arXiv subjects

Jun Li

Publications and source records attributed to Jun Li.

At least 127 records · Page 7Linked to original sources

arXiv2026

A Particle-in-Cell Simulation Framework for Thomson Scattering Analysis in Inertial Confinement Fusion

In inertial confinement fusion (ICF), Thomson scattering (TS) is a widely used diagnostic technique for probing plasma conditions. We present a first-principles numerical approach to obtaining scattered light signals of ion acoustic features with high resolution in angle and frequency space using particle-in-cell simulations under typical ICF conditions. Our method demonstrates good agreement with existing theories for thermal collective TS. In the super-thermal collective regime, the results align with theory when the driven plasma modes are well-matched in wave vectors to the probe and collecting beams. Moreover, we also find that TS signals can remain significant even under imperfect wave-vector matching-a result that contradicts the conventional expectation that the TS spectrum strictly follows the plasma density spectrum. We attribute this discrepancy to a beating wave mechanism arising from the interaction between the probe beam and driven plasma density modulations. Our work thus provides a practical framework for interpreting TS signals from driven ion modes, a common yet complex feature in ICF plasmas.

physics.plasm-ph↗

arXiv2026

Low-Latency Federated Fine-Tuning for Large Language Models Over Wireless Networks

Recently, federated large language models (LLMs) have drawn significant attention thanks to coupled capabilities of LLMs and federated learning (FL) that address privacy concerns in collaborative fine-tuning. However, due to large-scale parameters of LLMs, existing federated LLM fine-tuning frameworks incur significant challenges in resource-constrained clients characterized by heterogeneous computing capabilities and random wireless channels. To address this issue, we propose a joint client-specific pruning and bandwidth allocation (JCPBA) framework for federated LLMs to improve the fine-tuning efficiency over the wireless networks. Specifically, we formulate a fine-tuning latency minimization problem by jointly optimizing pruning rates and bandwidth allocations. Furthermore, we solve this optimization problem using a block coordinate descent method. Extensive experiments on the datasets of Yahoo Answers and GSM8K demonstrate that the proposed framework significantly reduces wall-clock fine-tuning time compared with state-of-the-art baselines and gains equal or lower test loss at the cost of lower computation and communication overhead.

cs.DC↗

arXiv2026

SPLIT: Sparse Incremental Learning of Error Dynamics for Control-Oriented Modeling in Autonomous Vehicles

Accurate, computationally efficient, and adaptive vehicle models are essential for autonomous vehicle control. Hybrid models that combine a nominal model with a Gaussian Process (GP)-based residual model have emerged as a promising approach. However, the GP-based residual model suffers from the curse of dimensionality, high evaluation complexity, and the inefficiency of online learning, which impede the deployment in real-time vehicle controllers. To address these challenges, we propose SPLIT, a sparse incremental learning framework for control-oriented vehicle dynamics modeling. SPLIT integrates three key innovations: (i) Model Decomposition. We decompose the vehicle model into invariant elements calibrated by experiments, and variant elements compensated by the residual model to reduce feature dimensionality. (ii) Local Incremental Learning. We define the valid region in the feature space and partition it into subregions, enabling efficient online learning from streaming data. (iii) GP Sparsification. We use bayesian committee machine to ensure scalable online evaluation. Integrated into model-based controllers, SPLIT is evaluated in aggressive simulations and real-vehicle experiments. Results demonstrate that SPLIT improves model accuracy and control performance online. Moreover, it enables rapid adaptation to vehicle dynamics deviations and exhibits robust generalization to previously unseen scenarios.

cs.RO↗

arXiv2026

OMGs: A multi-agent system supporting MDT decision-making across the ovarian tumour care continuum

Ovarian tumour management has increasingly relied on multidisciplinary tumour board (MDT) deliberation to address treatment complexity and disease heterogeneity. However, most patients worldwide lack access to timely expert consensus, particularly in resource-constrained centres where MDT resources are scarce or unavailable. Here we present OMGs (Ovarian tumour Multidisciplinary intelligent aGent System), a multi-agent AI framework where domain-specific agents deliberate collaboratively to integrate multidisciplinary evidence and generate MDT-style recommendations with transparent rationales. To systematically evaluate MDT recommendation quality, we developed SPEAR (Safety, Personalization, Evidence, Actionability, Robustness) and validated OMGs across diverse clinical scenarios spanning the care continuum. In multicentre re-evaluation, OMGs achieved performance comparable to expert MDT consensus ($4.45 \pm 0.30$ versus $4.53 \pm 0.23$), with higher Evidence scores (4.57 versus 3.92). In prospective multicentre evaluation (59 patients), OMGs demonstrated high concordance with routine MDT decisions. Critically, in paired human-AI studies, OMGs most substantially enhanced clinicians' recommendations in Evidence and Robustness, the dimensions most compromised when multidisciplinary expertise is unavailable. These findings suggest that multi-agent deliberative systems can achieve performance comparable to expert MDT consensus, with potential to expand access to specialized oncology expertise in resource-limited settings.

