arXiv subjects

Jun Li

Publications and source records attributed to Jun Li.

At least 181 records · Page 10Linked to original sources

arXiv2025

WeatherCycle: Unpaired Multi-Weather Restoration via Color Space Decoupled Cycle Learning

Unsupervised image restoration under multi-weather conditions remains a fundamental yet underexplored challenge. While existing methods often rely on task-specific physical priors, their narrow focus limits scalability and generalization to diverse real-world weather scenarios. In this work, we propose \textbf{WeatherCycle}, a unified unpaired framework that reformulates weather restoration as a bidirectional degradation-content translation cycle, guided by degradation-aware curriculum regularization. At its core, WeatherCycle employs a \textit{lumina-chroma decomposition} strategy to decouple degradation from content without modeling complex weather, enabling domain conversion between degraded and clean images. To model diverse and complex degradations, we propose a \textit{Lumina Degradation Guidance Module} (LDGM), which learns luminance degradation priors from a degraded image pool and injects them into clean images via frequency-domain amplitude modulation, enabling controllable and realistic degradation modeling. Additionally, we incorporate a \textit{Difficulty-Aware Contrastive Regularization (DACR)} module that identifies hard samples via a CLIP-based classifier and enforces contrastive alignment between hard samples and restored features to enhance semantic consistency and robustness. Extensive experiments across serve multi-weather datasets, demonstrate that our method achieves state-of-the-art performance among unsupervised approaches, with strong generalization to complex weather degradations.

cs.CV↗

arXiv2025

Diff-Reg v2: Diffusion-Based Matching Matrix Estimation for Image Matching and 3D Registration

Establishing reliable correspondences is crucial for all registration tasks, including 2D image registration, 3D point cloud registration, and 2D-3D image-to-point cloud registration. However, these tasks are often complicated by challenges such as scale inconsistencies, symmetry, and large deformations, which can lead to ambiguous matches. Previous feature-based and correspondence-based methods typically rely on geometric or semantic features to generate or polish initial potential correspondences. Some methods typically leverage specific geometric priors, such as topological preservation, to devise diverse and innovative strategies tailored to a given enhancement goal, which cannot be exhaustively enumerated. Additionally, many previous approaches rely on a single-step prediction head, which can struggle with local minima in complex matching scenarios. To address these challenges, we introduce an innovative paradigm that leverages a diffusion model in matrix space for robust matching matrix estimation. Our model treats correspondence estimation as a denoising diffusion process in the matching matrix space, gradually refining the intermediate matching matrix to the optimal one. Specifically, we apply the diffusion model in the doubly stochastic matrix space for 3D-3D and 2D-3D registration tasks. In the 2D image registration task, we deploy the diffusion model in a matrix subspace where dual-softmax projection regularization is applied. For all three registration tasks, we provide adaptive matching matrix embedding implementations tailored to the specific characteristics of each task while maintaining a consistent "match-to-warp" encoding pattern. Furthermore, we adopt a lightweight design for the denoising module. In inference, once points or image features are extracted and fixed, this module performs multi-step denoising predictions through reverse sampling.

cs.CV↗

arXiv2025

AMLA: MUL by ADD in FlashAttention Rescaling

Multi-head Latent Attention (MLA) significantly reduces KVCache memory usage in Large Language Models while introducing substantial computational overhead and intermediate variable expansion. This poses challenges for efficient hardware implementation -- especially during the decode phase. This paper introduces Ascend MLA (AMLA), a high-performance kernel specifically optimized for Huawei's Ascend NPUs. AMLA is built on two core innovations: (1) A novel FlashAttention-based algorithm that replaces floating-point multiplications with integer additions for output block rescaling, leveraging binary correspondence between FP32 and INT32 representations; (2) A Preload Pipeline strategy with hierarchical tiling that maximizes FLOPS utilization: the Preload Pipeline achieves Cube-bound performance, while hierarchical tiling overlaps data movement and computation within the Cube core. Experiments show that on Ascend 910 NPUs (integrated in CloudMatrix384), AMLA achieves up to 614 TFLOPS, reaching 86.8% of the theoretical maximum FLOPS, outperforming the state-of-the-art open-source FlashMLA implementation, whose FLOPS utilization is up to 66.7% on NVIDIA H800 SXM5. The AMLA kernel has been integrated into Huawei's CANN and will be released soon.

