Peer Revenue Publish or profit

Papers

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

All of arXiv

Monday 7 September

Materials Science

Excitons probe intrinsic flat band Mottness in a van der Waals heterostructure

Excitons provide a sensitive optical probe of electronic correlations in nearby two-dimensional materials, yet their coupling to intrinsic flat-band Mott systems remains largely unexplored. Here we combine gate-tunable optical spectroscopy with first-principles calculations to study monolayer WSe$_2$ in direct contact with the van der Waals Mott insulator Nb$_3$Cl$_8$. The gate evolution of WSe$_2$ excitonic resonances reveals signatures of a correlation-reconstructed Mott gap in Nb$_3$Cl$_8$ that is absent from the single-particle band picture. In the electron-doped regime, the WSe$_2$ 2s Rydberg exciton undergoes a multistage evolution and develops into interlayer attractive and repulsive polaron branches, showing that a Rydberg exciton can be dressed by strongly correlated flat-band electrons in an adjacent Mott layer. Under an out-of-plane magnetic field, spin-polarized Nb$_3$Cl$_8$ states further induce valley-selective exciton coupling, producing a strongly enhanced circular polarization of the WSe$_2$ exciton emission. These results extend exciton-based sensing and exciton-polaron physics to intrinsic flat-band Mott materials, providing an optical route to probe and engineer correlation-driven interfacial quasiparticles.
Read the full abstract

All authors Xinyue Huang, Xintong Tan, Haowei Chen, Yingzhou Huang, Yushen Zhou, Yuchen Gao, Zhijie Ma, Chengxin Xiao, Kenji Watanabe, Takashi Taniguchi, Jianpeng Liu, Zuxin Chen, Youguo Shi, Wang Yao, Yu Ye

Machine Learning

Temporal Heterogeneous Graph Transformer for Credit Card Fraud Detection

Credit card fraud detection typically relies on tabular features, while repeated attributes can also provide useful relational signals. This paper proposes THGT-FD, a Temporal Heterogeneous Graph Transformer for Fraud Detection. Each transaction is represented using one transaction token and six types of relation tokens and incorporates Time2Vec encoding into the transaction representation. A Transformer learns the interactions among these tokens within each individual transaction and then outputs a fraud probability. Experiments were conducted on 150,000 transactions sampled from the IEEE-CIS Fraud Detection dataset and chronologically partitioned according to TransactionDT. On the test set, THGT-FD achieved an AUC-ROC of 0.8536, an average precision of 0.4164, and a Recall@5% of 0.4708. The class-weighted histogram-based gradient-boosting baseline achieved an AUC-ROC of 0.8722. The results indicate that relation tokens provide useful information for fraud-risk ranking, although the current model does not yet incorporate entity-level historical aggregation.
Read the full abstract

Computer Vision

Ambient @ EgoProactive 2026 : Proactive Egocentric Assistance with Visually Grounded Supervision

We present our submission to the EgoProactive track of the ECCV 2026 Wearable AI Challenge, which ranked first in the large-model division and second in the <=2B division. The task requires a wearable assistant to decide after each eight-second segment of egocentric video whether to intervene or remain silent. Our approach has two main components. First, we reformulate intervention timing as single-token classification. Rather than generating either $interrupt$<utterance> or $silent$, the model predicts yes or no, and we derive the decision from the renormalised probabilities of these two tokens. This formulation improved macro-F1 by 0.249 and G-mean by 0.30 over free-form generation. Second, because labelled data were limited to the released validation set, we generated additional supervision using a tool-calling video agent that inspects each clip and assigns intervention timestamps. A narration-only alternative was four times larger and ten times cheaper, but transferred worse than supervision from an unrelated real corpus, suggesting that visual grounding is more important than annotation volume for this task.
Read the full abstract

Combinatorics

Bounded Ratios of Lorentzian Polynomials II: The Complete Quadratic Local-to-Global Classification

Every quadratic Hessian slice of a Lorentzian polynomial yields bounded monomial ratios among the normalized coefficients of the polynomial. We determine exactly for which pairs $(n,d)$ these quadratic-slice ratios generate the full bounded-ratio cone for every $M$-convex support $S\subseteq\Delta_n^d$. For $d\geq 2$, this quadratic local-to-global principle holds universally if and only if \[ n\leq 3,\qquad d=2,\qquad\text{or}\qquad (n,d)=(4,3). \] In every remaining case, the principle fails already for Lorentzian polynomials with full support: for cubics in $n\geq 5$ variables and for polynomials of degree $d\geq 4$ in $n\geq 4$ variables. We identify the two minimal obstructions, at $(n,d)=(4,4)$ and $(n,d)=(5,3)$, and propagate them throughout the failure region by degree and variable aggregation.
Read the full abstract

