Science & Mathematics

Neural operators offer an efficient data-driven approach to rapid prediction of complex automotive aerodynamic flows. However, spectral bias and other factors limit accurate reconstruction of fine-scale flow variations near complex geometries. Existing studies improve the representation and fitting of fine-scale information mainly through information extraction and training objectives. Further study is needed to determine how extracted local details can be used continually in subsequent operator computations and contribute effectively to the overall flow-field representation. We propose CIS-Transolver, a cross-layer incremental-state neural operator that extends fine-scale modeling from local information extraction to its use across layers. The method extracts fine-scale information through local multiscale analysis and encodes it as geometric moment states that are propagated and updated across layers. By combining information from previous layers with feedback to the value representations in Physics-Attention, the method allows local details to contribute continually to subsequent modeling of global physical correlations. We systematically evaluate CIS-Transolver on three public automotive aerodynamics benchmarks, ShapeNet-Car, AhmedML, and DrivAerNet++, covering surface pressure fields and volumetric pressure and velocity fields. CIS-Transolver achieves the lowest mean relative L2 error among the evaluated methods for all five prediction metrics across the three benchmarks. The relative error reductions average 8.72% across the five metrics compared with the best baseline for each metric, reaching 16.38% for ShapeNet-Car pressure prediction and 15.60% for AhmedML velocity prediction. These results provide a new approach to fine-scale flow prediction for complex geometries.

Recent advances in Gaussian splatting, 3D foundation models, and video-passthrough mixed reality offer unprecedented fidelity for capturing and blending physical and virtual worlds. However, established XR methods rely on polygon-based pipelines that treat physical and digital environments as separate domains, making seamless integration across recent novel representations difficult. Because splats lack explicit surface geometry, they impede standard graphics techniques necessary for cross-reality interactions, such as shadow or collision handling. Furthermore, while 3D foundation models can reconstruct physical objects on-the-fly into digital splats, current XR systems fall short of leveraging this capability for user experiences that transition seamlessly between physicality and virtuality. To address this, we organize XR as a multi-layered reality of interacting primitives, letting pixels and objects move across physical and digital boundaries. We contribute SplatXR, a Unity toolkit that operationalizes this structure as building blocks for interactive, immersive apps leveraging 3D Gaussian splatting: a standalone renderer, splat-to-mesh reconstruction, splat physics, splat blending, and offloaded splat object reconstruction via on-headset capture. SplatXR unites vision and graphics techniques to make splat-based XR development broadly available, and we validate it through demonstrative applications and a small developer study.

We present a controlled, large-scale evaluation of how corpus scale, document representation, and reranking affect the effectiveness and efficiency of biomedical Retrieval-Augmented Generation (RAG) systems. Retrieval collections were progressively expanded while retaining the relevant source documents, enabling consistent comparisons across corpus sizes. We compared abstract, flat full-text, and hierarchical full-text representations across first-stage retrieval, reranking, and end-to-end multiple-choice question answering. Configurations selected on development data were evaluated on five external biomedical benchmarks and a held-out test set, while retrieval latency and index storage were measured to quantify operational cost. Performance patterns observed with small or controlled collections did not consistently persist as collections grew, and the usefulness of reranking depended on the corpus and retrieval setting. RAG improved three of four model configurations across the external benchmarks, with mean gains of up to 9.6 percentage points, and improved all four configurations on the held-out test set by 10.2-30.3 points. Abstract retrieval performed best when reference evidence was abstract-aligned, whereas full-text representations were generally more effective when richer evidence was needed. Hierarchical representation matched or outperformed flat full-text retrieval in most end-to-end comparisons while requiring 67% fewer indexed units and 44-56% less storage. These findings show that corpus scale should be treated as an explicit evaluation variable in biomedical RAG, and that the most appropriate design depends on evidence format, model behavior, and operational constraints.

Earth observation satellites acquire more data than they can downlink. Extracting information on board with artificial intelligence (AI) turns the satellite into a data-reduction stage that downlinks compact products, such as masks or detections, instead of full scenes; deploying it in a mission, however, involves the whole processing chain and not only the neural network. We present a generic on-board AI processing chain along three axes, with wildfire segmentation as the worked example among ship detection and cloud segmentation. (i) The AI model is tuned (spectral-band selection, input normalization), quantized to 8-bit integers (int8) and ported to an AMD Versal VCK190, with its excellent segmentation performance preserved on the board. (ii) A modular tiler-inference-mosaicer workflow, reused unchanged across applications, brings the tile-level model to scene level, turning a full acquisition into a scene-level mask or detection list. (iii) Each step is packaged as a unikernel and orchestrated on board by ORCHIDE, a cloud-native orchestrator. The chain is characterized end to end, from the standalone execution on the target to its execution under orchestration with parallel workflows on a representative test bench. Beyond AI inference on individual tiles, this approach aims at increasing the maturity of on-board AI deployment by addressing a complete processing chain, from the acquired scene to the mission-level execution.

