arXiv:2606.13429v1 Announce Type: new Abstract: We propose a novel iterative solution framework for the unsteady Stokes equations in the pseudo-stress formulation. When solving this class of problems by using implicit time-integration schemes, standard solvers suffer from deteriorating convergence properties for small time steps, independently of the chosen space discretisation method. This is due to the singular modes of the dev-dev operator. For this reason, we introduce a computational framework obtained by combining a deflated Conjugate Gradient method with a W-cycle multigrid scheme that employs a Restricted Additive Schwarz smoother. The key point is to choose the deflation subspace so that the inner system to be solved within a deflated Conjugate Gradient scheme corresponds to a Laplace problem defined on the singular modes of the original dev-dev operator. This results to be independent of the spatial discretisation method and allows one to use efficient multigrid iterative solvers. Numerical experiments show that the proposed strategy significantly accelerates the Conjugate Gradient convergence and provides stable performance with respect to the time step, confirming its robustness for solving linear systems in the pseudo-stress framework.
Science Journals
arXiv:2606.13431v1 Announce Type: new Abstract: Maritime chokepoints concentrate shipping traffic. Disruptions to this traffic can have a widespread impact on the global economy. However, the way in which these impacts are shaped by the shipping sector's adaptive behavior is not well understood. Here, we introduce an empirically calibrated full-scale agent-based model of the global commercial shipping fleet, representing 35,954 active ships moving among 1,651 ports. We use the model to quantify how rerouting changes arrival losses under chokepoint closures. Static route exposure alone does not predict realized losses. In the adaptive model, rerouting reduces losses at some directly exposed ports, while delayed vessel cycles create losses at later port calls and in dependent regions. Cumulative net shipping-day losses therefore continue to rise with closure duration because longer routes keep ships delayed after the initial adjustment. Each additional closure day reduces global shipping arrivals by 3.0% for Suez and 7.7% for simultaneous Suez, Panama, and Malacca closures. These losses are unevenly distributed in exposed regions and ports. Disruptions with known duration show different loss profiles from unexpected shocks with unknown duration, revealing that end-date information can reduce avoidable short-run losses. The results show that chokepoint risk is a dynamic problem of routing, timing, and regional exposure and not a static property of maritime-network topology.
arXiv:2606.13432v1 Announce Type: new Abstract: Cloning camera motion from reference videos is an important task in video generation, as videos provide intuitive and precise control. Existing methods either directly use parametric representations that fail to handle multi-shot generation or synthesize cross-paired data, which suffer from data scarcity, resulting in poor performance in complicated camera motion cloning. To address these issues, we introduce a general camera motion representation that encodes cameras as grid motion videos. This camera grid represents the camera parameters visually and supports the integration of diverse trajectories for multi-shot video generation. Building upon this, we propose OmniDirector, a unified framework trained on a million-scale camera grid-video pairs that coordinates characters, actions, and cameras to provide director-level control for multimodal diffusion transformers. Furthermore, we design a novel hierarchical prompt expansion agent that harmoniously integrates different control signals by systematically describing camera motion and visual content through understanding signal relationships. Extensive experiments demonstrate the superior performance and outstanding controllability of our framework. Project page: https://ymlinfeng.github.io/OmniDirector.github.io/
arXiv:2606.13434v1 Announce Type: new Abstract: Numerical algorithms for computing electronic structure of incommensurate 2D materials using ab initio models is critical for predicting material properties and guiding experiment. For bilayers, momentum space and continuum models have been introduced to approximate observables of ab initio tight-binding models using a momenta description despite the lack of periodicity in the tight-binding model required for Bloch theory. A similar structure has been introduced for double-incommensurate trilayers using a continuum model, where the three lattices are all mutually incommensurate. However, this description leads to a four-dimensional lattice space, and numerical convergence of the density of states was observed to have poor convergence. In this work, we introduce a momentum space framework for double incommensurate trilayer graphene, and introduce an efficient truncation scheme of the four-dimensional lattice to drastically improve convergence of the density of states and momentum local density of states (a parallel object to classical band structure). We implement this algorithm on an ab initio model of twisted trilayer graphene and validate convergence estimates. We further verify numerically that the momentum space algorithm, inherently higher order than the continuum model as it is an exact transformation of the tight-binding model, captures altered band behavior near the flat bands at magic angles.
