Searcharxiv⌕ Search

arXiv subjects

Yan Xu

Publications and source records attributed to Yan Xu.

At least 73 records · Page 4Linked to original sources
arXiv2025

Novel View Synthesis from A Few Glimpses via Test-Time Natural Video Completion

Given just a few glimpses of a scene, can you imagine the movie playing out as the camera glides through it? That's the lens we take on \emph{sparse-input novel view synthesis}, not only as filling spatial gaps between widely spaced views, but also as \emph{completing a natural video} unfolding through space. We recast the task as \emph{test-time natural video completion}, using powerful priors from \emph{pretrained video diffusion models} to hallucinate plausible in-between views. Our \emph{zero-shot, generation-guided} framework produces pseudo views at novel camera poses, modulated by an \emph{uncertainty-aware mechanism} for spatial coherence. These synthesized frames densify supervision for \emph{3D Gaussian Splatting} (3D-GS) for scene reconstruction, especially in under-observed regions. An iterative feedback loop lets 3D geometry and 2D view synthesis inform each other, improving both the scene reconstruction and the generated views. The result is coherent, high-fidelity renderings from sparse inputs \emph{without any scene-specific training or fine-tuning}. On LLFF, DTU, DL3DV, and MipNeRF-360, our method significantly outperforms strong 3D-GS baselines under extreme sparsity.

cs.CV↗
arXiv2025

Cyber-Resilient Fault Diagnosis Methodology in Inverter-Based Resource-Dominated Microgrids with Single-Point Measurement

Cyber-attacks jeopardize the safe operation of inverter-based resource-dominated microgrids (IBR-dominated microgrids). At the same time, existing diagnostic methods either depend on expensive multi-point instrumentation or stringent modeling assumptions that are untenable under single-point measurement constraints. This paper proposes a Fractional-Order Memory-Enhanced Attack-Diagnosis Scheme (FO-MADS) that achieves timely fault localization and cyber-resilient fault diagnosis using only one VPQ (voltage, active power, reactive power) measurement point. FO-MADS first constructs a dual fractional-order feature library by jointly applying Caputo and Grünwald-Letnikov derivatives, thereby amplifying micro-perturbations and slow drifts in the VPQ signal. A two-stage hierarchical classifier then pinpoints the affected inverter and isolates the faulty IGBT switch, effectively alleviating class imbalance. Robustness is further strengthened through Progressive Memory-Replay Adversarial Training (PMR-AT), whose attack-aware loss is dynamically re-weighted via Online Hard Example Mining (OHEM) to prioritize the most challenging samples. Experiments on a four-inverter IBR-dominated microgrid testbed comprising 1 normal and 24 fault classes under four attack scenarios demonstrate diagnostic accuracies of 96.6% (bias), 94.0% (noise), 92.8% (data replacement), and 95.7% (replay), while sustaining 96.7% under attack-free conditions. These results establish FO-MADS as a cost-effective and readily deployable solution that markedly enhances the cyber-physical resilience of IBR-dominated microgrids.

eess.SY↗
arXiv2025

SkinCaRe: A Multimodal Dermatology Dataset Annotated with Medical Caption and Chain-of-Thought Reasoning

With the widespread application of artificial intelligence (AI), particularly deep learning (DL) and vision large language models (VLLMs), in skin disease diagnosis, the need for interpretability becomes crucial. However, existing dermatology datasets are limited in their inclusion of concept-level meta-labels, and none offer rich medical descriptions in natural language. This deficiency impedes the advancement of LLM-based methods in dermatologic diagnosis. To address this gap and provide a meticulously annotated dermatology dataset with comprehensive natural language descriptions, we introduce \textbf{SkinCaRe}, a comprehensive multimodal resource that unifies \textit{SkinCAP} and \textit{SkinCoT}. \textbf{SkinCAP} comprises 4,000 images sourced from the Fitzpatrick 17k skin disease dataset and the Diverse Dermatology Images dataset, annotated by board-certified dermatologists to provide extensive medical descriptions and captions. In addition, we introduce \textbf{SkinCoT}, a curated dataset pairing 3,041 dermatologic images with clinician-verified, hierarchical chain-of-thought (CoT) diagnoses. Each diagnostic narrative is rigorously evaluated against six quality criteria and iteratively refined until it meets a predefined standard of clinical accuracy and explanatory depth. Together, SkinCAP (captioning) and SkinCoT (reasoning), collectively referred to as SkinCaRe, encompass 7,041 expertly curated dermatologic cases and provide a unified and trustworthy resource for training multimodal models that both describe and explain dermatologic images. SkinCaRe is publicly available at https://huggingface.co/datasets/yuhos16/SkinCaRe.