Recursive self-improvement (RSI) requires agents to diagnose system bottlenecks and coordinate improvements across training and inference. We refer to the agents' capability to jointly optimize training data, the trainer, and the inference harness as cross-component co-design. We introduce VibeTraining-Bench, a benchmark for evaluating this capability across tasks spanning terminal operation, software engineering, multi-turn tool use, and mathematical reasoning. Within each task, research agents start from an identical initial agent system under a fixed resource budget. Guided by validation feedback, they iteratively refine data, trainers, and inference harnesses, then select and submit a final agent system for evaluation on the test set. To prevent reward hacking, we enforce three safeguards: test data isolation, separate research and grading containers, and network restrictions for research agents. Our evaluation results show that frontier research agents substantially improve target agent systems. For example, on SWE-bench Verified, GPT-6-Astra raises the target agent system's performance from 32.9% to 53.7%, comparable to the human-developed instruct model's 55.4%. However, there is still room to improve their ability to co-design and iteratively refine data, trainer, and harness to achieve cumulative performance gains. We hope that VibeTraining-Bench will serve as a unified platform for advancing RSI, deepening our understanding of cross-component optimization for agent systems. Code and data are available at https://github.com/D2I-ai/VibeTraining-Bench.

Reconfigurable intelligent surface (RIS)-empowered transmitters offer a promising basis for hardware-efficient multi-antenna architectures in next-generation wireless networks. In particular, combining these transmitters with low-complexity reception techniques such as receive antenna selection (RAS) can further simplify the receiver architecture. In this paper, we investigate the impact of practical phase-shift constraint on RIS-empowered multi-antenna systems employing maximal ratio combining (MRC) and RAS. Since each RIS element can apply only a single phase-shift at a given time, the system performance degrades substantially when multiple receive antennas are active simultaneously. To quantify this effect, closed-form outage probability expressions are derived and asymptotic analyses are carried out over Rician fading channels. We show that the practical phase-shift constraint of RIS-based transmitters can degrade the diversity order of MRC and in addition, RAS outperforms MRC in the examined system models, contrary to the conventional performance ordering of these techniques.

Emergency vehicles such as police vehicles, fire engines, and ambulances require a stable and efficient route optimization to reach life-critical events in the shortest times possible. However, the dynamic nature of urban traffic makes this ideal scenario much more challenging, causing critical life-threatening delays. Traditional civilian navigation systems cannot account for real-time traffic changes alongside vehicle-specific constraints, creating inadequate routing scenarios. This research presents the Constraint-Aware Real-Time Routing System (CARRS), a unified software architecture tailored for emergency coordination. The CARRS architecture is engineered to fuse together classical pathfinding algorithms such as A*, Dijkstra, and Bidirectional A*, with Constraint Satisfaction Problem methodologies. The Design Science Research Methodology is the research framework, and prototyping is the iterative process. The artifact was empirically evaluated using OpenStreetMap data for major metropolitan road networks, combined with simulated dynamic traffic scenarios. Performance metrics focused on route optimality, constraint adherence, rerouting responsiveness, and the potential for future system scalability. Findings revealed the system's capability to generate optimal routes for operationally relevant subgraphs in under 35 seconds, with incremental rerouting responsiveness below 20 seconds. Furthermore, the system achieved more than 90% adherence to vehicle-specific constraints, making the CARRS architecture viable for real-time dynamic routing requirements. This study delivers a routing platform that bridges artificial intelligence-driven optimization with software engineering operational compliance, presenting a practical transformative step toward adaptive emergency vehicle coordination in dynamic urban environments.

Non-invasive brain-computer interfaces are in practice limited by signal-to-noise ratio, and recent state-of-the-art methods address this limitation by aggregating over repeated trials of the same stimulus. Yet multi-trial aggregation has been treated inconsistently across the literature: some work averages brain signals at the model's input, other work averages predictions at the model's output, and neither treats aggregation as an architectural choice in its own right. We provide the first systematic comparison of aggregation strategies for non-invasive neural decoding, evaluating six paradigms (including logit averaging during training and three attention-based variants we introduce) across three datasets, four tasks, and two group sizes. We find that aggregation choice matters substantially: training-time logit averaging, the simplest method we propose, outperforms all others in 6 of 7 settings. On the phoneme classification setting of the 2025 PNPL competition, it sets a new state of the art, significantly outperforming the competition winner by 17.4 absolute percentage points in balanced accuracy. We further draw a theoretical connection between logit averaging and joint ensemble training, which has been shown to underperform independent ensembling in the deep learning literature; our results call into question recent explanations of why joint ensemble training fails. Together, these findings establish multi-trial aggregation as a key design choice for non-invasive neural decoding and identify training-time logit averaging as a simple, effective default.