arXiv:2606.13435v1 Announce Type: new Abstract: Human communication is inherently multimodal, where language is often accompanied by non-verbal cues such as gestures to convey intentions. However, current Vision-Language-Action (VLA) models treat robotic manipulation as a pure text-driven task, overlooking the important role of gestures in Human-Robot Interaction (HRI). This often leads to inaccurate intent grounding and unreliable manipulation when language instructions are ambiguous or underspecified. To address this challenge, we propose GIVE (Gesture Intent via Visual-Semantic Enhancement), an effective approach that enhances pre-trained VLA models with human gesture understanding without architectural modifications. Specifically, GIVE incorporates gesture information through two complementary pathways: a visual pathway that overlays hand skeletons and fingertip rays onto robot observations for explicit object grounding, and a semantic pathway that generates high-level descriptions of human gestures and task instructions for robust intent grounding. By jointly leveraging visual and semantic guidance, GIVE enables VLA policies to better associate gestures with manipulation behaviors and adapt to dynamic interaction intents. In real-world HRI experiments, GIVE substantially outperforms the baseline, improving target object recognition accuracy by 40% and overall task success rate by 80%, while demonstrating strong robustness and generalization to unseen spatial layouts and diverse participants.
arXiv:2606.13438v1 Announce Type: new Abstract: Retrieval-Augmented Generation (RAG) has become a common approach for improving the factuality of Large Language Models (LLMs), yet its reliability remains highly sensitive to how external evidence is retrieved and used. Semantically equivalent queries with different syntactic forms may lead to different retrieval results, while irrelevant or misleading documents can further induce hallucinated answers. Existing multi-path reasoning methods improve robustness by sampling multiple candidate answers and applying voting- or confidence-based selection, but they still face two limitations: diversity is often injected through uncontrollable decoding randomness, and answer evaluation is usually confined to a single query-induced evidence view. To address these limitations, we propose a Cross-Query Consistency Hypothesis: correct answers tend to maintain high confidence across semantically equivalent but syntactically diverse queries, whereas noise-induced hallucinations exhibit unstable confidence under such query variations. Based on this hypothesis, we introduce CQC-RAG, a framework that co-designs query-level diversity injection with cross-query consistency evaluation. CQC-RAG rewrites the original question into diverse but meaning-preserving queries, reranks a shared document pool to construct query-conditioned reasoning contexts, applies an evidence-grounded protocol to extract answer-evidence pairs and selects answers according to their confidence stability across these contexts. This design enables self-evaluation without external supervision and does not rely on expanded retrieval coverage. Experiments on four open-domain question answering benchmarks show that CQC-RAG outperforms the strongest previous multi-query baseline by +4.76 pp EM on TriviaQA and +9.12 pp EM on MuSiQue, validating the effectiveness of cross-query consistency for filtering noise-induced hallucinations.
arXiv:2606.13443v1 Announce Type: new Abstract: Neural operators have emerged as a powerful data-driven approach for solving time-dependent PDEs. Among recent advances, memory-augmented neural operators explicitly incorporate past states and have achieved remarkable performance under low-resolution observation settings. However, existing approaches apply a fixed memory weight regardless of observation conditions, such as resolution or physical parameters, limiting their adaptability. Our preliminary experiments reveal that optimal memory weight varies with resolution and viscosity, implying that a fixed memory weight cannot simultaneously optimize performance across diverse settings. We propose AMGFNO, which dynamically modulates memory weight through a learnable gate. On the Kuramoto-Sivashinsky and Burgers' equations, AMGFNO achieves 55-79% nRMSE reduction over at low resolution, with the learned gate value automatically decreasing from $\bar{g} \approx 0.7$ to near-zero as resolution increases.