cs.LG↗

arXiv2025

Hierarchical Federated Learning for Social Network with Mobility

Federated Learning (FL) offers a decentralized solution that allows collaborative local model training and global aggregation, thereby protecting data privacy. In conventional FL frameworks, data privacy is typically preserved under the assumption that local data remains absolutely private, whereas the mobility of clients is frequently neglected in explicit modeling. In this paper, we propose a hierarchical federated learning framework based on the social network with mobility namely HFL-SNM that considers both data sharing among clients and their mobility patterns. Under the constraints of limited resources, we formulate a joint optimization problem of resource allocation and client scheduling, which objective is to minimize the energy consumption of clients during the FL process. In social network, we introduce the concepts of Effective Data Coverage Rate and Redundant Data Coverage Rate. We analyze the impact of effective data and redundant data on the model performance through preliminary experiments. We decouple the optimization problem into multiple sub-problems, analyze them based on preliminary experimental results, and propose Dynamic Optimization in Social Network with Mobility (DO-SNM) algorithm. Experimental results demonstrate that our algorithm achieves superior model performance while significantly reducing energy consumption, compared to traditional baseline algorithms.

cs.LG↗

arXiv2025

Self-Alignment Learning to Improve Myocardial Infarction Detection from Single-Lead ECG

Myocardial infarction is a critical manifestation of coronary artery disease, yet detecting it from single-lead electrocardiogram (ECG) remains challenging due to limited spatial information. An intuitive idea is to convert single-lead into multiple-lead ECG for classification by pre-trained models, but generative methods optimized at the signal level in most cases leave a large latent space gap, ultimately degrading diagnostic performance. This naturally raises the question of whether latent space alignment could help. However, most prior ECG alignment methods focus on learning transformation invariance, which mismatches the goal of single-lead detection. To address this issue, we propose SelfMIS, a simple yet effective alignment learning framework to improve myocardial infarction detection from single-lead ECG. Discarding manual data augmentations, SelfMIS employs a self-cutting strategy to pair multiple-lead ECG with their corresponding single-lead segments and directly align them in the latent space. This design shifts the learning objective from pursuing transformation invariance to enriching the single-lead representation, explicitly driving the single-lead ECG encoder to learn a representation capable of inferring global cardiac context from the local signal. Experimentally, SelfMIS achieves superior performance over baseline models across nine myocardial infarction types while maintaining a simpler architecture and lower computational overhead, thereby substantiating the efficacy of direct latent space alignment. Our code and checkpoint will be publicly available after acceptance.

eess.SP↗

arXiv2025

When Color-Space Decoupling Meets Diffusion for Adverse-Weather Image Restoration

Adverse Weather Image Restoration (AWIR) is a highly challenging task due to the unpredictable and dynamic nature of weather-related degradations. Traditional task-specific methods often fail to generalize to unseen or complex degradation types, while recent prompt-learning approaches depend heavily on the degradation estimation capabilities of vision-language models, resulting in inconsistent restorations. In this paper, we propose \textbf{LCDiff}, a novel framework comprising two key components: \textit{Lumina-Chroma Decomposition Network} (LCDN) and \textit{Lumina-Guided Diffusion Model} (LGDM). LCDN processes degraded images in the YCbCr color space, separately handling degradation-related luminance and degradation-invariant chrominance components. This decomposition effectively mitigates weather-induced degradation while preserving color fidelity. To further enhance restoration quality, LGDM leverages degradation-related luminance information as a guiding condition, eliminating the need for explicit degradation prompts. Additionally, LGDM incorporates a \textit{Dynamic Time Step Loss} to optimize the denoising network, ensuring a balanced recovery of both low- and high-frequency features in the image. Finally, we present DriveWeather, a comprehensive all-weather driving dataset designed to enable robust evaluation. Extensive experiments demonstrate that our approach surpasses state-of-the-art methods, setting a new benchmark in AWIR. The dataset and code are available at: https://github.com/fiwy0527/LCDiff.

