Peer Revenue Play credits. Real papers.

Papers

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

All of arXiv

Wednesday 9 September

Machine Learning

In Medical Claims Data, Enhancing Predictive Performance for Major Adverse Cardiovascular Events Using Cross Attention

Medical claims data comprise the financial details, including the expenses and billing information, as well as the clinical information, such as the diagnoses and treatments, of patients visiting medical facilities. Recently, it has been acknowledged that large databases can be constructed from medical claims data for medical research purposes. However, the clinical information within these datasets is often medically unstructured, limiting its application in comprehensive analyses. This study enhances predictive model performance for major adverse cardiovascular events (MACE), a leading cause of death worldwide. Models that predict MACE are crucial to clinical practice guidelines. We utilize a cross-attention mechanism to develop a method that effectively weights the relationships between diagnoses and treatments. Effectively repre- senting the clinical information contained in medical claims data, this approach generates more representative features for predicting MACE. The ROC-AUC score of our proposed cross-attention-based model was 0.7720, higher than other benchmark models including the conventional atherosclerotic cardiovascular disease model, the light gradient boosting machine, and a self-attention-based model. These results indicate that integrating the clinical structure of medical claims data using a cross-attention mechanism significantly enhances the performance of predictive models.
Read the full abstract

All authors Yuhei Fujioka, Daitaro Misawa, Tatsuyoshi Ikenoue, Shingo Fukuma

Hardware Architecture

AMEND: Audited Margins Enable Nonblocking Drops in GPU-PIM LLM Decoding

Autoregressive large language model (LLM) decoding re-reads a growing key-value (KV) cache at every step, so long-context attention is bound by graphics processing unit (GPU) memory bandwidth. Block-sparse attention skips low-contribution KV blocks, but a selector that decides after the current query-key (QK) product, such as max-relative block thresholding (BLASST), still reads every K block, and a processing-in-memory (PIM) filter that decides from the current query places a serial PIM stage on the critical path. We present AMEND, a GPU-PIM attention design that removes both dependencies. AMEND predicts each block's BLASST verdict from margins audited at earlier steps, so the GPU fetches only predicted survivors while near-bank PIM units in high-bandwidth memory (HBM) concurrently score the omitted complement. A stack-level controller merges both observations, updates the predictor, and eagerly generates the next step's mask, so every predicted drop is re-observed without blocking the current token. Operating points are selected offline by constrained Bayesian optimization under a false-drop budget. Across LongBench and RULER runs, AMEND preserves near-baseline task quality; in simulation at batch size 8, it achieves $1.40$-$3.63\times$ end-to-end decode speedup and 28-66% lower dynamic decode energy than dense attention across 8K-64K contexts.
Read the full abstract

All authors Zuxiong Tan, Will Wei-Jen Wang, Wei Shao, Ali Karkehabadi, Houman Homayoun, Avesta Sasan

Solar and Stellar Astrophysics

Validity of the CRD limit for modeling scattering polarization in the photospheric Sr I 4607 Å line

Context. Scattering polarization in the Sr I 4607 Å line is a key diagnostic for small-scale, unresolved magnetic fields in the quiet solar photosphere, inaccessible to Zeeman-based techniques. The complete frequency redistribution (CRD) limit is commonly used to model this line, but its validity and impact on magnetic sensitivity have not been systematically investigated. This is timely given new facilities observing this line with unprecedented accuracy, and future synoptic programs targeting long-term quiet-Sun magnetism. Aims. We assess the CRD limit against the general partial frequency redistribution (PRD) description of scattering for the Sr I 4607 Å polarization, focusing on its magnetic sensitivity via the Hanle effect. Methods. We solved the non-LTE radiative transfer problem for polarized radiation in a semi-empirical 1D plane-parallel static atmosphere, with CRD and PRD scattering, across magnetic field configurations relevant for Hanle diagnostics. Results. For small-scale unresolved fields, where the field acts only via Hanle depolarization of Q/I, CRD and PRD emergent profiles and Hanle sensitivity are essentially identical. For deterministic fields, CRD remains accurate when polarization signals exceed about 0.3%; for weaker signals, PRD effects can appreciably affect the emergent profiles. Conclusions. CRD is adequate for modeling Sr I 4607 Å scattering polarization in the vast majority of observationally relevant cases; PRD effects matter only for weak polarization signals.
Read the full abstract

All authors Simone Riva, Gioele Janett, Franziska Zeuner, Luca Belluzzi, Fabio Riva, Pietro Benedusi

Graphics

InstantMimic: A High Performance System for Learning Physics-based Skills in Seconds

Physics-based character control is a long-standing challenge in computer graphics and robotics, requiring policies that satisfy complex dynamics while producing realistic motion. Recent Deep RL approaches, particularly imitation learning methods such as DeepMimic, have had broad impact beyond animation, influencing robotics by enabling agile and expressive behaviors. While these approaches achieve impressive results, they remain computationally inefficient to train in practice. Despite GPU-accelerated simulation, we find that end-to-end pipelines often underutilize hardware due to overheads outside the physics solver, caused by fragmented GPU kernels and CPU memory access in the critical path. We present InstantMimic, a system that addresses these inefficiencies by making the entire training loop GPU-native. Built on a GPU-native physics backend, our unified pipeline integrates simulation, environment computation, policy inference, and policy updates within a single execution flow. As a result, InstantMimic reduces training time for diverse physics-based skills to a few seconds and makes LLM-agent-driven hyperparameter search practical.
Read the full abstract