All authors Dijia Chen, Bowen Gan, Ivy Liu, Zemeng Wang, Chengzhi Wu

Logic in Computer Science

Optimising Metamath Proofs for Human Working Memory

Mathematical proofs vary in legibility. While most proof optimisation techniques seek to minimise proof size, the strategic reordering of inferences can reduce the working memory demand of proof checking without altering overall size. Metamath serves as a prime case study for this approach: its verification architecture requires proof steps to be ordered in a manner that prioritises algorithmic efficiency over readability. In this paper, we introduce algorithms to minimise both peak and cumulative memory consumption, applying the latter as a novel proxy for sustained human cognitive effort. We achieve this by representing proofs as directed acyclic graphs and modelling their execution as a pebbling game. Finding an optimal ordering via brute force is computationally infeasible, so we use heuristics to provide approximations. We apply these algorithms across Metamath's ZFC set theory library and present case studies demonstrating how automated reordering systematically improves the presentation of formal mathematics.
Read the full abstract

Robotics

RoboDreamer: Anticipatory Humanoid Locomotion with Predictive State-Space Models

Humanoid locomotion requires control policies that remain stable under imperfect sensing while exploiting temporal context for consistent motion. We present RoboDreamer, a two-stage teacher–student framework that combines next-observation consistency with randomized continuous temporal masking. A teacher is first trained on clean observations, and a student is then distilled under masked recent observations, encouraging the policy to infer missing current information from history. At inference, the same masking interface is reused for implicit closed-loop action refinement and optional multi-step action chunking. Mamba is used as the temporal backbone, while matched ablations show that masking/distillation provides a substantial part of the gain and Mamba contributes additional tracking improvements with real-time latency. Experiments in IsaacLab, MuJoCo, and on a Unitree G1 demonstrate robust motion tracking under observation masking and successful real-world deployment.
Read the full abstract

All authors Zhe Li, Yangyang Wei, Xichen Yuan, Zhenzhe Zhang, Weihao Yuan, Shanghang Zhang, Jianfei Yang

Artificial Intelligence

Risk Is Not Review Value: Wrong-Answer Exposure Under Bounded Review Budgets

LLM assistants often produce more answers than humans can review before users see them. Most evaluations ask whether an answer is wrong, unsupported, or low-confidence. Bounded review budgets instead ask which answers should be checked first under a fixed review budget. Risk alone is not enough: a high-risk answer may be hard to repair, while a moderately risky answer may be directly correctable from available evidence. For generated-answer evaluation, we model review prioritization as exposure reduction, where review value combines estimated wrongness, intervention affordance, impact, and cost. We evaluate review queues with Wrong-Answer Exposure Ratio (WAER), the fraction of wrong answers left unreviewed, and post-repair residual exposure (PRRE), the fraction still exposed after deterministic benchmark-supported repairs. PRRE uses repairability rules that do not numerically reuse the affordance scores used for ranking. On a 720-item TAT-QA/SciFact stress benchmark, review-value ranking keeps answer-level WAER nearly unchanged at 20% budget (0.605 vs. 0.600) but lowers PRRE from 0.881 to 0.716. These results show that trustworthy LLM evaluation should measure not only error detection, but also how limited review capacity reduces exposed wrong answers.
Read the full abstract

All authors SangJin Park, Myungsub Choi, Jineok Kim, Minseung Kang

Computer Vision

CASCADE: A Spatio-Temporal-Causal Reasoning Representation and Dataset for Driving

Reasoning is a promising route to the generalization that autonomous driving requires in the long tail, as it can infer how the elements of a scene depend on one another and traverse those dependencies to conclusions beyond what is observed. Yet it is hard to tell whether a model's conclusions follow the scene's dependencies, because no driving representation makes them explicit enough to test against. Text-based reasoning traces lack spatio-temporal grounding, spatio-temporal scene graphs lack causal links, and reasoning annotations at scale are increasingly model-generated and hard to verify. To this end, we introduce CASCADE (Causal Spatio-Temporal Analysis of Driving Environments), which encompasses two components: (1) a structured scene representation for reasoning in driving scenes and (2) a human-annotated dataset built on it. For every actor that interacts with the ego vehicle, the CASCADE representation records frame-by-frame, for as long as the actor is visible, what action is taken, where it occurs, and how it depends on the actions and states of others. The resulting structure makes reasoning predictions machine-verifiable: they can be scored against it element by element, without relying on (M)LLM judges. The CASCADE dataset provides comprehensive human annotations for 2,066 driving clips of the PhysicalAI dataset, with over 34K elements that establish the spatio-temporal and causal context of each scene, including 8.6K time-stamped ego and agent actions, 3.7K causal links and 2.9K potential influences, and 6.1K annotations for agents, objects, traffic lights, and environments. Being entirely human-annotated, CASCADE provides the reference for this comparison: benchmarking the reasoning abilities of Physical AI models, and verifying the quality of automatically generated reasoning labels. The CASCADE dataset is available at https://huggingface.co/datasets/nvidia/cascade.
Read the full abstract

All authors Jenny Schmalfuss, Despoina Paschalidou, Simon Gerstenecker, German Ros, Jose M. Alvarez

