These results suggest that safety failures in LRMs can arise from a transient breakdown at the critical transition from understanding to generation.
Read the full abstract ▾
All authors Yizheng Yang, Haining Yu, Yuechen Wang, Yikai Hou, Xing Fu, Jinbo Yang, Tianqing Zhu
Multimedia
Divide and Conquer: Mixture-of-Bottleneck Experts in Informative Ordinal Space for Video-based Multimodal Sentiment Analysis
Ronghao Lin, Qiaolin He, Zefeng Lu et al. (+5 more)
Video-based
Multimodal sentiment analysis (MSA) must handle information from text, audio, and image sequence in human speaking
videos, yet current methods often fail to integrate modalities with task awareness. Most models treat
video sentiment prediction as a single task, overlooking its ordinal nature, and their fusion strategies struggle to capture diverse unique and synergic cues across modalities. To address these limitations, we adopt a divide-and-conquer perspective by reformulating MSA as an ordinal regression problem and decoupling it into polarity recognition and intensity prediction. Driven by information theory, we introduce a Mixture-of-Bottleneck (MoB) framework that assigns different latents to polarity- and intensity-specific experts for different modalities. With the learning of information bottleneck, each expert learns compact and task-relevant representations while filtering out redundancy and noise. A
multimodal bottleneck routing fusion module then fuses these expert latents with hard mining strategy, guiding the prediction in the ordinal sentiment space. Extensive experiments on 4 MSA
datasets and 4
language models show that MoB effectively leverages informative latents from diverse modalities and captures general sentiment structure. Beyond stronger performance, MoB comprehensively captures fine-grained intra- and inter-modal dynamics, enabling more trustworthy localization of nuanced
video sentiment signals.
Read the full abstract ▾
All authors Ronghao Lin, Qiaolin He, Zefeng Lu, Yichu Liu, Li Huang, Sijie Mai, Haifeng Hu, Yap-peng Tan
Combinatorics
Antidirected forests in digraphs
Gengtao Liu and Yunshu Gao
A digraph is antidirected if every vertex has indegree zero or outdegree zero. Let $k\ge2$, and let $F$ be an antidirected forest with $k$ arcs and no isolated vertices. We prove that every digraph $D$ of order $n$ with more than $g_k(n):=2\max\left\{\binom{2k-1}{2}, (k-1)\left(n-\frac{k}{2}\right)\right\}$ arcs contains $F$ as a subdigraph. For $n\ge2k-1$, this threshold equals $2\mathrm{ex}(n,kK_2)$ and is attained by symmetric digraphs arising from extremal $kK_2$-free graphs. Consequently, the maximum directed extremal number over all such forests is $2\mathrm{ex}(n,kK_2)$. The proof combines a counting inequality for rooted antidirected forests,
embeddings extending vertex-disjoint arcs, and vertex deletion. In the remaining case, the Gallai–Edmonds decomposition of the underlying graph gives the required bound on the number of arcs.
Read the full abstract ▾
Atomic Physics
Improved systematic uncertainty evaluation of the $^{171}$Yb optical lattice clock NMIJ-Yb1 with uncertainty of $2.6\times10^{-17}$
Takumi Kobayashi, Akiko Nishiyama, Ikuhiko Saito et al. (+4 more)
We report a systematic uncertainty evaluation of the $^{171}$Yb optical
lattice clock NMIJ-Yb1 (NMIJ: National Metrology Institute of Japan) with a fractional frequency uncertainty of $2.6\times10^{-17}$, improved by a factor of 3.8 compared with our previous uncertainty. The uncertainty of the
lattice light shift is reduced to $1.6\times10^{-17}$ by preparing axially sideband-cooled atoms in a shallow optical
lattice. Different models to calculate the
lattice light shift are compared, taking account of possible model-dependent biases due to the treatment of the radial motion of atoms trapped in the optical
lattice. The uncertainty of the blackbody radiation shift reaches $7.0\times10^{-18}$ by measuring the radiative temperature at the position of atoms with an in-
vacuum temperature sensor. We also demonstrate the nearly continuous operation of NMIJ-Yb1 with an uptime of 91.3 $\%$ for 10 days, showing the potential for future improvement of the calibration uncertainty of International Atomic Time.