arXiv:2606.13444v1 Announce Type: new Abstract: Graph clustering - partitioning the node set of a graph into disjoint subsets that reflect some latent information - is a fundamental problem as it finds applications in a myriad of different scenarios. While this classic problem has been tackled for decades by different communities, a recent variation of the problem driven by real data considers the scenario where nodes have attributes that are also informative. This has triggered novel methods that simultaneously leverage network information (edges) and node information (attributed) in the design of novel clustering algorithms. This work proposes a novel framework that builds on prior works that have applied graph neural networks (GNN) to graph clustering. The proposed framework operates in rounds of self learning in a fully unsupervised setting. In each round, a GNN generates representations for nodes that are used to cluster the nodes. This clustering influences the graph used to generate the node representation in the next round. Moreover, a context graph built in each round using the original graph is used to generate the node representations. Empirical results show that the proposed methodology extracts information from both network edges and node attributes in synthetic data, outperforming algorithms focused solely on the network or attributes when neither are very informative. Multiple rounds of learning also improve the performance and always outperforms a long single round of training (i.e., classic GNN graph clustering). When considering real datasets, empirical results indicate that the proposed methodology is competitive to state-of-the-art methods when cluster sizes are balanced.
arXiv:2606.13451v1 Announce Type: new Abstract: Diffusion models have seen wide adoption for 3D molecular generation, yet they offer no principled signal of when a generated molecule is likely to be of low quality. We propose a post-hoc method for estimating per-sample uncertainty in pretrained molecular diffusion models. Building on a Laplace approximation of the denoising network, we measure the variability of the noise prediction across the generation trajectory. Empirically, we show that the resulting uncertainty score is informative of sample quality, exhibiting a negative correlation with established sample-level quality metrics. We further study how the proposed uncertainty score can be used to filter generated samples, improving model performance via test-time scaling.
arXiv:2606.13452v1 Announce Type: new Abstract: Novelty is a crucial metric for assessing the quality of academic papers. Scholars strive to highlight the novel aspects of their work, particularly in the title, abstract, and introduction. Peer review, serving as the gatekeeper of scientific rigor, rigorously evaluates the novelty of papers, yet a cognitive gap may exist between author self-promotion and reviewer evaluation. To investigate this, we analyzed 15,328 academic papers published in Nature Communications from 2016 to 2021, along with their peer-review comments. We found that both reviewers and authors emphasize result-oriented innovation, with reviewers adopting a more comprehensive evaluation perspective. Furthermore, by examining promotional intensity against inherent paper novelty, we found that its effect depends on the paper's actual innovation level. Highly innovative papers benefit from stronger promotional language, receiving more positive evaluations. We also found that promotional language significantly correlates with reviewer disagreement on novelty specifically for papers of moderate innovativeness, whereas it has negligible impact for papers with either very high or very low novelty. This reveals how promotional language operates most prominently in the gray area of academic evaluation.
arXiv:2606.13480v1 Announce Type: new Abstract: The posture of the vocal folds produced by laryngeal muscle activation plays a central role in determining the dynamics of voice production. Abnormal vocal fold configurations are frequently associated with inefficient phonation and a variety of voice disorders. Although diverse glottal closure patterns have been observed clinically, the biomechanical mechanisms governing their dynamic behavior and resulting phonatory characteristics remain incompletely understood. Moreover, existing numerical models that incorporate the effects of the intrinsic musculature on posturing and glottal conformation are computationally expensive, which limits their suitability for large-scale parametric investigations. In this work, we introduce a computationally inexpensive vocal fold (VF) model wherein the body and cover VF layers are treated as a composite beam and a coupled membrane, respectively. Intrinsic laryngeal muscle activation, in addition to positioning the arytenoid cartilages and cricothyroid joint, introduces moments at the boundaries of the structure that influence glottal conformation. The model produces phonatory characteristics that are qualitatively consistent with those reported in high-fidelity finite-element models and clinical studies, thereby supporting its predictive capability while offering substantial computational advantage. The proposed framework provides biomechanical insights into the influence of incomplete glottal closure on phonation dynamics and may serve as a computationally tractable tool for investigating mechanisms underlying certain voice disorders.