Computation and Language

Where to Look and What to Use: Retrieve-Localize-Generate for Long-Term Conversational Memory Question Answering

Retrieval-augmented generation (RAG) enables large language models (LLMs) to answer questions by accessing external knowledge and has been widely adopted for long-term conversational memory question answering. However, existing methods suffer from two key challenges: (1) fragmented evidence scattered across temporally distant sessions, and (2) noisy content within retrieved sessions that triggers the lost-in-the-middle effect. To address these challenges, we propose MemLoc, a unified Retrieve-Localize-Generate framework for long-term conversational memory QA. For retrieval, MemLoc decomposes each session into multi-granularity memory units and performs query routing via an inner-memory graph with entropy-based granularity selection. It further models cross-session semantic and temporal dependencies through a cross-memory graph, enabling coarse-to-fine retrieval of top-K relevant memory candidates. For localization, we introduce a reasoning-based evidence locator trained with Self-reflective Hint Policy Optimization (SHPO), which performs progressive refinement by extracting query-relevant fragments within memory units to suppress noise and reranking across candidates to remove redundancy, producing a compact evidence set with lightweight location IDs. For generation, these IDs act as precise grounding signals that guide the LLM to the correct memory positions, mitigating the lost-in-the-middle effect while preserving original contextual integrity. Extensive experiments on four benchmarks demonstrate that MemLoc achieves state-of-the-art retrieval accuracy and response quality while maintaining efficiency. Our code is available at: https://github.com/Nikol-coder/MemLoc.
Read the full abstract

All authors Yifan Wang, Xinkui Lin, Yongxiu Xu, Shen Gao, Ruochen Yang, Kun Huang, Yubin Wang, Jie Wu, Wei Liu, Jian Luan, Hongbo Xu, Shuo Shang

Number Theory

The method of telescoping continued fractions

We give an approach to discover continued fractions for series of the form $$\sum_{k=0}^\infty \frac{\epsilon^k}{(x+k)^s},$$ where $\epsilon = \pm 1$. We find a continued fraction of the form \begin{equation*} \frac{a_1}{b_1(x)} \fplus \frac{a_2}{b_2(x)} \fplus \fdots \end{equation*} where $a_k$ are constants and $b_k(x)$ are polynomials. Our technique involves telescoping continued fractions. This provides a discovery approach to continued fractions given by Ramanujan for $$2\sum_{k=1}^\infty \frac{(-1)^{k+1}}{x+2k-1}, 2\sum_{k=0}^\infty \frac{1}{(x+2k+1)^2}, 2\sum_{k=0}^\infty \frac{(-1)^k}{(x+2k+1)^2}, \sum_{k=1}^\infty \frac{1}{(x+k)^3}, $$ and the like. We display the first few terms of several continued fractions obtained in this manner for larger values of $s$, including $s=5, 7, 9, 11$. They do not follow as simple a pattern as Ramanujan's continued fractions.
Read the full abstract

Robotics

Human-Aware Target Tracking and Navigation: Fusing Kinematic State Estimation with Structural Map Constraints

Autonomous mobile robots performing person-following tasks often suffer from temporary occlusions and sensor track loss in dynamic environments. This research presents an end-to-end autonomous navigation stack that addresses target occlusion through map-informed spatial reasoning. The proposed system features a multi-modal perception pipeline, fusing deep learning-based visual tracking with 2-dimensional LiDAR point clustering to maintain high-fidelity tracking of a tagged person. A continuous state estimator integrates this perception data with wheel odometry and IMU sensors for stable localization. When the active track is lost due to occlusion, the system activates a map-based recovery framework. Leveraging a predefined topological map, the system executes a graph-based search to propagate the target's last known trajectory along structurally defined walking lanes, adhering to left-hand regional conventions. By generating a discrete set of feasible future trajectories, the robot reasons about potential structural trajectory changes, such as continuing a heading or turning at an intersection. This map-informed prediction is fed directly to the local obstacle avoidance planner, enabling the robot to continue following its target safely and predictably until the person is visually reacquired. Real-world evaluations in dense multi-person environments demonstrate the system's robustness, achieving a 71.4% target reacquisition success rate during major occlusion events lasting up to 7 seconds.
Read the full abstract