cs.CV↗

arXiv2025

Towards Communication-efficient Federated Learning via Sparse and Aligned Adaptive Optimization

Adaptive moment estimation (Adam), as a Stochastic Gradient Descent (SGD) variant, has gained widespread popularity in federated learning (FL) due to its fast convergence. However, federated Adam (FedAdam) algorithms suffer from a threefold increase in uplink communication overhead compared to federated SGD (FedSGD) algorithms, which arises from the necessity to transmit both local model updates and first and second moment estimates from distributed devices to the centralized server for aggregation. Driven by this issue, we propose a novel sparse FedAdam algorithm called FedAdam-SSM, wherein distributed devices sparsify the updates of local model parameters and moment estimates and subsequently upload the sparse representations to the centralized server. To further reduce the communication overhead, the updates of local model parameters and moment estimates incorporate a shared sparse mask (SSM) into the sparsification process, eliminating the need for three separate sparse masks. Theoretically, we develop an upper bound on the divergence between the local model trained by FedAdam-SSM and the desired model trained by centralized Adam, which is related to sparsification error and imbalanced data distribution. By minimizing the divergence bound between the model trained by FedAdam-SSM and centralized Adam, we optimize the SSM to mitigate the learning performance degradation caused by sparsification error. Additionally, we provide convergence bounds for FedAdam-SSM in both convex and non-convex objective function settings, and investigate the impact of local epoch, learning rate and sparsification ratio on the convergence rate of FedAdam-SSM. Experimental results show that FedAdam-SSM outperforms baselines in terms of convergence rate (over 1.1$\times$ faster than the sparse FedAdam baselines) and test accuracy (over 14.5\% ahead of the quantized FedAdam baselines).

cs.LG↗

arXiv2025

Towards robust variational quantum simulation of Lindblad dynamics via stochastic Magnus expansion

In this paper, we introduce a novel and general framework for the variational quantum simulation of Lindblad equations. Building on the close relationship between the unraveled Lindblad dynamics, stochastic Magnus integrators, and variational quantum simulation, we propose a high-order scheme for solving the quantum state diffusion equation using exponential integrators. This formulation facilitates the simulation of wavefunction trajectories within the established framework of variational quantum algorithms for time evolution. Our algorithm significantly enhances robustness in two key aspects: the stability of the simulation with large time steps, and the reduction in the number of quantum trajectories required to accurately simulate the Lindblad dynamics in terms of the ensemble average. We demonstrate the effectiveness of our algorithm through numerical examples in both classical and quantum implementations, including the transverse-field Ising model (TFIM) with damping, the Fenna-Matthews-Olson (FMO) complex, and the radical pair model (RPM). The simulation accuracy can be systematically improved, and the algorithm remains reliable even in highly oscillatory regimes. These methods are expected to be applicable to a broader class of open quantum systems beyond the specific models considered in this study.

quant-ph↗

arXiv2025

Plane Detection and Ranking via Model Information Optimization

Plane detection from depth images is a crucial subtask with broad robotic applications, often accomplished by iterative methods such as Random Sample Consensus (RANSAC). While RANSAC is a robust strategy with strong probabilistic guarantees, the ambiguity of its inlier threshold criterion makes it susceptible to false positive plane detections. This issue is particularly prevalent in complex real-world scenes, where the true number of planes is unknown and multiple planes coexist. In this paper, we aim to address this limitation by proposing a generalised framework for plane detection based on model information optimization. Building on previous works, we treat the observed depth readings as discrete random variables, with their probability distributions constrained by the ground truth planes. Various models containing different candidate plane constraints are then generated through repeated random sub-sampling to explain our observations. By incorporating the physics and noise model of the depth sensor, we can calculate the information for each model, and the model with the least information is accepted as the most likely ground truth. This information optimization process serves as an objective mechanism for determining the true number of planes and preventing false positive detections. Additionally, the quality of each detected plane can be ranked by summing the information reduction of inlier points for each plane. We validate these properties through experiments with synthetic data and find that our algorithm estimates plane parameters more accurately compared to the default Open3D RANSAC plane segmentation. Furthermore, we accelerate our algorithm by partitioning the depth map using neural network segmentation, which enhances its ability to generate more realistic plane parameters in real-world data.