cs.CL↗

arXiv2026

Fully Differentiable Bidirectional Dual-Task Synergistic Learning for Semi-Supervised 3D Medical Image Segmentation

Semi-supervised learning relaxes the need of large pixel-wise labeled datasets for image segmentation by leveraging unlabeled data. The scarcity of high-quality labeled data remains a major challenge in medical image analysis due to the high annotation costs and the need for specialized clinical expertise. Semi-supervised learning has demonstrated significant potential in addressing this bottleneck, with pseudo-labeling and consistency regularization emerging as two predominant paradigms. Dual-task collaborative learning, an emerging consistency-aware paradigm, seeks to derive supplementary supervision by establishing prediction consistency between related tasks. However, current methodologies are limited to unidirectional interaction mechanisms (typically regression-to-segmentation), as segmentation results can only be transformed into regression outputs in an offline manner, thereby failing to fully exploit the potential benefits of online bidirectional cross-task collaboration. Thus, we propose a fully Differentiable Bidirectional Synergistic Learning (DBiSL) framework, which seamlessly integrates and enhances four critical SSL components: supervised learning, consistency regularization, pseudo-supervised learning, and uncertainty estimation. Experiments on two benchmark datasets demonstrate our method's state-of-the-art performance. Beyond technical contributions, this work provides new insights into unified SSL framework design and establishes a new architectural foundation for dual-task-driven SSL, while offering a generic multitask learning framework applicable to broader computer vision applications. The code will be released on github upon acceptance.

cs.CV↗

arXiv2026

Stability conditions in the mathematical Gauged Linear Sigma Model

The theory of Mixed-Spin-P (MSP) fields was introduced by Chang-Li-Li-Liu for the quintic threefold, aiming at studying its higher-genus Gromov-Witten invariants. Chang-Guo-Li has successfully applied it to prove conjectures including the BCOV Feynman rule, Yamaguchi-Yau's polynomiality conjecture and the Holomorphic Anomaly Equation. Meanwhile, Fan-Jarvis-Ruan introduced a mathematical theory of Gauged Linear Sigma Model (GLSM), associating a counting theory to a GIT quotient with a super-potential, under suitable assumptions. This paper provides a common generalization of both works, by introducing new stability conditions in the mathematical GLSM. We show that our stability condition guarantees the separatedness and properness of the cosection degeneracy locus in the moduli. It generalizes the MSP fields construction to more general GIT quotients, including Calabi-Yau global complete intersections in toric varieties. This hopefully provides a geometric platform to effectively compute their higher-genus Gromov-Witten invariants.

math.AG↗

arXiv2026

A boundedness theorem for principal bundles on curves

Let $G$ be a reductive group acting on an affine scheme $V$. We study the set of principal $G$-bundles on a smooth projective curve $\mathcal C$ such that the associated $V$-bundle admits a section sending the generic point of $\mathcal C$ into the GIT stable locus $V^{\mathrm{s}}(θ)$. We show that after fixing the degree of the line bundle induced by the character $θ$, the set of such principal $G$-bundles is bounded. The statement of our theorem is made slightly more general so that we deduce from it the boundedness for $ε$-stable quasimaps and $Ω$-stable LG-quasimap.

math.AG↗

arXiv2026

Beyond Rejection Sampling: Trajectory Fusion for Scaling Mathematical Reasoning

Large language models (LLMs) have made impressive strides in mathematical reasoning, often fine-tuned using rejection sampling that retains only correct reasoning trajectories. While effective, this paradigm treats supervision as a binary filter that systematically excludes teacher-generated errors, leaving a gap in how reasoning failures are modeled during training. In this paper, we propose TrajFusion, a fine-tuning strategy that reframes rejection sampling as a structured supervision construction process. Specifically, TrajFusion forms fused trajectories that explicitly model trial-and-error reasoning by interleaving selected incorrect trajectories with reflection prompts and correct trajectories. The length of each fused sample is adaptively controlled based on the frequency and diversity of teacher errors, providing richer supervision for challenging problems while safely reducing to vanilla rejection sampling fine-tuning (RFT) when error signals are uninformative. TrajFusion requires no changes to the architecture or training objective. Extensive experiments across multiple math benchmarks demonstrate that TrajFusion consistently outperforms RFT, particularly on challenging and long-form reasoning problems.

cs.CL↗

arXiv2026

Can Distillation Mitigate Backdoor Attacks in Pre-trained Encoders?

