Searcharxiv⌕ Search

arXiv subjects

Jing Wu

Publications and source records attributed to Jing Wu.

At least 19 recordsLinked to original sources

Stochastic Diversity Windows: Diagnosing Noise Influence in Robotic Diffusion Sampling

Diffusion models are increasingly adopted in robotics and control due to their capability to express multiple feasible solutions to a given task. However, the role of stochasticity in the diffusion model-sampler across denoising remains poorly understood. We introduce a per-timestep counterfactual diagnostic that tests how changing the sampled noise affects the final output. Across five model-sampler configurations spanning three robotics benchmarks, diverse architectures, domains, and conditioning mechanisms, we reveal highly varied temporal sensitivity profiles, where noise-induced deviations can amplify, partially contract, or undergo near-total attenuation downstream. From these profiles, we define stochastic diversity windows (SDWs): compact intervals that identify where sampler stochasticity has the strongest measured influence on the final output. Guided by these windows, we apply two heuristic interventions: targeted stochastic amplification, which increases sample diversity exploration of the solution space, and an SDW-hybrid sampler, which replaces DDPM steps with DDIM skips in low-impact regions, resulting in network evaluation reductions of up to 40-60% and wall-clock time by up to 85%, with no consistent changes detected in feasibility, cost, or diversity across held-out tasks.

cs.RO↗

An evidence-traceable benchmark of clinical reasoning in large language models

Benchmark scores for medical large language models (LLMs) are difficult to interpret when scoring criteria cannot be traced to clinical evidence. Here we developed Med Evidence Bench (MEB; https://www.medevidencebench.org.cn/), comprising 1,536 physician-constructed items across three levels: guideline knowledge, standardized clinical scenarios and real-world laboratory scenarios. Each criterion stating a medical claim is linked to a versioned source-text span. Across ten LLMs, the Evidence Trace Score (ETS) was 0.975-0.997 for guideline knowledge, 0.292-0.658 for standardized scenarios and 0.182-0.447 for real-world scenarios, and rankings varied across levels. For real-world scenarios, scores for identifying clinical information (mean, 0.770) exceeded those for matching evidence (0.165) or formulating recommendations (0.160). On 1,296 criterion-level judgments, automated grading agreed with the majority vote of three physicians (macro-F1, 0.930; 95% confidence interval, 0.915-0.945), although agreement was lower for evidence application. These findings from Chinese-language materials indicate that guideline recall does not ensure patient-specific evidence use and do not establish clinical effectiveness.

cs.CL↗

Photon-Counting Readout via Electro-Optic Microwave-to-Optical Transduction for Axion Haloscopes

Axion dark matter searches require highly sensitive detection schemes with high scan rate to efficiently explore the parameter space. Building upon the conventional Sikivie haloscope, where axions convert into microwave photons inside a magnetic-field-immersed resonant cavity, we present a photon-counting scheme based on an electro-optic microwave-to-optical transducer. Through a comprehensive theoretical model under realistic experimental constraints, we find that the optimal operating regime employs an overcoupled microwave cavity, trading peak transduction efficiency for reduced transducer noise. Furthermore, combining parameters already demonstrated in existing devices, in a device with optimized extrinsic coupling rates, would yield a scan-rate enhancement of 5-10 relative to quantum-limited homodyne detection. Improvements in quality factors, coupling rates, and thermal management push this enhancement beyond 50. While our quantitative model is benchmarked at 9 GHz, the scheme adapts well to haloscopes across the 5-12 GHz range, where shrinking cavity volumes severely degrade the achievable scan rate. These results demonstrate the potential of electro-optical photon detection as a high-speed detection technique for axion dark matter searches.

hep-ph↗

Learning from Fleet Experience: Collective Experience Learning Across 96 Robots

Robot deployment generates experience that can support policy improvement, but continued reliance on expert corrective actions makes learning difficult to scale across growing fleets. We propose learning from fleet experience: using deployed robots' own executions as a source of supervision for further policy improvement. We call this paradigm Collective Experience Learning (CEL), in which robots pool execution experience to update a shared policy that returns to the fleet. Following expert-supervised initialization, we realize CEL through failure-guided experience retrieval fine-tuning (FERT). FERT uses failure contexts to retrieve relevant segments from successful trajectories and prioritizes their recorded actions for supervised updates, without collecting additional expert demonstrations or corrective actions. Operators provide binary episode-outcome labels and environment resets. A fault-tolerant asynchronous training system (FATS) coordinates robot execution, cloud inference and policy optimization despite local interruptions. Across five real-world manipulation tasks, the resulting policies achieve success rates of 93.3-99.2%, evaluated over 120 trials per task on robots excluded from training data collection. With collection time per robot held fixed, larger fleets provide more experience and generally improve performance across the evaluated sizes. On multi-stage phone packaging, learning with 96 robots raises success from 79.2% to 98.3%. These results support CEL as a scalable approach to improving robot policy reliability through experience generated during deployment.