arXiv:2606.13482v1 Announce Type: new Abstract: We present a multilevel B-spline-based fast integral method for the solution of the electric field integral equation (EFIE), combining fast Fourier transformation (FFT)-compatible kernel interpolation with robust high-order interpolation. Existing FFT-accelerated global Lagrange-based approaches rely on equidistant interpolation points and can, therefore, suffer from Runge-type instabilities at high interpolation orders, limiting robust high-accuracy compression. In contrast, B-splines on equidistant knot vectors overcome these instabilities and enable robust high-order interpolation for accurate matrix compression. Replacing Lagrange interpolation by B-spline interpolation is, however, non-trivial: B-spline coefficients do not coincide with function values at the interpolation points, and the associated sampling matrices can become ill-conditioned. To address these challenges, we introduce a knot-removal stabilization strategy, combined with exact interlevel transfers based on knot insertion, yielding accurate, well-conditioned multilevel interpolation. Moreover, we propose a factorization strategy that preserves the null space of the scalar potential operator up to machine precision and is compatible with low-frequency preconditioning techniques. Numerical results for both canonical and realistic geometries demonstrate robust high-order interpolation without the breakdown observed for Lagrange-based approaches and confirm $\mathcal{O}(N)$ complexity.
arXiv:2606.13511v1 Announce Type: new Abstract: The frequency bands between 7 and 24 GHz, also known as upper midband or Frequency Range (FR) 3, are being considered as an enabler of 6th Generation (6G) mobile networks. This portion of the spectrum exhibits different propagation characteristics compared to frequencies above 24 GHz, while also offering the potential to provide larger bandwidth allocations for mobile systems than those available in the sub-6 GHz range. 6G technology and spectrum policy, however, will need to guarantee coexistence with the incumbents that already use these frequency bands, which include a variety of services, from radiolocation to satellite-based communications, remote sensing, and radioastronomy. In this paper, we consider the challenge of coexistence between 6G terrestrial systems and satellite incumbents in different portions of the FR3 bands. Using a large-scale 3D model of a terrestrial deployment in the city of Boston and an open-source ray tracing solution, we evaluate the level of Radio Frequency Interference (RFI) that tens of terrestrial Next Generation Node Bs (gNBs) generate toward satellites at different elevation angles. Our model, based on realistic obstruction, clutter, diffraction, and reflections, shows that sidelobes and Non-Line-of-Sight (NLoS) paths can significantly contribute to RFI. Besides directionality, the spatial distribution of gNBs also plays a key role in defining the RFI levels, suggesting that a careful design and operation of terrestrial deployments can create coexistence opportunities.
arXiv:2606.13515v1 Announce Type: new Abstract: World Action Models (WAMs) present a promising paradigm for robotic control via video prediction. However, current WAMs suffer from fundamental spatial bottlenecks: standard text inputs introduce referential ambiguity in cluttered scenes, while unstructured RGB predictions lack semantic grounding and remain biased by task-irrelevant backgrounds. To overcome these limitations, we introduce MaskWAM, an object-centric world-action model. By jointly integrating masks as both explicit inputs and predictions via a unified Mixture of Transformers (MoT), MaskWAM unlocks robust policy generalization. This design provides two key benefits: (1) predicting future masks yields object-centric semantic supervision that suppresses visual noise, significantly enhancing even standard text-conditioned WAMs; and (2) coupling this predictive supervision with first-frame visual prompts, such as target object masks, establishes a precise spatial anchor that substantially reduces language ambiguity. Crucially, as WAMs are inherently vision-driven architectures, direct mask conditioning yields substantially stronger guidance than text alone, establishing a precise and robust paradigm for manipulating unseen objects. Evaluations on LIBERO, RoboTwin, and real-world tasks demonstrate that MaskWAM significantly outperforms baselines in both language-clear and language-ambiguous tasks.
arXiv:2606.13516v1 Announce Type: new Abstract: Frequency Modulated Continuous Wave (FMCW) radar systems traditionally rely on Fourier-based methods, such as the Fast Fourier Transform (FFT), to estimate target range and velocity. While computationally efficient, these approaches require storing and processing large blocks of data, which can become a bottleneck in memory-constrained or low-latency applications. In this work, we propose a neuromorphic-inspired signal processing method based on adaptive resonate-and-fire (ARF) neurons formulated as a discrete-time dynamical system. Each neuron dynamically adjusts its internal frequency to match dominant frequency components of the input radar signal, enabling direct estimation of target ranges and velocities without computing the full frequency spectrum. The proposed model operates in a sample-by-sample fashion, resulting in memory requirements that scale with the number of tracked targets rather than the signal length. A feedback mechanism is also introduced to enable multiple neurons to lock on distinct frequency components in multi-target cases. Results on simulated and experimental data demonstrate that the method can successfully track multiple targets. Compared to conventional FFT-based approaches, the proposed method offers reduced memory usage proportional only to the number of tracked targets, making it suitable for resource-constrained and edge-based radar applications.