All authors Sagar Gupta, Don Gideon, Seng W. Loke, Kevin Lee, Bijo Sebastian

Quantum Physics

Complexity Amplification from Compression in Quantum Random Access Optimization

Compressed quantum encodings aim to overcome hardware limitations towards tackling challenging problems at scale, with many classical variables mapped onto noncommuting observables of fewer qubits. Classically, relaxations such as the semidefinite program formulation of MaxCut trade solution quality for computational efficiency. By contrast, quantum relaxations based on compression can amplify the worst-case complexity of the problem being solved. We study quantum random access optimization (QRAO), a special case of the Pauli correlation encoding (PCE) framework that assigns up to three binary variables to the Pauli $X$, $Y$, and $Z$ observables of each qubit, with the packing choices determining the compressed Hamiltonian to be optimized. We identify explicit QRAO optimal energy promise problems complete for NP, StoqMA, and QMA, with inverse-polynomial promise gaps for the latter two. Our problem reductions preserve inverse-polynomial promise gaps without requiring gadgets or ancillas. For any prescribed packing, we show that weighted MaxCut instances compress, up to a known shift and rescaling, to arbitrary nonnegative-weight pairwise Pauli couplings allowed by the packing. For QRAO, using one aligned axis gives an NP-complete energy problem. Using two or three positive aligned Pauli axes generally gives QMA-complete problems, with bipartite restrictions in BQP $\cap$ StoqMA. We show that this computational hardness survives compilation and is practically relevant. Notably, this result applies directly to the current QRAO compiler implementation in Qiskit Optimization 0.7.0, confirming our hardness results are not artifacts of artificial or contrived packing rules. Altogether our results identify worst-case complexity barriers arising from quantum compression, while making no broad claims about typical cases or the performance and trainability of algorithm pipelines that use it.
Read the full abstract

Quantum Physics

QCxSimulation: Scatter-Aware X-Ray Projection Radiography via Discrete-Time Quantum Walks

X-ray projection radiography is a non-invasive imaging technique used in medical diagnostics and industrial inspection. The simulation of X-ray projections is commonly used to optimise acquisition protocols and improve image quality before performing costly scans. Classical photon transport simulations that include realistic X-ray scattering physics are computationally expensive because they require the sampling of a large number of distinct scattering paths. This limits the practical exploration of parameter spaces such as beam energy. Quantum computing offers the potential to solve high-dimensional problems faster by making use of quantum properties such as superposition. This work introduces a discrete-time quantum walk algorithm that simulates the transport of X-ray photons through heterogeneous volumes. It approximates the physics of X-ray projection radiography, including processes such as photoelectric absorption and higher-order scattering, including Compton and Rayleigh scattering. The quantum walk encodes all admissible photon paths into a single quantum state, enabling all scattering histories to be propagated simultaneously via the superposition principle. This quantum state representation enables flexible readout of various imaging modalities, including the primary, i.e., unscattered, image, or images exclusively containing Rayleigh and Compton scattering of specified orders. A quantitative comparison with classically computed reference simulations shows that the proposed quantum walk accurately reproduces radiographic projections, given the limitations of the underlying physical model. These results indicate that quantum circuits for X-ray transport can produce accurate radiographic images and imply that, as quantum hardware scales up, these algorithms could outperform classical Monte Carlo-based approaches in large-scale, scatter-aware virtual imaging studies.
Read the full abstract

All authors Anja Heim, Theobald Fuchs, Thomas Lang, Dimitri Prjamkov, Kilian Dremel, Stefan Kasperl, Christoph Heinzl

Combinatorics

A Family of Gauss Type Hadamard Difference Sets

A Hadamard difference set (HDS) $D$ of order $u^2$ in an abelian group $G$ satisfies $|\chi(D)|=u$ for every nontrivial character $\chi$ of $G$. We call such a character value naive if it is divisible by $u$, i.e., if it is equal to $u$ times a root of unity. All previously known abelian HDSs only have naive character values. We show that for $d\ge 1$ and $u=3d$, a group $Z_3^2\times H$, with $H$ an abelian group of order $2^{2d+2}$, contains a HDS of order $u^2$ with non-naive character values if and only if $8\le\exp H\le 2^{d+2}$. All difference sets obtained are new. The proof rests on a specific HDS in $Z_3^2\times Z_8\times Z_2$, a covering extended building set on $Z_3^2\times Z_8\times Z_4$, and a variation of the Davis-Jedwab recursive construction.
Read the full abstract

Information Theory

Spatial-Code-Domain Grouped Index Modulation: Fluid-Antenna-Assisted System Design and BER Performance Analysis

Fluid antenna systems (FASs) provide reconfigurable spatial resources within compact apertures. In this paper, we introduce code-domain grouped index modulation (CGIM) and its spatial-code-domain extension, termed SCGIM, for FA-assisted transceivers. CGIM partitions the available orthogonal spreading codes into multiple subsets and jointly maps information onto their in-phase and quadrature indices and constellation symbols. In an Rx-FAS-assisted single-input multiple-output (SIMO) link, group-wise despreading separates the orthogonal code groups for parallel detection, while receive-port selection provides spatial diversity. SCGIM further associates interleaved Tx-FA port subsets with the code subsets, with the Tx-FAS conveying spatial-index information and the Rx-FAS providing selection diversity in a multiple-input multiple-output (MIMO) link. For SCGIM, we develop maximum-likelihood (ML), staged greedy (GD), and cross-domain index message-passing (CD-IMPD) detectors. CD-IMPD exchanges soft information over a cycle-free factor graph to account for the coupling between the spatial and code indices, requiring only one inward and one outward message pass. For CGIM, the BER is derived from the joint decision regions of the despread-domain observations and averaged over the Rx-FAS selected-gain distribution under Rayleigh, Nakagami-m, and additive white Gaussian noise channels. For SCGIM, an average-BER approximation is derived from a full-pair union bound using the selected-gain density ratio and exponentially tilted quadratic-form Laplace transforms. Simulation results validate the BER analysis and show that the proposed schemes achieve lower BER and higher throughput than the considered IM schemes, while CD-IMPD achieves near-ML BER performance with lower detection complexity.
Read the full abstract