cs.CV↗
arXiv2025

Benchmarking the Thinking Mode of Multimodal Large Language Models in Clinical Tasks

A recent advancement in Multimodal Large Language Models (MLLMs) research is the emergence of "reasoning MLLMs" that offer explicit control over their internal thinking processes (normally referred as the "thinking mode") alongside the standard "non-thinking mode". This capability allows these models to engage in a step-by-step process of internal deliberation before generating a final response. With the rapid transition to and adoption of these "dual-state" MLLMs, this work rigorously evaluated how the enhanced reasoning processes of these MLLMs impact model performance and reliability in clinical tasks. This paper evaluates the active "thinking mode" capabilities of two leading MLLMs, Seed1.5-VL and Gemini-2.5-Flash, for medical applications. We assessed their performance on four visual medical tasks using VQA-RAD and ROCOv2 datasets. Our findings reveal that the improvement from activating the thinking mode remains marginal compared to the standard non-thinking mode for the majority of the tasks. Their performance on complex medical tasks such as open-ended VQA and medical image interpretation remains suboptimal, highlighting the need for domain-specific medical data and more advanced methods for medical knowledge integration.

cs.CL↗
arXiv2025

A New Neural Network Paradigm for Scalable and Generalizable Stability Analysis of Power Systems

This paper presents a new neural network (NN) paradigm for scalable and generalizable stability analysis of power systems. The paradigm consists of two parts: the neural stability descriptor and the sample-augmented iterative training scheme. The first part, based on system decomposition, constructs the object (such as a stability function or condition) for stability analysis as a scalable aggregation of multiple NNs. These NNs remain fixed across varying power system structures and parameters, and are repeatedly shared within each system instance defined by these variations, thereby enabling the generalization of the neural stability descriptor across a class of power systems. The second part learns the neural stability descriptor by iteratively training the NNs with sample augmentation, guided by the tailored conservativeness-aware loss function. The training set is strategically constructed to promote the descriptor's generalizability, which is systematically evaluated by verification and validation during the training process. Specifically, the proposed NN paradigm is implemented for large-disturbance stability analysis of the bulk power grid and small-disturbance stability conditions of the microgrid system. Finally, numerical studies for the two implementations demonstrate the applicability and effectiveness of the proposed NN paradigm.

eess.SY↗
arXiv2025

Supra-arcade downflows in an extensive fan associated with a giant quiescent solar filament eruption

We investigate the aftermath of a giant quiescent solar filament eruption on December 24, 2023. One feature of the eruption is an extensive fan above the filament channel that is about three times as wide as similar structures that appear above active regions (ARs) during solar flares. The fan contains numerous supra-arcade downflows (SADs), and we investigate the largest SADs with continuous Hinode X-ray Telescope (XRT) observations. The measured maximum width of the SADs in this event is at least three times the maximum width of SADs observed in AR flares, whereas the velocities of the largest SADs are similar to the typical values of AR SADs. The kinetic characteristics of the largest SADs observed in this event align with previous model predictions, where SADs originate from the non-linear development of Rayleigh-Taylor type instabilities. In this scenario, the larger system size allows the existence of larger-scale instabilities, while the development of the velocities of these instabilities is expected to be independent of the system size. Compared to AR flares, the temperature and emission measure in this event are lower, and there is less overall radiation, resulting in no evident Geostationary Operational Environmental Satellite (GOES) signature. Similar to those in AR flares, SADs show lower temperatures compared to the surrounding fan plasma. Our observations show that SADs are present in a wide variety of eruptions. The reconnection mechanisms present in quiescent filament eruptions are similar to those driving more compact eruptions originating from ARs.