cs.RO↗

GE-Act 2.0: Pretraining and Scaling a World-Action Model for Robotic Manipulation

World-action models (WAM) predict future states to guide robot actions, enabling learning from both action-free video and action-labeled interaction. Most inherit pretrained video generators, leaving WAM pretraining and scaling underexplored. We introduce Genie Envisioner Act 2.0 (GE-Act 2.0), a world-action model whose trainable generative and action components are all initialized from scratch on manipulation data. It combines a control-oriented autoencoder (CoAE), a single-step visual planner (SVP), and an inverse dynamics model (IDM). CoAE retains action- and instruction-relevant information under aggressive compression, while SVP produces a complete future state in one differentiable pass, so visual planning and inverse dynamics can be pretrained separately on complementary data. The components are then jointly trained with knowledge-aligned selective optimization (KASO), which reduces mismatched supervision by selecting only predicted futures judged behaviorally compatible with the recorded action. We evaluate pretrained checkpoints directly, without per-task fine-tuning, on 100 tasks across 20 manipulation skill groups with held-out scenes, backgrounds, lighting, and object instances. Scaling co-training data from 300 to 30,000 hours raises success from 17.1% to 44.1% on G1-OP and from 13.4% to 31.1% on G2-90D; despite comprising less than 2% of the co-training data, G2-90D improves by 17.7 points, suggesting cross-embodiment transfer. Gains span 19/20 and 18/20 skill groups, and skill-specific coverage strongly correlates with zero-shot out-of-distribution (OOD) success (Pearson r=0.80; Spearman rho=0.85). Under the same protocol, the model grounds object, color, shape, and position references in at least 90% of trials and follows explicit instructions even when they conflict with an already-committed behavior or a conventional scene association.

cs.RO↗

BEE: Intervention-Adaptive Real-World Reinforcement Learning with Vision-Language-Action Models

Vision-language-action (VLA) models handle long-horizon manipulation, yet success hinges on a few precision-critical phases where millimeter-scale errors undo all prior progress. Online reinforcement learning (RL) can optimize exactly these actions, but free exploration is far too costly on real robots, which makes human corrections indispensable. However, existing online RL methods for VLAs either cannot incorporate such corrections or fold them into undifferentiated supervision. Yet human corrections are not uniformly noisy but reliable along some action dimensions and variable along others. Building on this, we introduce BEE, an intervention-adaptive framework for real-world RL on a frozen VLA that lets the policy go BEyond Expert imitation. We formulate human corrections not as actions to reproduce but as evidence about a constraint: a Correction Model predicts how a human would correct a given VLA proposal and how consistent the correction is along each action dimension. This predicted consistency sets the per-dimension tightness of a constraint on policy optimization. Where corrections are consistent the policy stays close to the human, and where they vary, the constraint relaxes. We evaluate BEE on three real-world manipulation tasks and one LIBERO-Pro simulation task at a matched online-data budget. BEE attains the highest success rate on every task, 91.2% on average against 57.5% for RLT and 42.1% for DSRL, and the lowest human intervention rate on all real-world tasks.

cs.RO↗

Transduction-Enabled Superconducting Quantum Repeater: Toward Deterministic Entanglement Distribution with High-Fidelity Gates

Long-distance entanglement distribution is hindered by photon loss in optical fibers and the nocloning theorem. Optical quantum repeater (QR) protocols rely on Bell state measurements (BSMs), they are intrinsically limited to probabilistic photon operations and fail 50% of the time. We propose a hybrid approach to building quantum repeaters that combines the high transmission speed of photonic qubits in optical fiber with the high-fidelity quantum processing capabilities enabled by superconducting circuits. The transduction-enabled superconducting QR (TESQR) architecture eliminates the need for probabilistic BSMs and allows deterministic processing operations. The TESQR framework always yields a final state at the remote nodes rather than aborting on photon loss, manifesting deterministic entanglement distribution within certain parameter regimes. We evaluate the performance by assessing output-state fidelities and success probabilities of entanglement distribution using realistic noise models. Additionally, we integrate an entanglement purification scheme and evaluate the performance through numerical simulations in QuTiP environment. Our results show that, for entanglement swapping, the proposed scheme improves the entanglement distribution rate by an average of 63% and by up to 159% compared with photonic-only architectures. Moreover, after purification, the end-to-end fidelities exceed 0.8 over distances up to 20 km.