All authors Peng Zhang, Jian Dang, Yao Ge, Miaowen Wen, Ziyang Liu, Liang Wu, Zaichen Zhang, Yudong Yao

Machine Learning

Conditioned Initialization for Attention

Transformers are a dominant architecture in modern machine learning, powering applications across vision, language, and beyond. At the core of their success lies the attention layer, where the query, key, and value matrices determine how token dependencies are captured. While considerable work has focused on scaling and optimizing Transformers, comparatively little attention has been paid to how the weights of the queries, keys and values are initialized. Common practice relies on random initialization or alternatives such as mimetic initialization, which imitates weight patterns from converged models, and weight selection, which transfers weights from a teacher model. In this paper, we argue that initialization can introduce an optimization bias that fundamentally shapes training dynamics. We propose conditioned initialization, a principled scheme that initializes attention weights to improve the spectral properties of the attention layer. Theoretically, we show that conditioned initialization can potentially reduce the condition number of the attention Jacobian, leading to more stable optimization. Empirically, it accelerates convergence and improves generalization across diverse applications, highlighting conditioning as a critical yet underexplored area for advancing Transformer performance. Importantly, conditioned initialization is simple to apply and integrates seamlessly into a wide range of Transformer architectures.
Read the full abstract

Computer Vision

Single Image to Textured 3D Object Generation in Frequency Domain: From Theory to Pipeline

Single-view 3D reconstruction, also known as image-to-3D, is a persistently challenging task due to the extreme lack of information. Recently, diffusion models pre-trained on large-scale datasets served as 2D priors are used to solve the ill-posed task but suffer from color deviation and view inconsistency, which can be curbed by using diffusion models fine-tuned with 3D annotated data served as 3D priors. However, 3D priors lack high-frequency details, which cannot be solved by direct complementation with 2D priors in spatial domain for introducing erroneous low-frequency 2D prior guidance. In this paper, we revisit the characteristics of different diffusion priors from the frequency perspective. Based on our observations, we theoretically present a unified framework of hybrid optimization using multiple diffusion priors in frequency domain. Under this framework, we further propose Morpheus3D, a pipeline of 3D object generation from any single unposed image in the wild. Morpheus3D enhances 3D prior with high-pass image-prompt 2D prior guidance to reconstruct high-quality 3D objects while effectively suppressing view inconsistency, low-frequency color deviation, and high-frequency lacking problems. Both quantitative and qualitative experiments on the public and our collected datasets with complex textures show that our method exhibits significant improvements in generation quality.
Read the full abstract

Quantum Physics

Ancilla-assisted nondestructive discrimination of distributed GHZ-class states

Nondestructive quantum state discrimination is a fundamental primitive in distributed quantum information processing, where shared multipartite entangled resources need to be identified without being consumed. In this work, we present a scalable ancilla-assisted protocol for the strict nondestructive discrimination of $n$-qubit GHZ-class states distributed among distant parties. By employing multipartite GHZ ancillary states and local unitary interactions, we show that the computational-pattern and relative-phase information of the system GHZ state can be coherently mapped onto two ancillary registers while leaving the system state unchanged. We explicitly derive the discrimination rules for three- and four-qubit GHZ-class states and develop a systematic extension to arbitrary $n$. We further investigate the discrimination protocol in the presence of depolarizing and amplitude-damping noise in the ancillary system and obtain analytical expressions for the success probability and the corresponding critical noise thresholds. The protocol is also implemented using the IBM Qiskit platform to demonstrate its experimental feasibility and reproduce the expected ancillary measurement signatures. The proposed framework provides a scalable approach to the strict nondestructive discrimination of multipartite GHZ-class states and suggests a broader interpretation of the ancillary GHZ resources as probes of decoherence, opening a possible connection to quantum decoherence sensing in distributed quantum systems.
Read the full abstract

All authors Abdul Q. Batin, Suman Chand, Rajiuddin Sk, Prasanta K. Panigrahi

Numerical Analysis

A covariance-based reduced-order framework for solving acoustic scattering problems