General Relativity and Quantum Cosmology

Clock-noise propagation and calibration for phase-locking configurations in space-based gravitational-wave detectors

In space-based gravitational-wave detectors, laser phase locking transfers the phase of weak received light to a local laser and maintains the interspacecraft heterodyne frequencies within the phasemeter bandwidth. Fluctuations of the onboard ultrastable oscillators introduce clock noise into these heterodyne phase measurements, requiring additional calibration in time-delay interferometry (TDI). Existing formulations based on six independently formed one-way measurements do not explicitly describe the common-reference clock structure of a master–slave phase-locking configuration. In this work, we formulate clock-noise propagation and calibration directly for this configuration. Starting from carrier and sideband readouts, we derive the clock-noise couplings and construct an ordered sideband-calibration rule. The rule retains noncommuting delays and applies to two-branch TDI combinations satisfying propagation and clock-coefficient closure. For mutually independent onboard clocks with identical fractional-frequency noise spectra, we compare matched realizations with and without phase locking. In the adopted frozen equal-arm model, 45 second-generation combinations with up to 16 links considered here have smaller clock-noise residuals in the phase-locking realization throughout 0.1–10~mHz, before final clock calibration. Numerical simulations of two representative 16-link combinations reproduce the analytical spectra and verify the predicted reduction and calibration. These results show that the phase-locking realizations reduce the clock-noise contribution before final calibration, while the cancellation of master-laser phase noise still relies on TDI.
Read the full abstract

Algebraic Topology

The hyperbolic class and vanishing laws in a TMF-valued four-manifold invariant

Under explicit hypotheses on the spectral Looijenga construction, we compute the hyperbolic-plane value of the zero-section invariant of Gukov, Krushkal, Meier, and Pei in periodic topological modular forms. The value is the Hopf element eta, and adjoining a hyperbolic plane acts by multiplication by this element. Earlier work obtains the hyperbolic value under an additional cobordism-duality assumption; here we derive it directly from the Looijenga restriction maps without that assumption. Together with annihilation and vanishing results for definite lattices, this gives a complete value formula on smooth closed simply connected spin four-manifolds. In particular, nonzero signature forces the invariant to vanish, so both orientations of a K3 surface have value zero.
Read the full abstract

Quantitative Methods

Multi-Task Bacterial Colony Detection and Classification Using YOLOv8 with Edge Optimization for Resource-Constrained Deployment

Manual counting and classification of bacterial colonies are critical yet labor-intensive tasks in microbiology, prone to human error particularly on densely populated plates. This work proposes a multi-task deep learning framework trained on the Annotated Germs for Automated Recognition (AGAR) dataset (18,000 images; 9,202 training / 3,067 testing) to automate Colony Forming Unit (CFU) enumeration and species classification. A custom multi-task CNN employing global regression served as the baseline, but demonstrated limited performance in clustered colony environments due to the absence of spatial localization. To address this, a YOLOv8 object detection architecture was adopted with high-resolution 1024x1024 inputs, enabling instance-level colony detection and label assignment. The model achieved a classification accuracy of 98.13% and a counting accuracy of 98.27% (within a 10-colony margin), demonstrating strong predictive capability. To bridge the gap between model performance and practical deployability, the trained model was optimized through unstructured and structured pruning, ONNX conversion, and reduced-precision inference (FP32, FP16, INT8). On a Raspberry Pi 4B, ONNX FP32 and FP16 variants offered the best balance between inference speed (~6.4s) and accuracy (MAE ~2.20). Unstructured pruning preserved predictive accuracy (MAE ~2.01) without runtime gains, while structured pruning resulted in significant accuracy degradation (MAE ~6.3), revealing the sensitivity of instance-level colony detection to architectural compression. These findings provide practical guidance for selecting optimization strategies in resource-constrained laboratory deployments.
Read the full abstract

All authors Belaguppa Manjunath Ashwin Desai, Rohan Rajesh, Shreyas Murthy, Pronama Biswas, Revathi Vaithiyanathan

Cryptography and Security

Lightweight Zero Trust via Automotive SDN

Zonal in-vehicle networks ship Ethernet, MACsec, and TSN, but treat the network itself as trusted: once configured at the factory, there is no standardized runtime way to easily revoke access, rotate keys, or contain a compromised ECU. Zero Trust Architecture targets exactly that gap, yet existing automotive ZTA proposals bolt on dedicated infrastructure that duplicates the SDN management plane already required to enable SDVs. Thus, ZTA is not yet adopted in the automotive domain, and the question remains: can we do better? We answer this in two steps. Step 1 analyses what Open Alliance TC17~v1.0 MACsec/MKA with pre-shared CAKs already provides in terms of NIST SP~800-207 ZTA tenets. Step 2 adds CORECONF/YANG management as proposed in Open Alliance TC19, maps the SDN Controller and Agents one-to-one onto NIST's PE, PA, and PEP. We then instantiate this with two YANG-based mechanisms: a network-access-control flow and a key-management scheme. The result fully covers five and two partially of the seven tenets with no ZTA-specific infrastructure added.
Read the full abstract

