arXiv subjects

Xiaosong Li

Publications and source records attributed to Xiaosong Li.

At least 37 records · Page 2Linked to original sources

arXiv2026

All-weather Multi-Modality Image Fusion: Unified Framework and 100k Benchmark

Multi-modality image fusion (MMIF) combines complementary information from different image modalities to provide a comprehensive and objective interpretation of scenes. However, existing fusion methods cannot resist different weather interferences in real-world scenes, limiting their practical applicability. To bridge this gap, we propose an end-to-end, unified all-weather MMIF model. Rather than focusing solely on pixel-level recovery, our method emphasizes maximizing the representation of key scene information through joint feature fusion and restoration. Specifically, we first decompose images into low-rank and sparse components, enabling effective feature separation for enhanced multi-modality perception. During feature recovery, we introduce a physically-aware clear feature prediction module, inferring variations in light transmission via illumination and reflectance. Clear features generated by the network are used to enhance salient information representation. We also construct a large-scale MMIF dataset with 100,000 image pairs comprehensively across rain, haze, and snow conditions, as well as covering various degradation levels and diverse scenes. Experimental results in both real-world and synthetic scenes demonstrate that the proposed method excels in image fusion and downstream tasks such as object detection, semantic segmentation, and depth estimation. The code is available at https://github.com/ixilai/AWFusion.

cs.CV↗

arXiv2026

CAWM-Mamba: A unified model for infrared-visible image fusion and compound adverse weather restoration

Multimodal Image Fusion (MMIF) integrates complementary information from various modalities to produce clearer and more informative fused images. MMIF under adverse weather is particularly crucial in autonomous driving and UAV monitoring applications. However, existing adverse weather fusion methods generally only tackle single types of degradation such as haze, rain, or snow, and fail when multiple degradations coexist (e.g., haze+rain, rain+snow). To address this challenge, we propose Compound Adverse Weather Mamba (CAWM-Mamba), the first end-to-end framework that jointly performs image fusion and compound weather restoration with unified shared weights. Our network contains three key components: (1) a Weather-Aware Preprocess Module (WAPM) to enhance degraded visible features and extracts global weather embeddings; (2) a Cross-modal Feature Interaction Module (CFIM) to facilitate the alignment of heterogeneous modalities and exchange of complementary features across modalities; and (3) a Wavelet Space State Block (WSSB) that leverages wavelet-domain decomposition to decouple multi-frequency degradations. WSSB includes Freq-SSM, a module that models anisotropic high-frequency degradation without redundancy, and a unified degradation representation mechanism to further improve generalization across complex compound weather conditions. Extensive experiments on the AWMM-100K benchmark and three standard fusion datasets demonstrate that CAWM-Mamba consistently outperforms state-of-the-art methods in both compound and single-weather scenarios. In addition, our fusion results excel in downstream tasks covering semantic segmentation and object detection, confirming the practical value in real-world adverse weather perception. The source code will be available at https://github.com/Feecuin/CAWM-Mamba.

cs.CV↗

arXiv2026

Towards Efficient Agents: A Co-Design of Inference Architecture and System

The rapid development of large language model (LLM)-based agents has unlocked new possibilities for autonomous multi-turn reasoning and tool-augmented decision-making. However, their real-world deployment is hindered by severe inefficiencies that arise not from isolated model inference, but from the systemic latency accumulated across reasoning loops, context growth, and heterogeneous tool interactions. This paper presents AgentInfer, a unified framework for end-to-end agent acceleration that bridges inference optimization and architectural design. We decompose the problem into four synergistic components: AgentCollab, a hierarchical dual-model reasoning framework that balances large- and small-model usage through dynamic role assignment; AgentSched, a cache-aware hybrid scheduler that minimizes latency under heterogeneous request patterns; AgentSAM, a suffix-automaton-based speculative decoding method that reuses multi-session semantic memory to achieve low-overhead inference acceleration; and AgentCompress, a semantic compression mechanism that asynchronously distills and reorganizes agent memory without disrupting ongoing reasoning. Together, these modules form a Self-Evolution Engine capable of sustaining efficiency and cognitive stability throughout long-horizon reasoning tasks. Experiments on the BrowseComp-zh and DeepDiver benchmarks demonstrate that through the synergistic collaboration of these methods, AgentInfer reduces ineffective token consumption by over 50%, achieving an overall 1.8-2.5 times speedup with preserved accuracy. These results underscore that optimizing for agentic task completion-rather than merely per-token throughput-is the key to building scalable, efficient, and self-improving intelligent systems.