cs.CV↗

arXiv2025

When MoE Meets Blockchain: A Trustworthy Distributed Framework of Large Models

As an enabling architecture of Large Models (LMs), Mixture of Experts (MoE) has become prevalent thanks to its sparsely-gated mechanism, which lowers computational overhead while maintaining learning performance comparable to dense LMs. The essence of MoE lies in utilizing a group of neural networks (called experts) with each specializing in different types of tasks, along with a trainable gating network that selectively activates a subset of these experts to handle specific tasks. Traditional cloud-based MoE encounters challenges such as prolonged response latency, high bandwidth consumption, and data privacy leakage. To address these issues, researchers have proposed to deploy MoE over distributed edge networks. However, a key concern of distributed MoE frameworks is the lack of trust in data interactions among distributed experts without the surveillance of any trusted authority, and thereby prone to potential attacks such as data manipulation. In response to the security issues of traditional distributed MoE, we propose a blockchain-aided trustworthy MoE (B-MoE) framework that consists of three layers: the edge layer, the blockchain layer, and the storage layer. In this framework, the edge layer employs the activated experts downloaded from the storage layer to process the learning tasks, while the blockchain layer functions as a decentralized trustworthy network to trace, verify, and record the computational results of the experts from the edge layer. The experimental results demonstrate that B-MoE is more robust to data manipulation attacks than traditional distributed MoE during both the training and inference processes.

cs.DC↗

arXiv2025

Analysis and Optimization of Wireless Multimodal Federated Learning on Modal Heterogeneity

Multimodal federated learning (MFL) is a distributed framework for training multimodal models without uploading local multimodal data of clients, thereby effectively protecting client privacy. However, multimodal data is commonly heterogeneous across diverse clients, where each client possesses only a subset of all modalities, renders conventional analysis results and optimization methods in unimodal federated learning inapplicable. In addition, fixed latency demand and limited communication bandwidth pose significant challenges for deploying MFL in wireless scenarios. To optimize the wireless MFL performance on modal heterogeneity, this paper proposes a joint client scheduling and bandwidth allocation (JCSBA) algorithm based on a decision-level fusion architecture with adding a unimodal loss function. Specifically, with the decision results, the unimodal loss functions are added to both the training objective and local update loss functions to accelerate multimodal convergence and improve unimodal performance. To characterize MFL performance, we derive a closed-form upper bound related to client and modality scheduling and minimize the derived bound under the latency, energy, and bandwidth constraints through JCSBA. Experimental results on multimodal datasets demonstrate that the JCSBA algorithm improves the multimodal accuracy and the unimodal accuracy by 4.06% and 2.73%, respectively, compared to conventional algorithms.

cs.DC↗

arXiv2025

IRDFusion: Iterative Relation-Map Difference guided Feature Fusion for Multispectral Object Detection

Current multispectral object detection methods often retain extraneous background or noise during feature fusion, limiting perceptual performance. To address this, we propose an innovative feature fusion framework based on cross-modal feature contrastive and screening strategy, diverging from conventional approaches. The proposed method adaptively enhances salient structures by fusing object-aware complementary cross-modal features while suppressing shared background interference. Our solution centers on two novel, specially designed modules: the Mutual Feature Refinement Module (MFRM) and the Differential Feature Feedback Module (DFFM). The MFRM enhances intra- and inter-modal feature representations by modeling their relationships, thereby improving cross-modal alignment and discriminative power. Inspired by feedback differential amplifiers, the DFFM dynamically computes inter-modal differential features as guidance signals and feeds them back to the MFRM, enabling adaptive fusion of complementary information while suppressing common-mode noise across modalities. To enable robust feature learning, the MFRM and DFFM are integrated into a unified framework, which is formally formulated as an Iterative Relation-Map Differential Guided Feature Fusion mechanism, termed IRDFusion. IRDFusion enables high-quality cross-modal fusion by progressively amplifying salient relational signals through iterative feedback, while suppressing feature noise, leading to significant performance gains. In extensive experiments on FLIR, LLVIP and M$^3$FD datasets, IRDFusion achieves state-of-the-art performance and consistently outperforms existing methods across diverse challenging scenarios, demonstrating its robustness and effectiveness. Code will be available at https://github.com/61s61min/IRDFusion.git.