Read the full abstract ▾
All authors Takumi Kobayashi, Akiko Nishiyama, Ikuhiko Saito, Daisuke Akamatsu, Akio Kawasaki, Shintaro Nagase, Masami Yasuda
Signal Processing
Massive MIMO ISAC Under Target-Angle Uncertainty: CRLB Outage Analysis and Robust Resource Allocation
Smriti Uniyal, Tianyu Fang, Van-Dinh Nguyen et al. (+3 more)
In integrated sensing and communications (
ISAC), the same spectral and hardware resources are shared for two functionalities. Most
ISAC designs assume perfect target-angle information neglecting angle estimation errors, which introduce steering-vector mismatches, degrade sensing accuracy, and may invalidate deterministic sensing guarantees. This paper investigates monostatic massive multiple-input-multiple-output (
MIMO)
ISAC systems under imperfect target-angle estimates. We derive closed-form expressions for the Cramér-Rao lower bounds (CRLBs) of target azimuth and elevation estimates in the presence of angle uncertainty. We characterize the cumulative distribution functions and outage probabilities of the CRLBs under Gaussian, generalized uniform, and von Mises angle-error models. Our analysis reveals that, in the small-error regime, the CRLBs increase quadratically with the angle errors due to transmit steering-vector mismatch. To ensure reliable sensing, we propose a robust power allocation framework that jointly optimizes pilot training and communications/sensing transmission powers to maximize the communications sum rate while satisfying CRLB outage constraints. The resulting nonconvex problem is solved using an alternating-optimization
algorithm based on successive convex approximation. Numerical results validate the developed analysis and show that the proposed robust design reduces azimuth and elevation CRLB outage probabilities by up to $60\%$ compared with conventional non-robust schemes. It attains up to $45\%$ higher sum rates than the non-robust design under strict CRLB thresholds.
Read the full abstract ▾
All authors Smriti Uniyal, Tianyu Fang, Van-Dinh Nguyen, Hien Quoc Ngo, Markku Juntti, Nhan Thanh Nguyen
Machine Learning
Spatially Adaptive Noise Injection
Frantzeska Lavda, Maciej Falkiewicz, Van Khoa Nguyen et al. (+1 more)
Diffusion samplers reverse a learned noising process using either
stochastic (DDPM) or deterministic (DDIM) updates, which represent endpoints of a single family controlled by a scalar noise-injection variance that is applied identically at every spatial location. This uniform approach neglects the geometry of natural images: high-
curvature regions such as edges and textures, where the denoiser is uncertain, benefit from
stochastic correction, whereas smooth regions, where the score is precise, are degraded by injected noise. This work investigates whether each pixel requires
stochastic correction at a given timestep and introduces Spatially Adaptive Noise Injection (SANI), a novel sampling framework that dynamically adjusts noise application on a per-pixel basis. SANI integrates a probabilistic gating mechanism with a derived spatially adaptive variance, ensuring that noise is injected precisely where needed to refine complex features while preserving well-formed structures. Experimental results and decoupling ablations demonstrate that SANI consistently improves Fréchet Inception Distance (FID) over the vanilla DDPM and DDIM endpoint samplers across diverse sampling timesteps, while remaining competitive with variance-learning baselines, highlighting the importance of spatial adaptivity in
diffusion sampling.
Read the full abstract ▾
All authors Frantzeska Lavda, Maciej Falkiewicz, Van Khoa Nguyen, Alexandros Kalousis
Computer Vision
GeoCond: A Conditioning-Aware Reliability Adapter for Feed-Forward 3D Reconstruction
David Ahmedt-Aristizabal, Mohammad Ali Armin, Russell Tsuchida et al. (+1 more)
Feed-forward 3D foundation models such as VGGT predict cameras, depth, and point maps in a single pass, but can fail silently under low overlap, low parallax, and extreme relative rotation. Stratified analyses over these factors show that these failures are governed by
geometric conditioning and are poorly captured by native aleatoric confidence. We introduce GeoCond, a lightweight reliability adapter for frozen feed-forward 3D backbones. GeoCond reads
geometric conditioning cues and the backbone's own stability signals and outputs pose-level uncertainty and a refinement gate. During training, it can be supervised by frame-permutation orbit variance, ground-truth pose error when labels are available, or cycle residuals from unlabelled independent pose graphs. At inference, the default head requires only one backbone pass and a small MLP. On VGGT, GeoCond improves out-of-distribution (OOD) AUSE (area under the sparsification-error curve; lower is better) from $0.32$ to $0.20$ over native confidence on held-out outdoor scenes (and to $0.21$ when trained on indoor data only) and avoids the collapse caused by applying bundle adjustment uniformly, without improving on feed-forward poses there. For three of six backbones, cycle-distilled variants provide a ground-truth-free adaptation route, including cases where permutation variance vanishes on equivariant models. The same reliability signal supports gated refinement, pose-graph weighting, calibration, curation, and capture decisions. Reliable feed-forward 3D reconstruction requires not only predicting geometry, but also knowing when that geometry should be trusted.
Read the full abstract ▾
All authors David Ahmedt-Aristizabal, Mohammad Ali Armin, Russell Tsuchida, Lars Petersson