arXiv:2606.13528v1 Announce Type: new Abstract: Saliency-guided training is a paradigm in visual recognition that encourages models to focus on the most relevant image regions during learning. While its application in biometric presentation attack detection (PAD) has shown strong benefits in robustness and generalization, adoption is often limited by the high cost, domain specificity, and limited scalability of existing saliency acquisition methods, such as human annotations over a limited dataset. We present a novel, cost-efficient, and highly-scalable approach to saliency acquisition using maps inspired by classical dimensionality reduction techniques: PCA and LDA. Our proposed methods generate saliency maps directly from raw training data, requiring no human annotation nor domain knowledge. We contextualize the effectiveness of these saliency sources in three saliency-explored domains (iris PAD, synthetic face detection, fingerprint PAD) and demonstrate its scalability in two saliency-novel domains (fingerprint vein PAD and ID card PAD). Across all domains tested, models trained using dimensionality reduction-sourced saliency maps exceed baseline and sometimes SOTA saliency methods without any resource investment or domain-specific tooling. Our findings overcome an important yet unaddressed barrier to saliency-guided training for biometric attack detection and beyond.
arXiv:2606.13529v1 Announce Type: new Abstract: Post-traumatic stress disorder (PTSD) in veterans is characterized by persistent hyperarousal and comorbid anxiety and depressive symptoms that are difficult to monitor and manage outside clinical settings. Thirteen veterans participating in a Project Hero cycling event in Texas were randomized by computer-generated sequence in a naturalistic setting to two arms: (1) digital intervention plus physical activity, or (2) physical activity only, plus a third at-home monitoring control cohort consisting of 7 veterans selected from the broader Project Hero veteran community. Continuous smartwatch sensing combined heart rate and accelerometer features to detect hyperarousal events, which were confirmed in real time by participants. Weekly self-report measures of anxiety, depression, and PTSD severity were collected. Generalized additive mixed models characterized nonlinear trajectories over time. Baseline-normalized hyperarousal trajectories differed significantly across conditions, with the digital intervention group (n=7) showing structured stabilization compared to late-study escalation in the physical-only group (n=3). Both cycling groups exhibited acute symptom improvements during the endurance event; however, the digital intervention group demonstrated a higher overall maintenance of gains. The at-home control group (n=4) showed gradual symptom declines. Perceived precision of ML detections varied substantially across individuals and was positively associated with symptom severity, with higher-severity participants confirming a greater proportion of detected events. These results suggest that coupling wearable detection with digital self-management tools may support stabilization of hyperarousal and symptom improvement while emphasizing the importance of personalization and human-centered design in wearable mental health systems.
arXiv:2606.13532v1 Announce Type: new Abstract: Cloud-computing relies on large-scale networks which are inherently complex systems. In this paper, we present a novel approach to root cause analysis (RCA) of cloud network incidents, leveraging graph-based causal discovery techniques. Our method addresses the limitations of rule-based automation by introducing a spatiotemporal grouping strategy and an automation ontology to reduce the dimensionality of the problem. We construct a causal graph from binary time series data using bivariate Granger causality and conditional independence tests. For inference, we introduce a probabilistic method that assigns edge-specific conditional probabilities as a function of time lag, allowing for interpretable, time-aware root cause scoring via causal graph traversal. We evaluated the system using a labeled dataset of 35 production incidents from a major cloud provider. The model successfully recalled the correct root cause in 85.7% of incidents and produced an exact match in 74.3%. In production, the deployed system has been used in over 800 real-world incidents, with positive qualitative feedback from network engineers. These results highlight the practicality of a data-driven, causal approach to RCA in dynamic and large-scale operational environments.