cs.CV↗

arXiv2025

Observational Evidence of Dust Evolution in Supernova Remnants: Size Redistribution Toward Larger Grains in Early Sedov Phase

Recent observations have revealed that dust is widespread and abundant in galaxies up to $z\,{\backsimeq}\,8$, significantly influencing their appearance and spectral properties. In the early Universe, dust is thought to form primarily in supernova (SN) ejecta, but also undergoes destruction by the reverse shock. Studying dust in local supernova remnants (SNRs) of different sizes and ages thus provides key constraints on dust formation and evolution during cosmic dawn. Using the newly released 3D $R_{\rm V}$ map, we derived local $R_{\rm V} \equiv A_{\rm V}/E(B-V)$ values for dust in 14 Galactic SNRs in the early Sedov phase and their surrounding interstellar medium (ISM). For the first time, we detect a moderately strong positive correlation between the difference in SNR and ISM $R_{\rm V}$ ($ΔR_{\rm V}$) and SNR radius ($R_{\rm SNR}$), with a Spearman coefficient of $r_{\rm s}\,{=}\,0.62\,{\pm}\,0.14$. This trend offers direct observational evidence for a redistribution of dust grain sizes toward larger grains during SN shock processing, consistent with theoretical models. Our findings provide essential observational constraints on dust size evolution in SNRs and important implications for understanding the rapid enrichment and survival of dust in the early Universe.

astro-ph.GA↗

arXiv2025

Genus Two Stable Maps, Local Equations and Modular Resolutions

We provide a geometric construction of a sequence of modular blowups of the Artin stack parameterizing pre-stable pairs consisting of a genus-two nodal curve and a smooth divisor. The resulting stack locally diagonalizes the tautological derived objects associated with the moduli of stable maps from genus-two curves to projective space. As a consequence, the singularities of the main component of the moduli space of stable maps are resolved, and the entire space admits only normal crossing singularities. Our approach is expected to generalize to higher genera.

math.AG↗

arXiv2025

A biologically inspired separable learning vision model for real-time traffic object perception in Dark

Fast and accurate object perception in low-light traffic scenes has attracted increasing attention. However, due to severe illumination degradation and the lack of reliable visual cues, existing perception models and methods struggle to quickly adapt to and accurately predict in low-light environments. Moreover, there is the absence of available large-scale benchmark specifically focused on low-light traffic scenes. To bridge this gap, we introduce a physically grounded illumination degradation method tailored to real-world low-light settings and construct Dark-traffic, the largest densely annotated dataset to date for low-light traffic scenes, supporting object detection, instance segmentation, and optical flow estimation. We further propose the Separable Learning Vision Model (SLVM), a biologically inspired framework designed to enhance perception under adverse lighting. SLVM integrates four key components: a light-adaptive pupillary mechanism for illumination-sensitive feature extraction, a feature-level separable learning strategy for efficient representation, task-specific decoupled branches for multi-task separable learning, and a spatial misalignment-aware fusion module for precise multi-feature alignment. Extensive experiments demonstrate that SLVM achieves state-of-the-art performance with reduced computational overhead. Notably, it outperforms RT-DETR by 11.2 percentage points in detection, YOLOv12 by 6.1 percentage points in instance segmentation, and reduces endpoint error (EPE) of baseline by 12.37% on Dark-traffic. On the LIS benchmark, the end-to-end trained SLVM surpasses Swin Transformer+EnlightenGAN and ConvNeXt-T+EnlightenGAN by an average of 11 percentage points across key metrics, and exceeds Mask RCNN (with light enhancement) by 3.1 percentage points. The Dark-traffic dataset and complete code is released at https://github.com/alanli1997/slvm.