This paper presents a physics-aware reduced-order method (ROM) for the efficient computation of wave-scattering problems. Standard model order reduction techniques, typically treating scattering as generic parameterized systems, frequently overlook the underlying physical structure, limiting their effectiveness in practice. To address this limitation, we propose an algorithmic framework that utilizes the intrinsic low-rank structure of the induced contrast source density. By modeling the incident wave as a random variable governed by a specified prior probability measure, we formulate the contrast source as a spatial random field whose covariance function captures essential spatial correlation and physical interactions. The reduced-order basis is then constructed via the Karhunen-Loève (KL) expansion, effectively extracting the dominant features from the scattering process to resolve multiple scattering scenarios. A central algorithmic contribution is the efficient reconstruction of the covariance matrix for arbitrary scatterer geometries and specified incident wave priors. To circumvent the prohibitive computational cost of assembling high-fidelity covariance matrices, we introduce a non-intrusive, physics-informed graph neural network (GNN) surrogate to learn the complex mapping from scatterer geometry to the source correlation kernel, enabling a highly efficient offline-online computational paradigm suitable for large-scale scattering configurations. Extensive numerical experiments demonstrate that the proposed framework achieves robust computational acceleration over full-order models without sacrificing accuracy.
Read the full abstract

Optics

Broad-angle photon-pair generation from high-$Q$ flatband resonant 3R-MoS$_{2}$ metasurfaces

Spontaneous parametric down-conversion (SPDC) in ultrathin optical resonant metasurfaces offers a promising platform for integrated quantum light sources. However, conventional resonances suffer from steep momentum dispersion, limiting photon-pair generation to narrow excitation and emission angles. Here, we demonstrate a high-$Q$ flatband resonant 3R-MoS$_{2}$ metasurface for high-efficiency, broad-angle SPDC photon-pair generation. By combining Brillouin zone folding with in-plane symmetry breaking, we engineer a high-$Q$ flatband resonance in the near-infrared regime, maintaining a robust $Q$-factor across a wide range of incident angles. Exploiting the strong local electric field enhancement of resonance and the large second-order susceptibility of 3R-MoS$_{2}$, the metasurface boosts the degenerate SPDC pair-generation rate with orders of magnitude enhancement over an unpatterned film of identical thickness. Crucially, the near-zero band dispersion extends the effective emission angle to $\pm 5^{\circ}$ with negligible spectral drift $\sim 0.1$ nm and high directional tolerance. Under pulsed excitation with a finite spectral bandwidth, the metasurface maintains a substantial bandwidth-averaged photon-pair yield, while demonstrating strong spectral locking of the resonant emission channel against pump wavelength detuning despite the expected trade-off in generation efficiency. This work establishes a versatile strategy for momentum-independent quantum light sources, opening new avenues for scan-free quantum ghost imaging and quantum information processing.
Read the full abstract

All authors Tingting Liu, Huifu Qiu, Xintong Shi, Jumin Qiu, Shuyuan Xiao

Computer Vision

SupGRPO: Enhancing GRPO with Matching-based Online SFT for Text Spotting

Text spotting requires both accurate text recognition and precise spatial localization. Current specialised spotters excel at predicting tight bounding boxes in natural scenes, but falter on complex or artistic text, whereas multimodal large language models (MLLMs) possess strong recognition capabilities yet remain weak at localisation. To equip the text spotter with general and powerful recognition capabilities and to maximize its localization ability, we explore two MLLM-based fine-tuning methods: Supervised Fine-Tuning (SFT) and reinforcement learning fine-tuning based on Group Relative Policy Optimisation (GRPO). An interesting finding is that SFT is less effective than GRPO at enhancing recognition, while GRPO is less effective than SFT at enhancing detection. To compensate for each other's shortcomings, we introduce a joint training strategy, SupGRPO, which simultaneously optimizes the model using both SFT and GRPO. SupGRPO employs the specially designed reward functions and develops a matching-based online SFT applied solely to coordinate tokens. It both mitigates the reward sparsity problem of GRPO and avoids the instance order dependency problem of SFT. To evaluate particularly challenging cases, we curate ATS, a dataset for artistic text spotting. Experiments demonstrate that SupGRPO improves both text recognition and detection, and attains superior performance. Our code and dataset will be released at https://github.com/Psycho-9/SupGRPO.
Read the full abstract

All authors Xudong Xie, Yuzhe Li, Jing Shi, Zhifei Zhang, Curtis Wigington, Zhaowen Wang

Number Theory

Upper bounds for arithmetic functions over polynomial values via Chebotarev–Sato–Tate distributions

We establish logarithmic upper bounds for nonnegative multiplicative functions evaluated at values of multivariable polynomials. The polynomial contribution is encoded by the permutation character on the geometric irreducible components of the corresponding hypersurface, while the arithmetic contribution is described by a class function on a joint Chebotarev–Sato–Tate group. This yields a unified framework for symmetric-power coefficients of CM and non-CM modular forms, Dedekind zeta-function coefficients, and mixed automorphic–Galois weights. As a further application, we obtain quantitative divisibility results for Fourier coefficients along polynomial values, including explicit formulas in the full residual-image and Eisenstein congruence cases.
Read the full abstract