arXiv:2606.13543v1 Announce Type: new Abstract: Can a learned model capture how faults propagate through a large-scale network and use this knowledge to causally attribute customer impact to its underlying root cause? Existing root cause analysis techniques often rely on static rules, correlation heuristics, or topology-local reasoning, which struggle to generalize in dynamic environments where faults propagate across complex physical and logical dependencies. We present NetCause, a self-supervised learning-based framework that models network incidents as graph-temporal processes and uses counterfactual simulation to rank candidate root causes. This approach produces an interpretable ranking of root cause hypotheses and integrates naturally with operator-defined mitigation and remediation actions. We train the model on over 1,500 incidents collected over six months from a leading cloud provider's production network and evaluate it on 31 expert-labeled incidents. NetCause consistently improves root cause ranking quality in the regime most relevant to operational decision-making, achieving a 16.1% accuracy improvement over a rule-based heuristic baseline. While training is computationally intensive, inference is lightweight, requiring only seconds of GPU runtime per incident (well below typical telemetry collection latencies).
arXiv:2606.13549v1 Announce Type: new Abstract: Singular volume integral operators associated with constant-coefficient partial differential operators extend the applicability of potential theory to inhomogeneous problems, for example arising from nonlinearities or variable coefficients. Typically the PDE kernels in these operators give rise to singularities at all $\mathcal{O}(1/h^3)$ volume discretization/evaluation points in a mesh of characteristic size $h$, while the slowly-decaying nature of such kernels give rise to long-range interactions that require coupling to fast summation algorithms. The presented method uses Green's identities to regularize a wide variety of both scalar-valued and vector-valued volume integral operators by use of a certain regularizing volume density interpolant. The analysis shows how the regularizing effect of the interpolant is global in the sense that the interpolation quality increases in an exactly compensatory fashion as the distance to the Green's function singularity decreases. High-order convergence estimates with tabulated simplex quadratures are established, including with exact representation of curved domains.
arXiv:2606.13580v1 Announce Type: new Abstract: Event-based vision has drawn increasing attention owing to its distinctive properties, including ultra-high temporal resolution and extreme dynamic range. Recent works have introduced it to video super-resolution (VSR) to enhance flow estimation and temporal alignment. In contrast, this paper shifts the focus of event signals from motion refinement to texture enhancement in VSR. We propose EvTexture++, the first event-driven framework dedicated to texture enhancement in VSR. It leverages high-frequency spatiotemporal details from events to improve texture recovery. EvTexture++ incorporates a customized texture enhancement branch, along with an iterative texture enhancement module that progressively exploits high-temporal-resolution event information for texture restoration. This enables gradual refinement of texture regions across iterations, yielding more accurate and detailed high-resolution outputs. Besides intra-frame texture recovery, large motions could degrade inter-frame temporal consistency, particularly in texture regions, leading to texture flickering. To mitigate this, we further exploit the continuous-time motion cues of events to enhance temporal consistency, introducing a temporal texture alignment module that estimates event-guided texture-aware flow for precise inter-frame texture alignment. Moreover, EvTexture++ is designed as a plug-and-play tool to flexibly boost the performance of existing VSR models. Experiments on five datasets demonstrate that EvTexture++ achieves state-of-the-art performance. When integrated into recent VSR models, it yields significant improvements, with gains of up to 1.55 dB in PSNR on the texture-rich Vid4 dataset. Code: https://github.com/DachunKai/EvTexture.
arXiv:2606.13581v1 Announce Type: new Abstract: Despite raising concerns about the mental health effects associated with the usage of TikTok, little is known about how related content is framed by creators and received by audiences. We collect the content of 28,341 TikTok videos and 80,130 comments from Mental Health Awareness Month (May) in 2023 and 2024 via the TikTok Research API, and study how the tone of awareness varies across topics and years. We characterize "tone" as the emotional and interpersonal framing of mental health discourse, operationalized through sentiment and toxicity measures. We extract topics from video text using BERTopic and log-odds keywords, then quantify topic-conditioned sentiment (XLM-T) and toxicity (Detoxify) separately for video transcriptions and comments. Sentiment captures the affective valence of content, while toxicity reflects the presence of harmful or abusive language. We find a stable set of recurring themes across years, spanning clinical conditions, emotional disclosure, self-care, and campaign-oriented content, with engagement highly skewed toward a small subset of topics. All sentiment and toxicity analyses are computed separately for video content and comments, allowing us to distinguish between content production and audience reception. Sentiment in videos is often negative for emotionally charged topics, while comments tend to shift toward more mixed or positive polarity, especially for suicide prevention. Toxicity is low in median overall, but exhibits longer-tailed outliers in comments than in videos that are more pronounced in comments and concentrated in specific topics (e.g., "Duet", "Suicide Prevention", and "Psychisch"). Overall, our results provide a topic-level decomposition of mental health discourse on TikTok during awareness-month campaigns.