Self-Supervised Learning (SSL) has become a prominent paradigm for pre-training encoders to learning general-purpose representations from unlabeled data and releasing them on third-party platforms for broad downstream deep learning tasks. However, SSL is vulnerable to backdoor attacks, where an adversary may train and distribute poisoned pre-training encoders to contaminate the downstream models. In this paper, we study a defense mechanism based on distillation against poisoned encoders in SSL. Traditionally, distillation transfers knowledge from a pre-trained teacher model to a student model, enabling the student to replicate or refine the teacher's learned representations. We repurpose distillation to extract benign knowledge and remove backdoors from a poisoned pre-trained encoder to produce a clean and reliable pre-trained model. We conduct extensive experiments to evaluate the effectiveness of distillation in mitigating backdoor attacks on pre-trained encoders. Based on two state-of-the-art backdoor attacks and four widely adopted image classification datasets, our results demonstrate that distillation reduces the attack success rate from 80.87% to 27.51%, with only a 6.35% drop in model accuracy. Furthermore, by comparing four teacher architectures, three student models, and six loss functions, we find that the distillation with fine-tuned teacher networks, warm-up-based student training, and attention-based distillation losses yield the best performance.

cs.LG↗

arXiv2026

ECGomics: An Open Platform for AI-ECG Digital Biomarker Discovery

Background: Conventional electrocardiogram (ECG) analysis faces a persistent dichotomy: expert-driven features ensure interpretability but lack sensitivity to latent patterns, while deep learning offers high accuracy but functions as a black box with high data dependency. We introduce ECGomics, a systematic paradigm and open-source platform for the multidimensional deconstruction of cardiac signals into digital biomarker. Methods: Inspired by the taxonomic rigor of genomics, ECGomics deconstructs cardiac activity across four dimensions: Structural, Intensity, Functional, and Comparative. This taxonomy synergizes expert-defined morphological rules with data-driven latent representations, effectively bridging the gap between handcrafted features and deep learning embeddings. Results: We operationalized this framework into a scalable ecosystem consisting of a web-based research platform and a mobile-integrated solution (https://github.com/PKUDigitalHealth/ECGomics). The web platform facilitates high-throughput analysis via precision parameter configuration, high-fidelity data ingestion, and 12-lead visualization, allowing for the systematic extraction of biomarkers across the four ECGomics dimensions. Complementarily, the mobile interface, integrated with portable sensors and a cloud-based engine, enables real-time signal acquisition and near-instantaneous delivery of structured diagnostic reports. This dual-interface architecture successfully transitions ECGomics from theoretical discovery to decentralized, real-world health management, ensuring professional-grade monitoring in diverse clinical and home-based settings. Conclusion: ECGomics harmonizes diagnostic precision, interpretability, and data efficiency. By providing a deployable software ecosystem, this paradigm establishes a robust foundation for digital biomarker discovery and personalized cardiovascular medicine.

q-bio.QM↗

arXiv2026

Does DINOv3 Set a New Medical Vision Standard? Benchmarking 2D and 3D Classification, Segmentation, and Registration

The advent of large-scale vision foundation models, pre-trained on diverse natural images, has marked a paradigm shift in computer vision. However, how the frontier vision foundation models' efficacies transfer to specialised domains such as medical imaging remains an open question. This report investigates whether DINOv3, a state-of-the-art self-supervised vision transformer (ViT) pre-trained on natural images, can directly serve as a powerful, unified encoder for medical vision tasks without domain-specific fine-tuning. To answer this, we benchmark DINOv3 across common medical vision tasks, including 2D and 3D classification, segmentation, and registration on a wide range of medical imaging modalities. We systematically analyse its scalability by varying model sizes and input image resolutions. Our findings reveal that DINOv3 shows impressive performance and establishes a formidable new baseline. Remarkably, it can even outperform medical-specific foundation models like BiomedCLIP and CT-Net on several tasks, despite being trained solely on natural images. However, we identify clear limitations: The model's features degrade in scenarios requiring deep domain specialisation, such as in whole-slide images (WSIs), electron microscopy (EM), and positron emission tomography (PET). Furthermore, we observe that DINOv3 does not consistently follow the scaling law in the medical domain. Its performance does not reliably increase with larger models or finer feature resolutions, showing diverse scaling behaviours across tasks. Overall, our work establishes DINOv3 as a strong baseline, whose powerful visual features can serve as a robust prior for multiple medical tasks. This opens promising future directions, such as leveraging its features to enforce multiview consistency in 3D reconstruction.