astro-ph.SR↗
arXiv2025

ARCH: Hierarchical Hybrid Learning for Long-Horizon Contact-Rich Robotic Assembly

Generalizable long-horizon robotic assembly requires reasoning at multiple levels of abstraction. While end-to-end imitation learning (IL) is a promising approach, it typically requires large amounts of expert demonstration data and often struggles to achieve the high precision demanded by assembly tasks. Reinforcement learning (RL) approaches, on the other hand, have shown some success in high-precision assembly, but suffer from sample inefficiency, which limits their effectiveness in long-horizon tasks. To address these challenges, we propose a hierarchical modular approach, named Adaptive Robotic Compositional Hierarchy (ARCH), which enables long-horizon, high-precision robotic assembly in contact-rich settings. ARCH employs a hierarchical planning framework, including a low-level primitive library of parameterized skills and a high-level policy. The low-level primitive library includes essential skills for assembly tasks, such as grasping and inserting. These primitives consist of both RL and model-based policies. The high-level policy, learned via IL from a handful of demonstrations, without the need for teleoperation, selects the appropriate primitive skills and instantiates them with input parameters. We extensively evaluate our approach in simulation and on a real robotic manipulation platform. We show that ARCH generalizes well to unseen objects and outperforms baseline methods in terms of success rate and data efficiency. More details are available at: https://long-horizon-assembly.github.io.

cs.RO↗
arXiv2025

Symbolic Equation Modeling of Composite Loads: A Kolmogorov-Arnold Network based Learning Approach

With increasing penetration of distributed energy resources installed behind the meter, there is a growing need for adequate modelling of composite loads to enable accurate power system simulation analysis. Existing measurement based load modeling methods either fit fixed-structure physical models, which limits adaptability to evolving load mixes, or employ flexible machine learning methods which are however black boxes and offer limited interpretability. This paper presents a new learning based load modelling method based on Kolmogorov Arnold Networks towards modelling flexibility and interpretability. By actively learning activation functions on edges, KANs automatically derive free form symbolic equations that capture nonlinear relationships among measured variables without prior assumptions about load structure. Case studies demonstrate that the proposed approach outperforms other methods in both accuracy and generalization ability, while uniquely representing composite loads into transparent, interpretable mathematical equations.

eess.SY↗
arXiv2025

DySTop

Federated Learning (FL) has emerged as a potential distributed learning paradigm that enables model training on edge devices (i.e., workers) while preserving data privacy. However, its reliance on a centralized server leads to limited scalability. Decentralized federated learning (DFL) eliminates the dependency on a centralized server by enabling peer-to-peer model exchange. Existing DFL mechanisms mainly employ synchronous communication, which may result in training inefficiencies under heterogeneous and dynamic edge environments. Although a few recent asynchronous DFL (ADFL) mechanisms have been proposed to address these issues, they typically yield stale model aggregation and frequent model transmission, leading to degraded training performance on non-IID data and high communication overhead. To overcome these issues, we present DySTop, an innovative mechanism that jointly optimizes dynamic staleness control and topology construction in ADFL. In each round, multiple workers are activated, and a subset of their neighbors is selected to transmit models for aggregation, followed by local training. We provide a rigorous convergence analysis for DySTop, theoretically revealing the quantitative relationships between the convergence bound and key factors such as maximum staleness, activating frequency, and data distribution among workers. From the insights of the analysis, we propose a worker activation algorithm (WAA) for staleness control and a phase-aware topology construction algorithm (PTCA) to reduce communication overhead and handle data non-IID. Extensive evaluations through both large-scale simulations and real-world testbed experiments demonstrate that our DySTop reduces completion time by 51.8% and the communication resource consumption by 57.1% compared to state-of-the-art solutions, while maintaining the same model accuracy.

cs.DC↗
arXiv2025

All-in-One Medical Image Restoration with Latent Diffusion-Enhanced Vector-Quantized Codebook Prior