General Relativity and Quantum Cosmology

Thermodynamics of Kerr-Newman-Bertotti-Robinson black holes

In this work, we extend the thermodynamic analyses of the neutral and specially charged Kerr-Bertotti-Robinson black holes to the general Kerr-Newman-Bertotti-Robinson family, in which the electric and external-field parameters are independent. Using covariant surface charges and canonical integrability methods, we determine the total angular momentum and the canonical mass. The angular momentum follows analytically from the combined gravitational and electromagnetic surface charges. However, the infinitesimal charge associated with coordinate time translations is not integrable in solution space, so the mass must be associated with a more general symmetry generator. Imposing the canonical integrability conditions on this generator, together with the Kerr-Newman mass as the zero-field boundary condition, selects the Christodoulou-Ruffini mass and determines the associated thermodynamic potentials. The first law and Smarr formula take the same form as the ones in usual Kerr-Newman case, and the two previously studied Kerr-Bertotti-Robinson cases are recovered as special limits.
Read the full abstract

Artificial Intelligence

UnitBoost: Managing Compound LLM Systems with a Merge Operator, Not a Model

Compound LLM systems often solve a coordination problem by adding a higher-level LLM. The resulting meta-agent reads workers' outputs, writes the final answer, allocates later calls, and decides when to stop. It is expressive, but it also concentrates three control decisions in an opaque, order-sensitive model call. We ask whether the manager needs to be generative at all. UnitBoost replaces that model with a defined meta-level operator: a task-given unit map turns worker outputs into slot-value proposals, a constrained argmax assembles the output, and the slots left unfilled or unsupported become an explicit residual for the next round. The operator is order-free, records unit provenance, and gives a simple guarantee: without coupling constraints, unit-wise maximization under the same admission score dominates selection of any complete candidate. On three held-out benchmarks, it exceeds the best single candidate chosen with gold labels by 0.060-0.195 absolute task-score points and input-matched generative managers by 0.048-0.076. Replacing only the management step improves six compound-system configurations by 0.013-0.182. Residual-directed rounds raise FanOutQA cell F1 from 0.4778 to 0.5524; matched controls show that the true residual outperforms random targets and ordinary rereading, while a label-free supply signal flags exhaustion after one unproductive round. The same analysis measures three conditions in which no such gain is available (one indivisible unit, unavailable unit identity, and an endpoint that charges for every emitted unit) and quantifies cross-unit coupling as a repair cost. The manager gives up semantic freedom and gains order invariance, unit provenance, and testable failure conditions.
Read the full abstract

All authors Xing Zhang, Guanghui Wang, Yanwei Cui, Mengdie Flora Wang, Peiyang He

Chemical Physics

Comprehensive Study of L-Menthol and Octanoic Acid as a Hydrophobic Eutectic Solvent

Hydrophobic eutectic solvents (HES) based on natural compounds represent promising green alternatives to conventional solvents. In this work, we investigate the physicochemical, structural, and dynamical properties of an ES formed by L-menthol and octanoic acid using a combined experimental and molecular dynamics simulation approach. Five compositions with molar ratios from 1:3 to 3:1 were studied with molecular dynamics simulation in the temperature range 15 degrees C to 35 degrees C. Experimental measurements of density and viscosity in the temperature range from 5 degrees C to 35 degrees C were complemented with results obtained from MD simulations employing the OPLS force field. Structural analyses based on radial distribution functions and Kirkwood-Buff integrals reveal that the dominant interactions in the mixture are hydrogen bonds between L-menthol and octanoic acid molecules. Dynamic properties, including self-diffusion coefficients and hydrogen-bond lifetimes, indicate that intermolecular hydrogen bonds between the two components are stronger and longer-lived than bonds between identical species. These findings provide molecular-level insight into the structure and transport properties of menthol-based ESs relevant for green solvent applications.
Read the full abstract

All authors Bernarda Lovrinčević, Martina Požar, Romana Cerc Korošec, Klara Švegelj, Peter Ogrin, Marija Bešter-Rogač, Tomaž Urbič

Differential Geometry

Outward minimizing $p$-capacity, horizons, and Schwarzschild rigidity

Let $(M^3,g)$ be a complete Riemannian manifold diffeomorphic to $\R^3\setminus\{0\}$, with nonnegative scalar curvature. Assume that a distinguished end is asymptotically flat, with ADM mass $m_+$. For each $p\in(1,3)$, define $c_{O,p}$ as the infimum of the Schwarzschild-normalized $p$-capacity over outward-minimizing finite-perimeter boundaries separating the two ends. We prove that $m_+\ge c_{O,p}$ whenever $A(g)>0$, where $A(g)$ is the infimum of the areas of boundaries separating the two ends. Equality at a single exponent produces a least-area horizon, forces its exterior to be the Schwarzschild exterior of mass $m_+$, and yields equality at every exponent. In the equality case, if the second end is also asymptotically flat, its mass satisfies $m_-\ge m_+$, with equality precisely for the two-sided spatial Schwarzschild manifold. We also show that a strict gap between $c_{O,p}$ and the unconstrained capacity infimum $c_{M,p}$ detects a horizon.
Read the full abstract