cs.CV↗

arXiv2026

V2X-Radar: A Multi-modal Dataset with 4D Radar for Cooperative Perception

Modern autonomous vehicle perception systems often struggle with occlusions and limited perception range. Previous studies have demonstrated the effectiveness of cooperative perception in extending the perception range and overcoming occlusions, thereby enhancing the safety of autonomous driving. In recent years, a series of cooperative perception datasets have emerged; however, these datasets primarily focus on cameras and LiDAR, neglecting 4D Radar, a sensor used in single-vehicle autonomous driving to provide robust perception in adverse weather conditions. In this paper, to bridge the gap created by the absence of 4D Radar datasets in cooperative perception, we present V2X-Radar, the first large-scale, real-world multi-modal dataset featuring 4D Radar. V2X-Radar dataset is collected using a connected vehicle platform and an intelligent roadside unit equipped with 4D Radar, LiDAR, and multi-view cameras. The collected data encompasses sunny and rainy weather conditions, spanning daytime, dusk, and nighttime, as well as various typical challenging scenarios. The dataset consists of 20K LiDAR frames, 40K camera images, and 20K 4D Radar data, including 350K annotated boxes across five categories. To support various research domains, we have established V2X-Radar-C for cooperative perception, V2X-Radar-I for roadside perception, and V2X-Radar-V for single-vehicle perception. Furthermore, we provide comprehensive benchmarks across these three sub-datasets. We will release all datasets and benchmark codebase at https://huggingface.co/datasets/yanglei18/V2X-Radar and https://github.com/yanglei18/V2X-Radar.

cs.CV↗

arXiv2026

AnyECG: Evolved ECG Foundation Model for Holistic Health Profiling

Background: Artificial intelligence enabled electrocardiography (AI-ECG) has demonstrated the ability to detect diverse pathologies, but most existing models focus on single disease identification, neglecting comorbidities and future risk prediction. Although ECGFounder expanded cardiac disease coverage, a holistic health profiling model remains needed. Methods: We constructed a large multicenter dataset comprising 13.3 million ECGs from 2.98 million patients. Using transfer learning, ECGFounder was fine-tuned to develop AnyECG, a foundation model for holistic health profiling. Performance was evaluated using external validation cohorts and a 10-year longitudinal cohort for current diagnosis, future risk prediction, and comorbidity identification. Results: AnyECG demonstrated systemic predictive capability across 1172 conditions, achieving an AUROC greater than 0.7 for 306 diseases. The model revealed novel disease associations, robust comorbidity patterns, and future disease risks. Representative examples included high diagnostic performance for hyperparathyroidism (AUROC 0.941), type 2 diabetes (0.803), Crohn disease (0.817), lymphoid leukemia (0.856), and chronic obstructive pulmonary disease (0.773). Conclusion: The AnyECG foundation model provides substantial evidence that AI-ECG can serve as a systemic tool for concurrent disease detection and long-term risk prediction.

eess.SP↗

arXiv2026

Hausdorff dimension and quasisymmetric minimality of homogeneous Moran sets

In this paper, we study the quasisymmetric Hausdorff minimality of homogeneous Moran sets. First, we obtain the Hausdorff dimension formula of two classes of homogeneous Moran sets which satisfy some conditions. Second, we show two special classes of homogeneous Moran sets with Hausdorff dimension 1 are quasisymmetrically Hausdorff minimal.