Sound

Comparing Self-Supervised and Domain-Invariant Features for Cross-Domain Voice Phishing Detection

Voice phishing detection faces three critical challenges: real criminal recordings are unavailable due to privacy constraints; when available, only a handful of samples exist, insufficient for fine-tuning; and lightweight acoustic-only detection is needed as an alternative to large self-supervised models. We compare domain-invariant prosodic features and self-supervised representations (HuBERT, wav2vec2.0) through cross-domain evaluation-training on scenario-based actor recordings and testing on authentic criminal calls. Domain-invariant prosodic features achieve 69.5% F1 zero-shot and 71.0% with 5-shot learning. HuBERT achieves highest performance (94.2% F1, 5-shot), while wav2vec2.0 exhibits a precision-oriented detection profile (90.2% F1 with 99.4% precision, 5-shot). These findings reveal fundamental trade-offs: domain-invariant features enable zero-shot deployment when no real data exists, while SSL methods achieve higher performance but require real samples and compute.
Read the full abstract

All authors Jeongmin Lee, Seung Yun, Minkyu Lee, Ran Han, Yoonkyu Woo, Jinxia Huang

Computer Vision

One for All: Generalist Foundation Model for Cross-Sensor Skeleton Representation Learning

For learning generalizable motion representations from large-scale unlabeled data, Self-supervised learning (SSL) has become a widely adopted methodology. However, existing approaches are primarily limited by the inherent heterogeneity of skeleton data—characterized by varying joint counts, indexing protocols, and topological structures across different sensors—which typically necessitates training separate, sensor-specific, or even entirely dataset-specific models. To overcome this, we introduce SOfA (Skeleton One for All), the first generalist foundation model designed to achieve sensor-unified skeleton representation learning across diverse sensors. To accommodate the dimensional gap caused by varying joint counts, we introduce a fixed-size set of learnable Canonical Joint Slots, acting as a universal vessel that seamlessly accommodates arbitrary skeletal topologies. SOfA fills these slots via an attention mechanism that dynamically aggregates skeletal information from sensor-specific inputs. Furthermore, we resolve joint index misalignment between various sensors by introducing a Semantic Joint Embedding derived from a pre-trained text encoder, rather than relying on absolute positional embeddings. To validate our approach, we standardized ten 3D skeleton datasets for unified training. Extensive experiments demonstrate that SOfA can serve as a truly universal encoder, achieving state-of-the-art (SOTA) performance across a wide range of downstream tasks and sensor types, often outperforming dataset-specific specialist models with a single foundation model.
Read the full abstract

Statistics Theory

Network recovery from aggregated relational data: alignment and resolution

We prove that labeled edge probabilities can be recovered uniformly from complete aggregated relational data (ARD) at a given community composition if and only if the sum of trait alignment and partition resolution diverges. We study a two-community logistic network with fixed known weak signal and $K$ trait groups of size $m$. If $b_k$ is the difference between the two community counts in group $k$, alignment $\Lambda=\|b\|_2^2/m$ measures separation in the mean counts. Resolution $\rho=K/m^2$ measures the information that accumulates across groups through joint count fluctuations. Without knowing $b$, the full count likelihood gives a mean squared error bounded by $Ce^{-c(\Lambda+\rho)}+O((Km)^{-2})$ under bounded-energy activity perturbations. At exact group balance and zero activity, the minimax risk is $\exp\{-\Theta(\rho)\}$ as $\rho\to\infty$. Uniform recovery over all globally balanced community assignments holds exactly when $\rho\to\infty$. At zero activity, when the community assignment is uniform over all balanced assignments, the optimal mean squared error averaged over assignments and networks vanishes exactly when $K\to\infty$. The population count law identifies every edge probability, and full adjacency permits uniform recovery throughout the model.
Read the full abstract

2609.07077 · 7 Sep · 13 pages; 38-page supplement in anc/supplement.pdf. Retitled and substantially expanded. The two-group impossibility result for labeled edge probabilities is the $K=2$ case of Corollary 2.4. New: a necessary and sufficient criterion for uniform recovery at a given group composition, exponential risk bounds, and a recovery criterion for uniformly random balanced labels. Submitted to Bernoulli

Quantum Physics

Single-tone drive-enhanced CROT gate for bosonic quantum error correction

Bosonic error correction provides a hardware-efficient route to protected qubits for accurate quantum information processing. A critical component for error correction of rotation-symmetric bosonic (RSB) codes, like the well-known cat codes, is the two-mode controlled-rotation (CROT) operation. The CROT gate is useful because of its code-agnostic generality and its error-propagation properties. While the CROT gate can in principle be composed from existing bosonic physical primitives, the desirable properties are typically lost when composed as a sequence of imperfect primitive operations. Past work has focused on direct implementations that rely on features of specific codes. Here, we present a direct route to CROT within the circuit-QED architecture that retains its code-agnostic and error-propagation characteristics. By driving a transmon simultaneously coupled to two microwave cavities, we controllably enhance the effective nonlinearities between the microwave cavities and thus engineer the necessary two-mode interaction underlying the CROT gate. Using only a single drive frequency, we can achieve an on-off ratio sufficient for gate implementation, while minimizing mode distortions. We provide simple analytical formulas that permit the identification of potential working regimes, further refined by exact numerics, and illustrate the efficacy of our CROT approach in an error-correction example involving different RSB codes within the same circuit. Our CROT gate adds to the arsenal of direct two-mode gates available for general bosonic information processing.
Read the full abstract