Quantum Algebra

Classification of finite-dimensional Nichols algebras of rank two in twisted Yetter–Drinfeld categories

Let \(G\) be a finite non-abelian group, let \(\Phi\in Z^3(G,\mathbb C^\times)\) be normalized, and let \(V,W\) be finite-dimensional simple objects of \({}_G^G\mathcal{YD}^{\Phi}\). We classify, up to interchange, the braided-indecomposable pairs \((V,W)\) whose supports generate \(G\) and for which \(\mathcal B(V\oplus W)\) is finite-dimensional. The classification consists of eight cases with five possible support quandles. In every case the Cartan graph is standard of type \(A_2\), \(B_2\), or \(G_2\), and the dimension is determined explicitly. A new phenomenon occurs {in the \(\Gamma _2\) case}: the twisted setting admits a family of type \(G_2\) absent from the ordinary \(\Gamma _2\) classification and we construct an explicit example over a non-abelian group of order \(16\).
Read the full abstract

Category Theory

A Natural Fuzzy Order on Fuzzy Numbers

This paper introduces a natural fuzzy order on fuzzy numbers that extends the natural orders on real numbers and interval numbers. We investigate its completeness properties and show that the space of uniformly bounded fuzzy numbers is conically complete and conically cocomplete, and that it is complete if and only if the underlying continuous t-norm is the Gödel t-norm. Moreover, it is proved that this space constitutes a \([0,1]\)-enriched domain if and only if the underlying continuous t-norm satisfies the (S) condition. These results provide a foundation for ordering fuzzy numbers.
Read the full abstract

Optics

Resonant subspace approximation for photonic crystal slabs

Resonant approximations are indispensable for the interpretation and efficient modeling of photonic crystal slabs, yet most of them describe an eigenmode as a pole in the complex energy plane at a fixed set of the remaining parameters. Tracking such poles and their hybridization across the Brillouin zone or upon variation of structural parameters is labor-intensive and severely limits the use of resonant approximations in band-structure calculations and structural optimization. Here we introduce a resonant subspace approximation that treats the photon energy and all other parameters on an equal footing. Considering a wide class of photonic crystal slabs that can be split into two non-resonant parts, we show that their resonances arise solely from the round-trip propagation of coupled Fourier harmonics between these parts, in direct analogy with Fabry–Pérot and waveguide modes. We generalize the scalar round-trip phase to round-trip and phase matrices, whose smooth dependence on all parameters allows us to project the problem onto a small resonant subspace defined at a single anchor point and to extrapolate it throughout a local region of parameter space of arbitrary dimensionality. As a result, rigorous computations at only a few points suffice to reconstruct the band structure, modal linewidths, complex hybridization, and optical spectra as functions of energy, wavevector, geometric dimensions, or permittivity within seconds. We demonstrate the accuracy and versatility of the approach on a strong hexagonal silicon grating, resolving intricate mode hybridization, symmetry-protected features, and subtle geometry-controlled effects that are hardly accessible to straightforward computations. The method is fast, accurate, and readily extensible, offering a practical route to the exploration, design, and optimization of resonant photonic crystal slabs.
Read the full abstract

All authors Nikolay A. Gippius, Ilia M. Fradkin, Natalia S. Salakhova, Sergey A. Dyakov

Machine Learning

Online Inverse Integer Linear Optimization via Small-Gradient Skipping: Constant Regret and Finite Mistakes

In online inverse linear optimization, the learner predicts a weight at each round, observes the optimal action of the agent, and updates its prediction. In the general setting, the gap of $\log T$ between the regret upper bound $O(d \log T)$ and the lower bound $\Omega(d)$ is unresolved (here $T$ is the total number of rounds and $d$ is the dimension). When the action set is M-convex, the regret is known to be bounded by $O(d \log d)$, but the method attaining it computes a center of gravity at every round. This paper therefore proposes Small-Gradient Skipping (SGS), a mechanism that skips the update at rounds without a mistake, and applies it to online gradient descent, the online Newton step, and MetaGrad. When the forward problem is an integer linear program with a unique optimal solution, the number of mistakes is bounded, for all three, by a quantity independent of $T$; and for the online Newton step and for MetaGrad with SGS, the dimension dependence of the regret becomes $O(d^2)$, that is, the factor $\log T$ is removed. Moreover, when the action set is M-convex, the regret is bounded efficiently without computing a center of gravity.
Read the full abstract

Robotics

GTA-2: A Multi-VLM Framework for Synthesizing Robot Manipulation Skills via Grounded Task Axes