math.GN↗

arXiv2026

Experimental realization of quantum Zeno dynamics for robust quantum metrology

Quantum Zeno dynamics (QZD), which restricts the system's evolution to a protected subspace, provides a promising approach for protecting quantum information from noise. Here, we explore a practical approach to harnessing QZD for robust quantum metrology. By introducing strong inter-particle interactions during the parameter encoding stage, we overcome the typical limitations of previous QZD studies, which have largely focused on single-particle systems and faced challenges where QZD could interfere with the encoding process. We experimentally validate the proposed scheme on a nuclear magnetic resonance platform, achieving near-optimal precision scaling under amplitude damping in both parallel and sequential settings. Numerical simulations further demonstrate the scalability of the approach and its compatibility with other control techniques for suppressing more general types of noise. These findings highlight QZD as a powerful strategy for noise-resilient quantum metrology.

quant-ph↗

arXiv2026

Continuous Unitary Designs for Universally Robust Quantum Control

Unitary designs are unitary ensembles that emulate Haar-random unitary statistics. They provide a vital tool for studying quantum randomness and have found broad applications in quantum technologies. However, existing research has focused on discrete ensembles, despite that many physical processes, such as in quantum chaos, thermalization, and control, naturally involve continuous ensembles generated from continuous time-evolution. Here we initial the study of continuous unitary designs, addressing fundamental questions about their construction and practical utility. For single-qubit system, we construct explicit unitary 1-design paths from spherical 2-design curves and Hopf fibration theory. For arbitrary dimensions, we develop two systematic construction frameworks, one based on topological bundle theory of the unitary group and the other based on the Heisenberg-Weyl group. On the practical front, our unitary design paths provide analytical solutions to universally robust quantum control. Simulations show they outperform conventional pulse techniques in mitigating arbitrary unknown static noises, demonstrating immediate utility for quantum engineering. Extending unitary designs to the continuous domain not only introduces powerful geometric and topological tools that complement conventional combinatorial and group-theoretic methods, but also enhances experimental feasibility over discrete counterparts which usually involve instantaneous pulses. As an outlook, we anticipate that this work will pave the way for using continuous unitary designs to explore complex quantum dynamics and devise quantum information protocols.

quant-ph↗

arXiv2026

Parallel Quantum Gates via Scalable Subsystem-Optimized Robust Control

Accurate and efficient implementation of parallel quantum gates is crucial for scalable quantum information processing. However, the unavoidable crosstalk between qubits in current noisy processors impedes the achievement of high gate fidelities and renders full Hilbert-space control optimization prohibitively difficult. Here, we overcome this challenge by reducing the full-system optimization to crosstalk-robust control over constant-sized subsystems, which dramatically reduces the computational cost. Our method effectively eliminates the leading-order gate operation deviations induced by crosstalk, thereby suppressing error rates. Within this framework, we construct analytical pulse solutions for parallel single-qubit gates and numerical pulses for parallel multi-qubit operations. We validate the proposed approach numerically across multiple platforms, including coupled nitrogen-vacancy centers, a nuclear-spin processor, and superconducting-qubit arrays with up to 200 qubits. As a result, the noise scaling is reduced from exponential to linear for parallel single-qubit gates, and an order-of-magnitude reduction is achieved for parallel multi-qubit gates. Moreover, our method does not require precise knowledge of crosstalk strengths and makes no assumption about the underlying qubit connectivity or lattice geometry, thereby establishing a scalable framework for parallel quantum control in large-scale quantum architectures.

quant-ph↗

arXiv2026

A relativistic quantum Euler-Poisson system derived from the Klein-Gordon-Poisson equation: hyperbolic-elliptic structure

In the Klein-Gordon equation, quantum and relativistic parameters are strongly coupled, which poses significant analytical challenges in the derivation and analysis of related classical fluid models. In this paper, starting from the Klein-Gordon-Poisson system, we formally derive a relativistic quantum hydrodynamic (RQHD) system via the Madelung transformation, in which the relativistic and quantum correction terms in the Euler-Poisson framework are clearly exhibited. In particular, at a formal level, the RQHD system reduces to the relativistic hydrodynamics system in the semiclassical regime and to the quantum hydrodynamics system in the non-relativistic regime. These limiting procedures highlight the unified structure of the proposed model and clarify the role played by the coupled relativistic and quantum effects. From an analytical point of view, by reformulating the RQHD system as a coupled hyperbolic-elliptic system with a nonlocal Poisson interaction, we establish the local-in-time existence and uniqueness of classical solutions to the associated Cauchy problem. The initial density is assumed to be a small perturbation of a positive constant state, while the remaining initial data are taken to be general smooth functions. The analysis relies on energy estimates and suitable estimates for the nonlocal terms, and provides a rigorous well-posedness result in the natural energy space.

math.AP↗