All-in-one medical image restoration (MedIR) aims to address multiple MedIR tasks using a unified model, concurrently recovering various high-quality (HQ) medical images (e.g., MRI, CT, and PET) from low-quality (LQ) counterparts. However, all-in-one MedIR presents significant challenges due to the heterogeneity across different tasks. Each task involves distinct degradations, leading to diverse information losses in LQ images. Existing methods struggle to handle these diverse information losses associated with different tasks. To address these challenges, we propose a latent diffusion-enhanced vector-quantized codebook prior and develop \textbf{DiffCode}, a novel framework leveraging this prior for all-in-one MedIR. Specifically, to compensate for diverse information losses associated with different tasks, DiffCode constructs a task-adaptive codebook bank to integrate task-specific HQ prior features across tasks, capturing a comprehensive prior. Furthermore, to enhance prior retrieval from the codebook bank, DiffCode introduces a latent diffusion strategy that utilizes the diffusion model's powerful mapping capabilities to iteratively refine the latent feature distribution, estimating more accurate HQ prior features during restoration. With the help of the task-adaptive codebook bank and latent diffusion strategy, DiffCode achieves superior performance in both quantitative metrics and visual quality across three MedIR tasks: MRI super-resolution, CT denoising, and PET synthesis.

cs.CV↗
arXiv2025

Batched Adaptive Network Formation

Networks are central to many economic and organizational applications, including workplace team formation, social platform recommendations, and classroom friendship development. In these settings, networks are modeled as graphs, with agents as nodes, agent pairs as edges, and edge weights capturing pairwise production or interaction outcomes. This paper develops an adaptive, or \textit{online}, policy that learns to form increasingly effective networks as data accumulates over time, progressively improving total network output measured by the sum of edge weights. Our approach builds on the weighted stochastic block model (WSBM), which captures agents' unobservable heterogeneity through discrete latent types and models their complementarities in a flexible, nonparametric manner. We frame the online network formation problem as a non-standard \textit{batched multi-armed bandit}, where each type pair corresponds to an arm, and pairwise reward depends on type complementarity. This strikes a balance between exploration -- learning latent types and complementarities -- and exploitation -- forming high-weighted networks. We establish two key results: a \textit{batched local asymptotic normality} result for the WSBM and an asymptotic equivalence between maximum likelihood and variational estimates of the intractable likelihood. Together, they provide a theoretical foundation for treating variational estimates as normal signals, enabling principled Bayesian updating across batches. The resulting posteriors are then incorporated into a tailored maximum-weight matching problem to determine the policy for the next batch. Simulations show that our algorithm substantially improves outcomes within a few batches, yields increasingly accurate parameter estimates, and remains effective even in nonstationary settings with evolving agent pools.

econ.EM↗
arXiv2025

Sign-patterns of Certain Infinite Products

The signs of Fourier coefficients of certain eta quotients are determined by dissecting expansions for theta functions and by applying a general dissection formula for certain classes of quintuple products. A characterization is given for the coefficient sign patterns for \[ \frac{(q^i;q^i)_{\infty}}{(q^p;q^p)_{\infty}} \] for integers \( i > 1 \) and primes \( p > 3 \). The sign analysis for this quotient addresses and extends a conjecture of Bringmann et al. for the coefficients of \( (q^2;q^2)_{\infty}(q^5;q^5)_{\infty}^{-1} \). The sign distribution for additional classes of eta quotients is considered. This addresses multiple conjectures posed by Bringmann et al.

math.NT↗
arXiv2025

The angular momentum of stars reflects the relationship between star-forming environment and galactic evolution history

This study focuses on stars with masses above the Kraft break in the \textit{Kepler} field. Their rotational angular momenta are essentially the same as those at the zero-age main sequence. The angular momentum dissipation experienced by these stars during their pre-main sequence phase is also relatively weak, so their rotational angular momentum can reflect the parameters of their parental molecular clouds. The reliability of angular momentum measurements was evaluated based on the phenomenon of angular momentum conservation observed in stars before and after the turn-off point in observational data. We find that stars with masses between 1.4$M_\odot$ and 1.8$M_\odot$ show an inverse proportionality between angular momentum and isochrone age. We propose that the angular momentum-age correlation reflects changes in the star-forming environment in the Milky Way's history. Besides, the observed inverse proportionality implies that, as the Milky Way has evolved, the stars formed within it tend to possess greater rotational angular momenta. This trend would promote the fragmentation of stars during the pre-main sequence phase and inhibit the formation of massive stars, providing a useful perspective for explaining variations in the initial mass function.