arXiv:2606.13596v1 Announce Type: new Abstract: Planar turbulent mixing layers, formed by the interactions of two parallel streams with different velocities, have been studied far more than three dimensional (3D) turbulent mixing layers, in which the incoming streams are skewed, and not parallel. Yet many practical shear flows are 3D. Here, we develop and validate an experimental methodology to generate and characterize skewed turbulent mixing layers and to quantify how mean-flow skew modifies mixing layer dynamics. We introduce skew with a spanwise deflection of the mean flow using turning vanes mounted near the trailing edge of a splitter plate, and we use cross-wire anemometry to investigate the downstream evolution of the flow. Relative to the planar configuration, the skewed mixing layer exhibits systematic reductions in both mean and turbulent quantities, with deviations reaching approximately 40\%. Despite these quantitative differences, the fundamental characteristics of the mixing layer remain largely unchanged. Mean-velocity profiles collapse under similarity scaling, shear-layer thicknesses retain approximately linear downstream growth, and Reynolds-stress profiles preserve their characteristic near-Gaussian form. Townsend's structure parameter, which quantifies the efficiency of turbulent momentum transport, remains approximately invariant between the planar and skewed configurations, in contrast to skewed turbulent boundary layers, wherein comparable mean flow skewing reduces the parameter by approximately 30\%. These results indicate that mean flow skew modifies turbulent mixing layers quantitatively while exerting only a secondary influence on their underlying dynamics. This study establishes a controlled experimental framework and empirical benchmark for future investigations of three-dimensional free-shear turbulence.
arXiv:2606.13624v1 Announce Type: new Abstract: Large language models (LLMs) have enabled time series (TS) analysis by jointly modeling numerical observations and textual context through a shared token interface. However, TS tokens and prompt tokens exhibit fundamentally different information structures, making uniform token processing inefficient. In this paper, we study token efficiency in TS language modeling from an asymmetric-token perspective. We show that TS tokens have highly uneven spectral contributions, where many tokens share redundant frequency patterns while a small subset preserves critical temporal evidence. We also observe that prompt-token influence attenuates with model depth, suggesting that full prompt retention across all layers is unnecessary. Based on these findings, we develop an adaptive token budgeting framework that compresses TS tokens via frequency-domain structure and progressively reduces prompt tokens across layers. Experiments across forecasting, classification, imputation, and anomaly detection demonstrate up to \textit{\textbf{7.68$\times$}} inference acceleration and performance gains in \textit{\textbf{78\%}} of evaluated settings, showing the effectiveness of asymmetric token compression for scalable TS foundation models.
arXiv:2606.13625v1 Announce Type: new Abstract: Vehicle color recognition is an important cue for vehicle identification in surveillance systems, especially when license plates are illegible due to low resolution, occlusion, motion blur, or poor illumination. However, real-world vehicle color distributions are highly imbalanced, making overall accuracy insufficient to assess performance on rare but operationally relevant colors. This paper presents a comprehensive study of vehicle color recognition under severe class imbalance using UFPR-VeSV, a challenging real-world surveillance dataset. We investigate synthetic minority-class augmentation through two off-the-shelf generative strategies: text-conditioned image generation with RunDiffusion/JuggernautXL and image-conditioned color editing with Gemini 2.0 Flash. The curated synthetic data are combined with modern visual representations, loss reweighting, learning-rate scheduling, color-safe augmentation, foreground-aware preprocessing, and ensemble fusion. The bestperforming approach achieves 94.6% micro accuracy and 79.7% macro accuracy, improving macro accuracy by 8.2 percentage points over recent literature. A manual error analysis further shows that many remaining failures are visually ambiguous even for human annotators, highlighting the practical limits of color-based vehicle identification in unconstrained surveillance imagery. The generated images and source code are publicly available at https://github.com/viniciusorru/vcr-synthetic