All authors Shushen Qin, Long D. H. My, Adrian Copetudo, Amon M. Kasper, Yvonne Y. Gao, Hui Khoon Ng

Artificial Intelligence

A Hierarchical Consistency Framework for Auditing Retrieval-Augmented Generation Systems

Retrieval-augmented generation (RAG) is commonly evaluated by whether the final answer is correct. That test is insufficient: an answer can match its reference while the context that produced it contains a direct contradiction, leaving the contested evidence invisible to answer-only review and retrieval relevance scores. This paper presents the Hierarchical Consistency Framework (HCF), a post-hoc, model-agnostic audit of three distinct levels of a RAG process: the knowledge corpus, the final retrieved context, and the generated answer. HCF represents corpus conflicts as source-linked atomic facts, thereby identifying the documents responsible, and returns each Answer Consistency Score (ACS) with an explanation of supporting and contradictory contextual statements. We evaluate HCF on several controlled corpora spanning five domains and 100 query-corpus instances. A human evaluator compares every generated response with its supplied ground-truth response. The results show that the three diagnostic levels can dissociate: the corpus with the highest mean retrieval similarity has the lowest mean ACS, while a structurally degraded corpus performs worse at corpus level but better at answer level. Most importantly, HCF identifies contradictory retrieved evidence in several cases where the answer still matches the ground truth. HCF does not certify factual truth; it makes the evidence supporting and challenging an answer inspectable and attributable.
Read the full abstract

Algebraic Geometry

Fano Generalized Bott-Samelson Varieties

The Bott-Samelson varieties provide natural desingularizations of Schubert varieties, and their generalizations, called generalized Bott-Samelson varieties, were constructed by Nicolas Perrin as towers of locally trivial fibrations with fibers isomorphic to Schubert varieties. In this article, we give a complete characterization of Fano and weak Fano generalized Bott-Samelson varieties. As a consequence, we recover the corresponding results for ($G$-)Bott-Samelson varieties and minuscule generalized Bott-Samelson varieties.
Read the full abstract

Fluid Dynamics

A Universal Vortex-Formation Law for Insect Hovering

Flapping insect wings generate leading-edge vortices (LEVs) that produce much of the lift required for hovering, yet the frequency governing their formation has not been connected to classical vortex shedding. Roshko scaling for bluff-body wakes, together with the Sigurdson extension to wall-bounded separated flows, identifies a universal vortex-formation Strouhal-number range of approximately 0.15 to 0.17. Here we show that this same range governs LEV formation on both stationary wings and hovering insect wings. We first develop an approximate scaling for stationary flat-plate wings in which discrete LEVs are generated by periodic leading-edge perturbations, and show that maximum lift consistently occurs within the universal Strouhal range. In hovering insects, one LEV forms during each translational half-stroke, allowing the same framework to be applied to published flight data. Despite large differences in morphology, kinematics and Reynolds number, the insect data recover the same Strouhal range. These results reveal a common vortex-formation timescale linking classical separated flows, artificially generated LEVs and the natural LEVs that sustain insect hovering.
Read the full abstract

Computer Vision

Detect Anything in Graphic Design: Element-Level Rewards for Autoregressive Detection

Graphic designs, such as posters, advertisements, and infographics, are an important medium for communicating information and shaping understanding. Unlike natural images, they consist of layered elements with explicit compositional order. However, existing object detection models treat these elements as an unordered set, leaving compositional order unexploited. To address this limitation, we present Detect Anything in Graphic Design (DAD), a model that formulates graphic design detection as compositional deconstruction. It decodes elements in compositional order, using lower-layer elements to better detect higher-layer ones. The key feature of DAD is amodal detection, which predicts the full bounding box of each element, including regions occluded by elements placed above it. Building on this formulation, we propose Element Relative Policy Optimization (EleRPO), which extends GRPO from sequence-level supervision to element-level optimization. EleRPO provides fine-grained training signals that capture how each detected element contributes to overall detection quality, and works synergistically with compositional order to improve detection performance. To support training and evaluation, we build a dataset of 10 million graphic designs. Experiments show that DAD outperforms all baselines and achieves human-level performance in amodal detection, supporting effective image-to-layer decomposition. EleRPO consistently improves over GRPO across nine detection benchmarks.
Read the full abstract

All authors Jiangning Zhu, Bowen Li, Shenyu Qiao, Yima Gu, Zhao Zhang, Yuhui Yuan, Shixia Liu