The continuous evolution of next-generation power systems, characterized by the high penetration of renewable energy resources and dynamic topology variations, poses formidable challenges to traditional N-1/N-2 contingency analysis in terms of computational efficiency and scalability. While deep learning-based methods offer accelerated inference, existing studies predominantly treat contingency analysis as a pure regression problem, often overlooking the physical essence of grid operation and engineering requirements of real-world scenarios. To bridge this gap, this paper proposes a physics and prompt-guided graph foundation model (PhyPro-FM) for generalizable and engineering-viable contingency analysis. The proposed architecture integrates graph neural networks with a learnable global query mechanism and, notably, introduces a prompt-guided learning paradigm. By leveraging historical snapshots as few-shot prompts, the model dynamically captures correlations between target and historical operating conditions, significantly enhancing generalization capabilities under unseen topologies. Furthermore, we design a physics-informed training mechanism that prioritizes branch power flow errors over nodal voltage errors, aligning the optimization objective with critical grid security constraints. To ensure engineering applicability, a physics-informed inference mechanism is also developed, incorporating a residual-based correction module and a specialized post-processing strategy for singular branches to circumvent numerical instability. Comprehensive experiments demonstrate that PhyPro-FM outperforms state-of-the-art methods in accuracy and generalization, validating its superior physical consistency and robustness for practical engineering applications.

GNSS signal acquisition is the process of detecting and estimating signal parameters by exploiting correlation operations that provide processing gain and enable the detection of signals below the noise floor. The design of an acquisition engine requires balancing achievable sensitivity, computational complexity, memory requirements, power consumption, and implementation complexity. In particular, increasing coherent and non-coherent integration times improves sensitivity but can significantly increase receiver processing requirements, limiting its applicability in low-power and resource-constrained receivers. This paper presents a comparative analysis of L1-band signal acquisition complexity by evaluating the required integration time and corresponding implementation requirements for different signal designs under equivalent acquisition performance targets. The analysis considers the impact of signal characteristics, including spreading codes, secondary codes, modulation, and signal bandwidth, on coherent and non-coherent acquisition performance. For a given attenuation level and target probability of false alarm and detection, the required integration times are derived and translated into acquisition complexity metrics, including arithmetic operations, processing time, and memory requirements. The analysis is applied to Pulsar X1, a LEO-enabled PNT signal, and representative GNSS L1 signals to determine under which attenuation, assistance, and integration conditions the higher received power of LEO-PNT translates into reduced receiver acquisition workload. By relating the required integration to arithmetic operations, processing time, and memory requirements, the study provides a receiver-level comparison relevant to both conventional and snapshot-based receivers, where reducing acquisition processing can directly contribute to lower energy consumption.

Most existing studies on pinching-antenna systems optimize the antenna positions or activation states under a prescribed waveguide deployment, where mutually parallel waveguides are commonly adopted. However, it remains unclear whether, and under what conditions, such a deployment is optimal from a communication-performance perspective. To fill this gap, this paper investigates long-term waveguide deployment optimization for pinching-antenna systems over a rectangular service region. We first consider the fundamental two-waveguide case and derive a closed-form sufficient condition, jointly determined by the waveguide height and the service-region aspect ratio, under which a deployment parallel to the longer side of the service region is globally optimal. We then consider the general multi-waveguide case, where the orientation and offset of each waveguide are jointly optimized without prescribing any deployment pattern, and develop an adaptive exchange algorithm for the resulting nonconvex optimization problem. We further show that, within the parallel class, the waveguides should be aligned with the longer side of the service region, while in the large-height regime, the globally optimal unrestricted deployment asymptotically concentrates all waveguides on the centerline along the longer side. Numerical results validate the proposed algorithm and reveal how the preferred deployment geometry depends on the waveguide height and service-region shape. In particular, nonparallel deployments can provide noticeable gains in intermediate-height regimes, whereas parallel or center-concentrated deployments become favorable in other regimes.