quant-ph↗

Universal Optimization and Tighter Fidelity Bounds for Approximate Quantum Error Correction

Approximate quantum error correction (AQEC) extends the framework of discrete- and continuous-variable quantum error correction beyond the Knill-Laflamme (KL) conditions, where the recovery performance is quantified by entanglement fidelity. Recent studies have enabled efficient evaluation of near-optimal entanglement fidelity using transpose-channel recovery. Yet, determining the global optimal recovery map and its entanglement fidelity for general codes beyond the KL conditions remains a major computational challenge. Direct optimization becomes prohibitive as the number of noise Kraus operators grows rapidly with system size, and existing approaches lack rigorous guarantees for reducing this optimization to a tractable dimension. Here, we derive an explicit characterization of the optimal environmental state of complement channel, which transforms the optimization over recovery channels into an equivalent optimization over quotient unitaries. For a broader class of codes that satisfy only the orthogonality part of the KL conditions, we show that the optimal recovery map admits an explicit analytical form. Building on this form, we derive novel rigorous lower bounds of entanglement fidelity that strictly improve upon the transpose-recovery bound. We further develop a novel recovery strategy based on principle components, and derive a rigorous bound on the error introduced by noise truncation. Our approach enables efficient searches for approximate recovery maps for AQEC codes, avoiding the need to optimize over the full Kraus-operator space.

quant-ph↗

Magnetic Moments and Radiative Transitions of the $T_{cc}(3875)^+$ and its partner states

We systematically investigate the magnetic moments (MMs) and radiative decay widths of S-wave doubly heavy tetraquark states based on chromomagnetic interaction (CMI). They provide a complementary probe to distinguish compact from molecule structures in addition to the mass spectrum and strong decay properties. Using the CMI eigenvectors, we compute the MMs and M1 transition rates for the tetraquarks in the compact configuration. We predict the MM of the observed $T_{cc}(3875)^+$ with $I(J^P)=0(1^+)$ to be 0.45 $μ_N$ when treating it as a compact tetraquark state, whereas the MM of the $0(1^+)$ $DD^*$ molecule is about $-0.07\,μ_N$. Five radiative transition channels related to the $T_{cc}(3875)^+$ are identified with widths ranging from $6.07$ keV to $306.37$ keV. Our results show that the MMs of the $J^P=1^+$ $bc\bar{q}\bar{q}^\prime$ ($q/q^\prime=u,d,s$) and $J^P=1^+$ $QQ\bar{n}\bar{s}$ ($Q=b,c; n=u,d$) states are influenced by the diquark-spin mixing. Through the analyses of radiative transitions between different tetraquark states, we find that such processes in the $QQ\bar{n}\bar{n}^\prime$, $QQ\bar{s}\bar{s}$, and $bc\bar{n}\bar{s}$ cases may serve to reveal the tetraquark structures of the initial or final states. We also define the magnetic coupling matrices characterizing the MMs of the tetraquark system with $J^P=1^+$, with which the range of MM can be constrained. The present study provides a valuable reference point for the search of exotic states in future particle physics experiments.

hep-ph↗

OS-Marathon: Benchmarking Computer-Use Agents on Vast-Horizon, Repetitive Tasks

Vast-horizon, repetitive workflows are common in daily routines, e.g., processing expense reports from a stack of receipts, organising a collection of PDF annotations into structured notes, and are tedious for humans, with execution length scaling with the volume of data to process. They are, however, well-suited to autonomous agents, since their structured, recurring sub-workflows follow a logic that can be systematically learned. Yet no existing benchmark evaluates agents specifically on vast-horizon, repetitive capabilities, leaving the problem largely underexplored. To bridge this gap, we establish OS-Marathon, comprising 100 vast-horizon, repetitive tasks across 5 scenarios and 10 domains, on which our evaluation reveals that leading state-of-the-art (SOTA) CUAs struggle substantially. We further find that introducing a task orchestrator to decompose the workflow into per-instance subtasks fails to mitigate the challenge: errors accumulate and propagate across solver agents, indicating that naive decomposition is insufficient for these tasks. We then explore a cost-friendly personalisation strategy, i.e. GraphDemo, that adapts general agents to such tasks from a single human demonstration of the recurring sub-workflow logic. Extensive experiments show both the real-world challenge of vast-horizon, repetitive tasks and the improvement provided by human demonstration in this setting. Project website: https://os-marathon.github.io/.