Robotic manipulation tasks are often decomposed into behaviors or skills. However, one often needs to predefine these behaviors for specific tasks or try to cover a wide range of tasks using generic skills. As a result, these behaviors can remain too coarse to expose the geometric, control, and scene-dependent decisions required for execution. We introduce Grounded Task Axes v2 (GTA-2), a modular multi-VLM framework that constructs executable, task-bespoke manipulation skills from reusable object-centric task-axis components. Rather than predicting actions end-to-end or composing fixed task-level primitives, GTA-2 represents each skill as semantic subtasks comprising task-relevant keypoints and axes, controller compositions, and scene-dependent parameters. Four specialized VLM agents separately decompose the task, construct an abstract task-axis skill, assign controller parameters, and ground the required visual features from RGB-D observations. This abstraction-to-grounding factorization enables zero-shot skill generation without task-specific robot demonstrations, policy training, or fine-tuning. It also keeps intermediate decisions explicit, allowing targeted human feedback to refine an incorrect stage while preserving correct components. We evaluate GTA-2 on 14 real-robot manipulation tasks against a VLA policy pi_{0.5} and two Code-as-Policies baselines using task-axis controllers or conventional robot primitives. GTA-2 achieves an average zero-shot success rate of 73.9%, exceeding the strongest baseline by 31.4 percentage points, while targeted refinement raises GTA-2's average success rate to 90.7%. Project page: https://gta2-project.github.io/
Read the full abstract

All authors M. Yunus Seker, Shobhit Aggarwal, Ruwan Wickramarachchi, Jonathan Francis, Oliver Kroemer

Cryptography and Security

Quantifying IIoT Sensor Node Criticality by Fusing its Data Criticality and Security Vulnerability

The integration of the Industrial Internet of Things (IIoT) into manufacturing has transformed industrial operations by optimising production management and ensuring product quality through smart industrial sensors that regulate processes based on real-time data. However, these sensor nodes are highly vulnerable to cyber threats, posing significant security risks that compromise their reliability and integrity. While existing research explores cybersecurity vulnerabilities and cyberattack-based methods for ranking critical nodes, some studies assess node criticality based on the impact of sensor data on product quality. However, a comprehensive approach that integrates both data criticality and cybersecurity vulnerability remains unexplored. To bridge this gap, this study introduces a novel framework that evaluates IIoT sensor node criticality by leveraging Dempster–Shafer (D-S) theory to fuse data criticality and cybersecurity vulnerabilities. The proposed method is validated using a dataset from red wine production, demonstrating its effectiveness in ranking sensor nodes based on both factors. The results show that criticality rankings based on security vulnerability scores computed using CVSS version 4.0 differ significantly from those obtained with CVSS version 3.1, highlighting the influence of enhanced vulnerability assessment methodologies. While initially applied to wine manufacturing, this framework is adaptable to broader industrial applications with minimal modifications, offering a robust approach to securing IIoT-enabled production systems.
Read the full abstract

Quantum Physics

Dynamic Observable Subspaces: Fast Simulation of Observable Quantum Dynamics

By decomposing quantum dynamics across the Lie orbits of a list of observables, we find polynomial bounded classical simulations for the dynamics of the quantum system given a polynomial sized dynamic Lie algebra (DLA) for the generators of the quantum system. To do so, we describe how to construct the Dynamic Observable Subspace (DOS) that captures all the relevant dynamics for calculating expectation values for a specific observable for Pauli strings, diffusor mixers, and general local generators. Efficient sparse matrix representation, decoupling nonlinearity from such representation and basis construction with permissible pruning allows us to simulate the closed dynamics of such systems with hundreds of qubits. Moreover, we find that while restricted DOS circuits may not express the space of an associated Hamiltonian, they can dramatically outperform a fully expressive circuit due to the absence of the Barren Plateau. While classically simulatable, such circuits can still exhibit a form of quantum advantage through inference, act as warm starting for more expressive circuits, and find high quality trial states for Quantum Amplitude Amplification or Quantum Phase Estimation. In the simulation of quantum dynamics, we find that our restricted circuit can be simulated in polynomial time while producing high quality guiding state for downstream tasks like Quantum Phase Estimation with significantly lower energy than circuits with full expressivity.
Read the full abstract

Quantum Physics

Electrically tunable, two-photon interference from remote silicon-vacancy centers in industrial silicon carbide

Distributed quantum networks rely on spatially separated, independently operated quantum systems as network nodes, whose emitted photons must be interfered with high visibility to establish end-to-end entanglement. Crucially, for network-relevant applications, high visibilities must be achieved over prolonged timescales to reduce overheads for error correction, and to increase network rates. Here, we demonstrate experimentally that silicon vacancy $\mathrm{V_{Si}}$ color centers in silicon carbide (SiC) achieve these requirements, notably in a mass-deployable fashion. We integrate $\mathrm{V_{Si}}$ centers in different industrial-grade SiC p-i-n diodes, which are controlled via voltage biassing. This way, we demonstrate both, spectral overlapping of 19 randomly selected $\mathrm{V_{Si}}$ centers in different diodes, as well as spectral narrowing close to the lifetime limit, i.e., typically below 60 MHz. Notably, these performance parameters are long-term stable, e.g., readjusting the p-i-n diode bias is required only every 8.4 hours, which reduces significantly the overall experimental overhead. We then use these assets to demonstrate high-quality two-photon interference between $\mathrm{V_{Si}}$ centers located in two different cryostat setups, which are spatially separated by two meters. Notably, we perform a 26-days long measurement campaign, demonstrating two-photon interference with state-of-the-art raw interference visibilities of 82%, which aligns with the current state-of-the-art. These results establish $\mathrm{V_{Si}}$ centers in industry-grade SiC devices as a scalable, spectrally stable building block for distributed quantum networks.
Read the full abstract