cs.CV↗

arXiv2025

FlashRecovery: Fast and Low-Cost Recovery from Failures for Large-Scale Training of LLMs

Large language models (LLMs) have made a profound impact across various fields due to their advanced capabilities. However, training these models at unprecedented scales requires extensive AI accelerator clusters and sophisticated parallelism strategies, which pose significant challenges in maintaining system reliability over prolonged training periods. A major concern is the substantial loss of training time caused by inevitable hardware and software failures. To address these challenges, we present FlashRecovery, a fast and low-cost failure recovery system comprising three core modules: (1) Active and real-time failure detection. This module performs continuous training state monitoring, enabling immediate identification of hardware and software failures within seconds, thus ensuring rapid incident response; (2) Scale-independent task restart. By employing different recovery strategies for normal and faulty nodes, combined with an optimized communication group reconstruction protocol, our approach ensures that the recovery time remains nearly constant, regardless of cluster scale; (3) Checkpoint-free recovery within one step. Our novel recovery mechanism enables single-step restoration, completely eliminating dependence on traditional checkpointing methods and their associated overhead. Collectively, these innovations enable FlashRecovery to achieve optimal Recovery Time Objective (RTO) and Recovery Point Objective (RPO), substantially improving the reliability and efficiency of long-duration LLM training. Experimental results demonstrate that FlashRecovery system can achieve training restoration on training cluster with 4, 800 devices in 150 seconds. We also verify that the time required for failure recovery is nearly consistent for different scales of training tasks.

cs.DC↗

arXiv2025

Category-Aware 3D Object Composition with Disentangled Texture and Shape Multi-view Diffusion

In this paper, we tackle a new task of 3D object synthesis, where a 3D model is composited with another object category to create a novel 3D model. However, most existing text/image/3D-to-3D methods struggle to effectively integrate multiple content sources, often resulting in inconsistent textures and inaccurate shapes. To overcome these challenges, we propose a straightforward yet powerful approach, category+3D-to-3D (C33D), for generating novel and structurally coherent 3D models. Our method begins by rendering multi-view images and normal maps from the input 3D model, then generating a novel 2D object using adaptive text-image harmony (ATIH) with the front-view image and a text description from another object category as inputs. To ensure texture consistency, we introduce texture multi-view diffusion, which refines the textures of the remaining multi-view RGB images based on the novel 2D object. For enhanced shape accuracy, we propose shape multi-view diffusion to improve the 2D shapes of both the multi-view RGB images and the normal maps, also conditioned on the novel 2D object. Finally, these outputs are used to reconstruct a complete and novel 3D model. Extensive experiments demonstrate the effectiveness of our method, yielding impressive 3D creations, such as shark(3D)-crocodile(text) in the first row of Fig. 1. A project page is available at: https://xzr52.github.io/C33D/

cs.CV↗

arXiv2025

Adaptive CUSUM Chart for Simultaneous Monitoring of Mean and Variance

Simultaneously monitoring changes in both the mean and variance is a fundamental problem in Statistical Process Control, and numerous methods have been developed to address it. However, many existing approaches face notable limitations: some rely on tuning parameters that can significantly affect performance; others are biased toward detecting increases in variance while performing poorly for decreases; and some are computationally burdensome. To address these limitations, we propose a novel adaptive CUSUM chart for jointly monitoring the mean and variance of a Gaussian process. The proposed method is free of tuning parameters, efficient in detecting a broad range of shifts in both mean and variance, and well-suited for real-time monitoring due to its recursive structure. It also has a built-in post-signal diagnostics function that can identify what kind of distributional changes have occurred after an alarm. Simulation results show that, compared to existing methods, the proposed chart achieves the most favorable balance between detection power and computational efficiency, delivering the best overall performance.

stat.ME↗