cs.CV↗

SpatioLM: Towards General Physical Spatial Intelligence in Vision-Language Models

Vision-Language Models (VLMs) perform well on commonsense reasoning tasks but struggle with visual spatial reasoning. Most existing solutions introduce extra 3D prior inputs or external spatial encoders, which increase complexity and degrade the underlying VLMs' general-purpose capabilities after spatial fine-tuning. To this end, we propose a parameter-efficient \textit{\textbf{Spatio}-vision \textbf{L}anguage \textbf{M}odels (SpatioLM)}, that enhances spatial intelligence without extra 3D prior inputs or third-party spatial encoders. Concretely, we design a plug-and-play and non-invasive spatio-vision module that elicits the spatial knowledge inherent in VLMs. Furthermore, we innovatively leverage pseudo depth and camera information as supervision to guide the model in learning physically coherent representations. Extensive experiments show that SpatioLM achieves significant improvements in diverse tasks, including spatial perception and understanding while effectively limiting the degradation of general capabilities. Notably, the model achieves an impressive score of 71.6 on the VSI-Bench (the first model to surpass 70). In addition, it attains competitive performance when transferred to embodied manipulation tasks. Code is available at \href{https://github.com/xiaomi-research/spatio-lm}{\faGithub~spatio-lm}.

cs.CV↗

Object Detection for Autonomous Driving in Chinese Rural Scenes: An Experimental Study on Real-Synthetic Data Mixing and Model Evaluation

Currently, autonomous driving object detection models face significant data scarcity and generalization challenges when navigating complex Chinese rural traffic scenarios. To address these limitations, we propose a novel real-synthetic mixed object detection dataset tailored specifically for Chinese rural roads and systematically evaluate the performance of 13 mainstream detectors under different real-to-synthetic data ratios, thereby providing empirical evidence for model selection and data strategy design in rural autonomous driving scenarios. Our dataset combines real-world images captured in Weishi County, Henan Province, with parameterized virtual scenes generated via Unreal Engine. To accurately reflect the unique realities of rural traffic, we define a comprehensive 14-category object system encompassing region-specific elements such as electric tricycles, low-speed vehicles (LSVs), and roadside stalls. Under a unified training protocol, we systematically evaluate 13 mainstream detectors -- spanning the YOLOv5, YOLOv8, YOLO11, and YOLO26 series, as well as RT-DETR-L -- across three data configurations: an all-real baseline, a 1:0.5 real-to-virtual mix, and a 1:1 mix. Experimental results demonstrate that a moderate injection of synthetic data (1:0.5 ratio) effectively enhances detection performance, with YOLO11m achieving the highest mAP@0.5 of 0.758. However, a higher proportion of synthetic data (1:1) introduces domain shifts that offset the benefits of data scaling. While most models reliably identify distinct local vehicles, significant perceptual bottlenecks remain for long-tail, non-standard objects like stalls and railings. This research provides crucial empirical evidence and novel insights for model selection and synthetic data strategies, facilitating the practical deployment of autonomous driving perception systems in rural areas.

cs.CV↗

Quantitative Homogenization of PDEs with Neumann boundary conditions: a probabilistic approach

In this paper, we study quantitative homogenization for viscosity solutions of multi-scale semilinear second order partial differential equations (PDEs) on convex domains with Neumann boundary conditions. To this aim we use the probabilistic approach by studying the quantitative homogenization of backward stochastic differential equations (SDEs) associated with slow-fast systems of reflected SDEs.

math.PR↗

Beyond the LUMIR challenge: The pathway to foundational registration models

Medical image challenges have played a transformative role in advancing the field, catalyzing innovation and establishing new performance benchmarks. Image registration, a foundational task in neuroimaging, has similarly advanced through the Learn2Reg initiative. Building on this, we introduce the Large-scale Unsupervised Brain MRI Image Registration (LUMIR) challenge, a next-generation benchmark for unsupervised brain MRI registration. Previous challenges relied upon anatomical label maps, however LUMIR provides 4,014 unlabeled T1-weighted MRIs for training, encouraging biologically plausible deformation modeling through self-supervision. Evaluation includes 590 in-domain test subjects and extensive zero-shot tasks across disease populations, imaging protocols, and species. Deep learning methods consistently achieved state-of-the-art performance and produced anatomically plausible, diffeomorphic deformation fields. They outperformed several leading optimization-based methods and remained robust to most domain shifts. These findings highlight the growing maturity of deep learning in neuroimaging registration and its potential to serve as a foundation model for general-purpose medical image registration.