All authors Fedor Dzmitryevich Hrunski, Daniel Scheller, Maximilian Hollendonner, Kim Ullerich, Shravan Kumar Parthasarathy, Chiun Fu, Andre Pointner, Wolfgang Knolle, Florian Kaiser, Durga Bhaktavatsala Rao Dasari, Roland Nagy

Quantum Physics

Small-Bias Quantum Approximate Counting via the Multiplicative Adversary Method

We study the two-weight decision version of quantum approximate counting: given oracle access to $x\in\{0,1\}^N$, distinguish $|x|=M$ from $|x|=M+\Delta$ with success probability $1/2+\zeta$. Using the multiplicative adversary method, we prove $\Omega\left(\max\left\{\zeta\sqrt{(N-M)(M+\Delta)}/\Delta,\sqrt{\zeta N/\Delta}\right\}\right)$. The same parameter dependence follows from the polynomial-method characterization of the two-layer symmetric function by Podder, Yao, and Ye. Our contribution is a multiplicative-adversary derivation that tracks the progress produced by individual oracle queries. For the first term, after complementing the input if necessary, we assume $M+\Delta\le N-M$. We use the Hamming-layer subspaces from the eigenspace method of Ambainis, Spalek, and de Wolf and compose their adjacent-layer unitary maps to relate the two nonadjacent promise layers. After fixing the queried coordinate, the analysis block-diagonalizes into four-dimensional subspaces. An exact calculation of the one-query progress ratio gives the first lower bound. The same estimate also implies $\left\|(I-\widehat{\Pi}_{\mathrm{bad}})\lvert\Psi^T\rangle\right\|^2=O\left(T^2\Delta^2/((N-M)(M+\Delta))\right)$ for the coherent input superposition used in the adversary argument. For the second term, we prove directly using a three-eigenvalue multiplicative adversary that unique OR on $n$ bits with success probability $1/2+\zeta$ requires $\Omega(\sqrt{\zeta n})$ queries, and then reduce unique OR to the two-weight counting problem.
Read the full abstract

Earth and Planetary Astrophysics

Hybrid Spin-Orbit Tomography for Earth-like Planets: Simultaneous Mapping of Static Surfaces and Dynamic Clouds from Multicolor Light Curves

Photometric variability of directly imaged exoplanets encodes information on both persistent surface features and time-variable phenomena such as clouds. We present Hybrid Spin-Orbit Tomography, a method that simultaneously retrieves multiple static components and a single dynamic component, together with their reflection spectra, from multiband photometric variability. The method integrates spectral unmixing and dynamic spin-orbit tomography by introducing a time-dependent component into the forward model. We employ sparse regularization for static surface distributions, volume regularization for reflection spectra, and a Kronecker-sum kernel regularization for the dynamic component. Applying the method to a toy Earth model, we successfully recover static surface distributions together with a single time-dependent cloud component. Furthermore, we apply the method to DSCOVR/EPIC observations of Earth and retrieve a dynamic component consistent with real cloud distributions, together with components broadly interpretable as oceans, vegetation, and land surfaces. Relative to static spin-orbit unmixing, the hybrid model substantially reduces the residual scatter of the multiband light curves, demonstrating that explicit treatment of dynamic components improves global mapping from color variability. These results provide a step toward the simultaneous retrieval of static and dynamic planetary components, while highlighting the need for more physical cloud-surface masking models and uncertainty quantification.
Read the full abstract

High Energy Physics (Phenomenology)

Charmed baryon semileptonic decays in a relativistic three-quark model

We study the spin-$1/2\to1/2$ semileptonic decays of singly charmed baryons ($\Lambda_c^+$, $\Xi_c^{0,+}$, and $\Omega_c^0$) into the light baryon octet within a relativistic three-quark model. The constituent quark masses and spatial wave functions are determined by the baryon mass spectrum. For the physical $\Xi_c$ states, the light flavor $\mathrm{SU}(3)$ breaking ($m_s > m_{u,d}$) naturally induces a coherent mixing between the flavor antitriplet and flavor sextet configurations, which is completely fixed by the mass eigenstates. Consequently, no adjustable parameters are introduced in calculating the weak transition amplitudes. Using these wave functions, we calculate the $c\to s,d$ helicity amplitudes with the Bakamjian–Thomas boost, including the spatial Jacobian and the Wigner rotations of the constituent spins. The branching fractions, $q^2$ distributions, longitudinal polarizations, and form factors are then obtained from these amplitudes. For the $\Lambda_c^+\to\Lambda\ell^+\nu_\ell$ and $\Lambda_c^+\to n\ell^+\nu_\ell$ modes, our branching fractions and form factors are in good agreement with the experimental data and Lattice QCD results. For $\Xi_c^0\to\Xi^-e^+\nu_e$, we obtain branch ratio $\mathcal B_{\ell} \simeq 4.0~\%$, which is consistent with recent Lattice QCD calculations but lies well above the current experimental average. Clarifying the origin of this discrepancy calls for further dedicated efforts from both experimental and theoretical sides. For the $\Omega_c^0\to\Xi^-\ell^+\nu_\ell$ decay, the spin-$1$ $ss$ spectator yields a positive longitudinal polarization of the final $\Xi^-$, in contrast to the negative polarizations in the predominantly antitriplet decay modes. Our predictions for the $q^2$ spectra, angular asymmetries, and final baryon polarizations may provide useful references for future measurements of singly charmed baryon semileptonic decays.
Read the full abstract