astro-ph.GA↗
arXiv2025

Well-balanced path-conservative discontinuous Galerkin methods with equilibrium preserving space for shallow water linearized moment equations

This paper presents high-order, well-balanced, path-conservative discontinuous Galerkin (DG) methods for the shallow water linearized moment equations (SWLME), designed to preserve both still and moving water equilibrium states. Unlike the multi-layer shallow water equations, which model vertical velocity variations using multiple distinct layers, the SWLME employs a polynomial expansion of velocity profiles with up to $N$ moments. This approach enables a more detailed representation of vertical momentum transfer and complex velocity profiles while retaining hyperbolicity. However, the presence of non-conservative terms and complex steady-state structures introduces significant numerical challenges. Addressing these challenges, we develop path-conservative DG schemes grounded in the Dal Maso-LeFloch-Murat (DLM) theory for non-conservative products. Our method balances flux gradients, non-conservative terms, and source terms through equilibrium-preserving spaces. For the still water equilibrium, we reformulate the equations into a quasilinear form that eliminates source terms, inherently preserving steady states. For the moving water equilibrium, we extend the DG method by transforming conservative variables into equilibrium variables and employing linear segment paths. Theoretical analysis and numerical experiments demonstrate that the proposed methods achieve exact equilibrium preservation while maintaining high-order accuracy, even in scenarios with vertical velocity variations and complex topographies.

math.NA↗
arXiv2025

Extriangulated factorization systems, $s$-torsion pairs and recollements

We introduce extriangulated factorization systems in extriangulated categories and show that there exists a bijection between $s$-torsion pairs and extriangulated factorization systems. We also consider the gluing of $s$-torsion pairs and extriangulated factorization systems under recollements of extriangulated categories.

math.CT↗
arXiv2025

Visual Textualization for Image Prompted Object Detection

We propose VisTex-OVLM, a novel image prompted object detection method that introduces visual textualization -- a process that projects a few visual exemplars into the text feature space to enhance Object-level Vision-Language Models' (OVLMs) capability in detecting rare categories that are difficult to describe textually and nearly absent from their pre-training data, while preserving their pre-trained object-text alignment. Specifically, VisTex-OVLM leverages multi-scale textualizing blocks and a multi-stage fusion strategy to integrate visual information from visual exemplars, generating textualized visual tokens that effectively guide OVLMs alongside text prompts. Unlike previous methods, our method maintains the original architecture of OVLM, maintaining its generalization capabilities while enhancing performance in few-shot settings. VisTex-OVLM demonstrates superior performance across open-set datasets which have minimal overlap with OVLM's pre-training data and achieves state-of-the-art results on few-shot benchmarks PASCAL VOC and MSCOCO. The code will be released at https://github.com/WitGotFlg/VisTex-OVLM.

cs.CV↗
arXiv2025

FedBKD: Distilled Federated Learning to Embrace Gerneralization and Personalization on Non-IID Data

Federated learning (FL) is a decentralized collaborative machine learning (ML) technique. It provides a solution to the issues of isolated data islands and data privacy leakage in industrial ML practices. One major challenge in FL is handling the non-identical and independent distributed (non-IID) data. Current solutions either focus on constructing an all-powerful global model, or customizing personalized local models. Few of them can provide both a well-generalized global model and well-performed local models at the same time. Additionally, many FL solutions to the non-IID problem are benefited from introducing public datasets. However, this will also increase the risk of data leakage. To tackle the problems, we propose a novel data-free distillation framework, Federated Bidirectional Knowledge Distillation (FedBKD). Specifically, we train Generative Adversarial Networks (GAN) for synthetic data. During the GAN training, local models serve as discriminators and their parameters are frozen. The synthetic data is then used for bidirectional distillation between global and local models to achieve knowledge interactions so that performances for both sides are improved. We conduct extensive experiments on 4 benchmarks under different non-IID settings. The results show that FedBKD achieves SOTA performances in every case.

cs.LG↗