eess.IV↗

Negative temperature coefficient of Gilbert damping in magnetic bilayers

The Gilbert damping of magnetic materials is an important magnetic parameter that determines the switching speed and energy dissipation of spintronic devices. In simple metals, the intrinsic Gilbert damping increases with temperature and diverges near the Curie temperature as a result of spin fluctuations. Here we present atomistic simulations and experimental measurements showing surprising and opposite behavior in Py/Nd bilayers, where the Gilbert damping decreases with increasing temperature. The effect arises because of the enhanced damping at the interface as a result of spin pumping, where elevated temperatures cause a dynamic separation of the interfacial and bulk magnetization during relaxation. Furthermore, the temperature dependence of the damping can be controlled by varying the thickness of the Nd capping layer. Our findings present a new spintronic effect that can be used to modify the dynamic properties of nanoscale materials and devices for enhanced energy efficiency or with improved switching dynamics.

cond-mat.mes-hall↗

Functional renormalization group study of the jet quenching parameter near the QCD critical end point

We investigate the jet quenching parameter $\hat{q}$ in the QCD phase diagram within a QCD-assisted low-energy effective theory using the functional renormalization group (fRG). Following the formalism that relates $\hat{q}$ to the spectral functions of the chiral order-parameter field, we compute the $σ$ and $π$ meson contributions to $\hat{q}$ at finite temperature and baryon chemical potential from analytically continued mesonic two-point functions. We find that $\hat{q}$ receives appreciable contributions mainly above the chiral phase boundary and exhibits a pronounced enhancement at large baryon chemical potential as the chiral crossover sharpens toward the critical end point (CEP), a behavior consistent with the picture of partonic critical opalescence (PCO), a pronounced enhancement of jet transverse momentum broadening induced by the critical $σ$ field fluctuations.

hep-ph↗

SurgRFO: Foundation Model Based Compositional Synthesis of Critical Retained Foreign Objects in Intraoperative Chest X-rays

Critical retained foreign objects (RFOs) on intraoperative chest radiographs are rare but high-risk events. Their scarcity limits robust automated detection model training and generalization. We introduce SurgRFO, a two-stage synthesis framework for generating realistic RFO-present intraoperative chest X-rays. In Stage 1, a Roentgen chest X-ray foundation model is fine-tuned on surgical-domain images to generate realistic RFO-free backgrounds that preserve anatomy, indwelling lines and tubes, and intraoperative imaging characteristics. In Stage 2, a lightweight generator trained on localized RFO patches from limited positive cases synthesizes diverse RFO instances, which are composited onto generated backgrounds using conditional Poisson fusion to improve photometric consistency. We evaluate SurgRFO through (i) a blinded clinician study assessing realism and clinical plausibility, and (ii) downstream detection experiments in which synthesized data are used to augment Faster R-CNN, YOLOv8, and RetinaNet. SurgRFO consistently improves sensitivity at low false-positive-per-image (FPPI) operating points on internal and external test sets. Clinician ratings indicate that the synthesized images achieve realism comparable to real intraoperative images. Ablation analyses further examine fusion strategies and synthesis scale. Ethical safeguards for synthetic surgical data are also discussed.

eess.IV↗

Self-Supervised Contrastive Learning for Cardiac MR Sequence Classification

Vision Transformer (ViT) models, utilizing self-attention mechanisms, have demonstrated robust generalization capabilities across various vision tasks, including image classification. However, these models, typically pretrained on general public datasets, often lack the specialized domain knowledge necessary for medical imaging applications. In this study, we investigate the adaptation of ViT models, specifically for cardiac magnetic resonance (MR) images, using an in-house dataset. We found that pretrained ViT features do not effectively transfer to the cardiac MR domain. To overcome this limitation, we introduce an adaptation strategy that utilizes image-based self-supervised contrastive learning, demonstrating superior performance compared to traditional supervised training approaches. Moreover, our adapted ViT model exhibits strong generalization to external MR datasets such as BraTS and ADNI. Through ablation studies, we further investigate the impact of batch size and dataset scale on performance. Ultimately, our adapted model achieves classification AUC exceeding 0.75 across the four most common cardiac MR sequences.

cs.CV↗