All authors Ru-Hui Ni, Zhen-Yang Wang, Jia-Jun Wu, Bing-Song Zou

Image and Video Processing

Morphological Decoupling-Based Skeletal Classification for Clinical Assessment of Malocclusion

Malocclusion skeletal grading is a fundamental task in orthodontics, critical for diagnosis and treatment planning. Traditionally, cone-beam computed tomography (CBCT) is used for visual measurement, and the reconstructed lateral cephalograms are handed over to expert dentists for diagnosis. However, manual review is time-consuming, labor-intensive, and subject to inter-operator variability. Therefore, an automatic CBCT-based system is needed for reliable malocclusion skeletal grading. In this case, we develop TeethGNN, a novel graph-based framework designed to combine CBCT image features with morphological information for accurate and efficient malocclusion grading. TeethGNN utilizes a decoupled learnable decoder to directly predict key morphological indicators from CBCT images, eliminating the need for manual measurements. These morphological features are then fused with image features using a graph neural network (GNN), which effectively models the relationships between the modalities. To further enhance robustness and calibration, we introduce a collaborative calibration strategy. This strategy combines multi-scale graph adversarial perturbation for explicit calibration and nonlinear topological graph calibration for implicit confidence adjustment. Extensive experiments and ablation studies on our collected clinical dataset demonstrate that our malocclusion measurement system achieves 77.08% in accuracy and 89.61% in AUC, outperforming the compared state-of-the-art methods. These results validate the effectiveness of graph-based multimodal fusion and collaborative calibration in improving malocclusion grading performance. Our system shows strong potential for advancing computer-aided orthodontic diagnosis, providing an accurate and reliable solution for vision-based clinical measurement and diagnosis.
Read the full abstract

All authors Zhichun Jin, Zhicheng He, Hao Xu, Dongyang Li, Lin Wang, Hongliang Ren, Long Bai

Hardware Architecture

HBFSim: Fast and Faithful Simulation of High-Bandwidth Flash Under Real GPU Execution

High-Bandwidth Flash (HBF) places high-capacity NAND beside HBM to relieve the memory-capacity bottleneck of LLM inference, yet its system-level behavior cannot be evaluated before hardware becomes available. Cycle-level GPU simulators are too slow for production-scale models. Trace replay has a further shortcoming: it cannot capture the allocation, migration, and execution changes induced by different HBM-HBF configurations. Our key insight is that HBF need not be evaluated by simulating the GPU: only the program-visible effects of HBF need to be modeled. And only a real LLM workload running on real hardware can answer the arguments about HBF. Hence the modeled service has to be injected into that running program, and the injection must not destroy the GPU concurrency that would hide the original I/O latency. We present HBFSim, an open-source HBF simulator that executes LLM workloads on a real GPU while modeling HBF timing, thermal, and other behaviors online. HBFSim rewrites the PTX of the workload's kernels and routes accesses inside a registered address range into the HBF simulator. It supports asynchronous TMA transfers and capacities beyond physical GPU memory. HBFSim leaves the model's run unaffected across ordinary-memory, TMA, and capacity-mode tests. The delay it injects matches the delay requested to within 0.152%. We also design a coupled thermal module that puts HBF, HBM, and the GPU in one advanced package, which is important for answering how severe the hot throttling problem becomes after HBF runs for a long time. Experiments with Qwen3-30B show how package heating, HBM-HBF allocation, and shared MoE demand jointly constrain the design space of future HBF accelerators. The source code of HBFSim is available at https://github.com/SlugLab/hbfsim/.
Read the full abstract

All authors Yanpeng Hu, Yiwei Yang, Yuanwu Zhu, Yusheng Zheng, Andi Quinn, Wei Zhang

Materials Science

uFlowCSP: Crystal Structure Prediction using Mean flow generative models

Crystal structure prediction (CSP) is fundamental to computational materials discovery. Generative models including CDVAE, DiffCSP, FlowMM, and CrystalFlow learn stable-crystal distributions directly, but diffusion and flow-matching inference requires tens to thousands of sequential network evaluations per candidate. We introduce uFlowCSP, a MeanFlow-based CSP model that learns the average, rather than instantaneous, probability-flow velocity. It generates a complete structure in one to five evaluations, delivering 5x-58x faster inference with equal or better performance. A chemistry- and symmetry-aware Transformer uses canonical atom ordering, global composition, and per-token chemistry embeddings. A coarse crystal-system token is used only during training; it provides additive gains, particularly improving space-group agreement despite being absent at inference, which remains formula-only. On MP-20 with 20 candidates per target, one step matches CrystalFlow (78.38% vs. 78.34%) with 100x fewer evaluations and about 10x lower wall-clock time. Five steps reach 83.64%, exceeding CrystalFlow (78.34% at 2,000 evaluations) and DiffCSP (77.93% at about 20,000), while using 20x fewer evaluations. uFlowCSP generates 10,000 structures in 0.39-1.31 minutes, versus 6.5 for CrystalFlow and 76.1 for DiffCSP. Under CSPBench's energy-ranked top-five structure-and-space-group criterion, five-step uFlowCSP reaches 72%/72%/65% structure, space-group, and consensus match rates. CrystalFlow reaches 78%/73%/68% at 100 steps but falls to 49%/32%/31% at five. Thus, uFlowCSP improves accuracy per network evaluation, not merely peak accuracy.
Read the full abstract