Multimodal survival prediction jointly analyzes histopathological whole-slide images (WSIs), genomic profiles, and clinical variables to stratify patients by risk. However, existing approaches often inadequately account for granularity differences across modalities, particularly in spatial organization and feature dimensionality. Within pathology, aggregating WSI patch features without preserving their spatial arrangement can obscure tissue architecture; across modalities, directly concatenating low-dimensional clinical variables with high-dimensional visual and genomic features may limit the contribution of clinical information to prediction. To address these challenges, we propose HGASurv, which introduces a hierarchical WSI encoding module and a clinical-guided graph propagation module for granularity-aware multimodal integration. The hierarchical WSI encoding module maps WSI patch features onto a spatial grid using their original coordinates and applies parallel convolutions to construct representations at patch, local-neighborhood, and regional scales. Genomic signature embeddings then query these scale-specific WSI representations through cross-modal co-attention, producing signature-conditioned morphological features. The clinical-guided graph propagation module uses clinical variables to define patient-similarity edges and propagates fused image-genomic node embeddings with graph attention, integrating clinical information through topology rather than feature concatenation. Across three TCGA cohorts, HGASurv achieved the strongest overall C-index performance, with ablation studies supporting the complementary contributions of the hierarchical WSI encoding and clinical-guided graph propagation modules. This design shows that matching representation and integration mechanisms to modality-specific granularity improves multimodal survival prediction.

Embodied agents are increasingly capable of perceiving, planning, and acting autonomously in dynamic environments, demonstrating strong performance on complex cooperative tasks. Recent advances typically assume that cooperative tasks are known in advance. In real-world applications, however, hidden-cooperation tasks are more common: the need for cooperation only becomes apparent after an agent observes and interacts with task-related targets. In this paper, we first present HiddenCoop, a benchmark for effectively evaluating agent cooperation in hidden-cooperation tasks. We further propose Discovery-Probability Graph Planning (DPGP), a method for enabling effective agent cooperation in such tasks. At the core of DPGP is a discovery-probability graph that encodes the probability distribution over the number and locations of task-related targets, facilitating effective and efficient information exchange among agents. A search algorithm is then used to derive action plans that minimize expected costs. Experiments show that DPGP achieves cooperative-task success rates of 99.3% on C-WAH and 58.3% on TDW-MAT, surpassing all compared baselines, while also reducing overall embodied execution time and substantially lowering LLM token usage. Our code is available at https://github.com/hnu-vis/DPGP.

This paper proposes an integrated setup for a multi-scale asymmetric collaboration scenario in the context of a remote-expert use case, focusing on the expert side. In this system, the remote expert (with an exocentric view in Mixed Reality (MR)) uses a Tangible User Interface (TUI) to precisely position and rotate their avatar within a scaled tabletop World-in-Miniature (WIM). Inside the WIM, the expert's avatar and the operator share the same scale and egocentric view, while the expert retains a global tabletop overview of the plant to support collaboration. Interviews with industry experts during the participatory design phase revealed the need for an egocentric view inside the plant on the remote-expert side to enable full collaboration. We designed and studied three interaction-visualisation combinations to integrate and control the expert avatar's gaze: VR (fully immersive with direct head control), Head (floating screen with direct head control), and Touch (on-avatar screen with touchpad control). We evaluated these methods with 29 participants across two fault-detection tasks: Same Machine, where both collaborators inspect the same machine, and Twin Machine, where they inspect identical machines in separate locations. The VR condition, which combines immersion, direct head control, and a full-scale view, yielded the strongest overall performance and user experience, while the tabletop avatar setup was well received for collaboration. Our findings inform the design of gaze-control techniques for asymmetric MR-VR collaboration and reveal challenges and opportunities for future work.

Document annotation is a fundamental yet labor-intensive step in building document information extraction systems. While Optical Character Recognition (OCR) and Large Language Models (LLMs) can automate parts of this process, we argue that annotation systems should be judged not by how completely they remove human effort, but by how effectively they redirect it toward review and correction. We propose a modular human-in-the-loop architecture that integrates OCR-based extraction, LLM-based semantic pre-annotation, and human validation in a single editable workflow, and present \textbf{GAIDA}, a research prototype serving as its reference implementation. GAIDA supports interchangeable OCR engines, LLM backends, and prompt configurations, and enables stage-specific measurement of annotation effort. In an initial single-annotator, stage-specific study on the FUNSD scanned-form and WildReceipt retail-receipt benchmarks, OCR assistance reduced region-creation and transcription time by up to \textbf{80.9\%}, and reviewing LLM pre-annotations reduced semantic-labeling time by up to \textbf{73.7\%}, despite imperfect pre-annotations of 78.86\% and 86.98\% Micro F1 on FUNSD and WildReceipt, respectively. Finally, we reuse the same implementation, adapting only OCR and prompt configuration, to build and openly release \textbf{AraReceipt}, a manually annotated dataset of 100 Arabic receipt images labeled with 25 key-information classes. The results indicate that imperfect automated outputs provide practical value when presented as editable suggestions under human control.