All authors Sourin Dey, Dipannoy Das Gupta, Lai Wei, Sadman Sadeed Omee, Jianjun Hu

Cryptography and Security

CS-Guard: Benchmarking LLM Guardrails for Code Generation Security

Large language models (LLMs) have been ex- ploited to generate malware, but the effective- ness of guardrails for code generation secu- rity remains unclear. We introduce CS-Guard, the first benchmark to systematically evalu- ate guardrails for code generation security. It covers 1) text-to-code generation with 1000 high-quality malware-generation prompts, 7 jailbreak attacks, and a novel fictional scenario attack (FSA) that embeds malicious intent in a legitimate fictional software-development sce- nario; and 2) code-to-code generation with 331 code prompts spanning code infilling, code completion, and code translation. We empiri- cally evaluate 9 guardrails across seven LLMs. We find that current guardrails perform poorly against malicious code-generation re- quests: for text-to-code, the average attack success rate (ASR) after jailbreaks reaches about 50% for many guardrails; for code-to- code, average ASR approaches 100% on base LLMs and remains high across many guardrails (14.4% to nearly 100%). Our FSA also achieves ASR close to 100% across many guardrails, raising major reliability concerns for real-world software development. To sup- port future research, CS-Guard uses a modular three-layer guardrail taxonomy that lets devel- opers register guardrails for evaluation. We release the benchmark and data to enable fur- ther community evaluation.
Read the full abstract

Information Theory

Learning-Aided Short Code Design for ISAC based on MIMO-OFDM

This paper proposes a deep learning (DL)-based coded waveform design for integrated sensing and communications (ISAC), enabling flexible trade-offs between communication reliability and ranging accuracy in short-block transmissions. The proposed scheme is built upon a practical multiple-input multiple-output orthogonal frequency-division multiplexing (MIMO-OFDM) architecture, where the communication channel state information and the angles of the static targets are assumed available at the transmitter. A transformer-based transmitter encodes input information bits directly into ISAC transmit waveforms to jointly optimize the bit error rate (BER) performance and the delay modified Cramer-Rao bound (MCRB). A corresponding transformer-based receiver is adopted at the communication side to recover the transmitted information bits. We further examine the learned codewords for communication-oriented and sensing-oriented designs, revealing that a balanced ISAC waveform naturally exhibits an intermediate structure between these two extremes. Numerical results illustrate these codeword structures and demonstrate that the proposed design provides substantial trade-off gains over conventional schemes based on standard channel coding and modulation.
Read the full abstract

All authors Mingcheng Nie, Shuangyang Li, Geng Wang, Peng Cheng, Shenghong Li, Chang Liu, Giuseppe Caire, Yonghui Li

Machine Learning

A practical DIRECT-type algorithm for medium-scale black-box global optimization

The DIRECT algorithm is a deterministic global optimization method known for its versatility and balanced exploration-exploitation strategy. However, DIRECT-type algorithms are primarily effective for low-dimensional problems and often exhibit slow convergence as dimensionality increases, limiting their applicability to more complex optimization tasks. To address this limitation, this paper introduces X-DTC-GL, a novel DIRECT-type algorithm that incorporates dynamic partitioning and hybridization techniques. The dynamic partitioning approach adaptively refines the search space based on local one-dimensional surrogate models, enabling rapid subdivision of promising hyper-rectangles. The hybridization strategy selectively employs a hill-climbing method to exploit promising regions identified by the surrogate models. Extensive experiments on four diverse benchmark suites demonstrate that X-DTC-GL significantly outperforms existing DIRECT-type baselines, achieving improvements of ~12% in solvability and ~27% in solution quality. Performance-profile analyses indicate the fastest convergence on up to ~40% of instances, the best runtime performance on ~17% of problems, and competitive overall execution times. By improving performance within the partition-based framework, these advances strengthen the algorithm's competitiveness in state-of-the-art black-box optimization.
Read the full abstract

Differential Geometry

On the Asymptotics of the Volume of Hitchin Moduli Spaces

Let $X$ be a compact Riemann surface of genus $g\geq2$, and let $\M$ be the moduli space of rank-two trace-free Higgs bundles with fixed determinant of odd degree. For the normalization of the Hitchin metric used in this paper, we prove that the volume of geodesic ball in the Hitchin moduli space is given by $$ \Vol_{g_{L^2}}B_{L^2}(p,R) =\frac{2^{4g-3}\pi^{9g-9}}{(3g-3)!}\,R^{6g-6}+o(R^{6g-6}) $$ for every $p\in\M$. We also determine the leading asymptotics of Hamiltonian sublevel volumes and exponentially weighted volumes. The proof combines homogeneity of the hyperkähler volume form with metric asymptotics on the regular Hitchin locus. Symplectic reduction and the Prym polarization evaluate the coefficient, which is independently recovered by equivariant localization.
Read the full abstract