cs.CL↗

arXiv2026

TSJNet: A Multi-modality Target and Semantic Awareness Joint-driven Image Fusion Network

This study aims to address the problem of incomplete information in unimodal images for semantic segmentation and object detection tasks. Existing multimodal fusion methods suffer from limited capability in discriminative modeling of multi-scale semantic structures and salient target regions, which further restricts the effective fusion of task-related semantic details and target information across modalities. To tackle these challenges, this paper proposes a novel fusion network termed TSJNet, which leverages the semantic information output by high-level tasks in a joint manner to guide the fusion process. Specifically, we design a multi-dimensional feature extraction module with dual parallel branches to capture multi-scale and salient features. Meanwhile, a data-agnostic spatial attention module embedded in the decoder dynamically calibrates attention allocation across different data domains, significantly enhancing the model's generalization ability. To optimize both fusion and advanced visual tasks, we balance performance by combining fusion loss with semantic losses. Additionally, we have developed a multimodal unmanned aerial vehicle (UAV) dataset covering multiple scenarios (UMS). Extensive experiments demonstrate that TSJNet achieves outstanding performance on five public datasets (MSRS, M\textsuperscript{3}FD, RoadScene, LLVIP, and TNO) and our UMS dataset. The generated fusion results exhibit favorable visual effects, and compared to state-of-the-art methods, the mean average precision (mAP@0.5) and mean intersection over union (mIoU) for object detection and segmentation, respectively, improve by 7.97\% and 10.88\%.The code and the dataset has been publicly released at https://github.com/XylonXu01/TSJNet.

cs.CV↗

arXiv2026

EAQuant: Enhancing Post-Training Quantization for MoE Models via Expert-Aware Optimization

Mixture-of-Experts (MoE) models enable scalable computation and performance in large-scale deep learning but face quantization challenges due to sparse expert activation and dynamic routing. Existing post-training quantization (PTQ) methods fail to address activation outliers, routing instability, and sparse expert calibration, leading to significant performance degradation. To address this, we propose EAQuant, a PTQ framework tailored for MoE architectures. Our method introduces three expert-aware innovations: (1) smoothing aggregation to suppress activation outliers, (2) routing consistency alignment to preserve expert selection post-quantization, and (3) calibration data balance to optimize sparsely activated experts. These strategies collectively enable robust, high-precision quantization of MoE models under ultra-low-bit constraints.Extensive experiments across several extreme quantization settings (e.g., W4A4/W3A4/W3A3/W2A4) demonstrate that EAQuant significantly outperforms existing methods, achieving average accuracy improvements of 1.15 - 13.81% across three diverse MoE architectures, with particularly pronounced gains in reasoning tasks and robust performance retention under aggressive quantization. By integrating these innovations, EAQuant establishes a new state-of-the-art for high-precision, efficient MoE model compression.Our code is available at https://github.com/darren-fzq1/EAQuant.

cs.CL↗

arXiv2026

EPD-Serve: A Flexible Multimodal EPD Disaggregation Inference Serving System On Ascend

With the widespread adoption of large multimodal models, efficient inference across text, image, audio, and video modalities has become critical. However, existing multimodal inference systems typically employ monolithic architectures that tightly couple the Encode, Prefill, and Decode stages on homogeneous hardware, neglecting the heterogeneous computational characteristics of each stage. This design leads to inefficient resource utilization and limited system throughput. To address these issues, we propose EPD-Serve, a stage-level disaggregated inference serving system for multimodal models. EPD-Serve decouples the inference pipeline into independent Encode, Prefill, and Decode stages, enabling logical isolation and flexible co-located deployment through dynamic orchestration. Leveraging the Ascend interconnect topology, EPD-Serve introduces asynchronous feature prefetching between Encode and Prefill stages and a hierarchical grouped KV cache transmission mechanism between Prefill and Decode stages to improve cross-node communication efficiency. In addition, EPD-Serve incorporates multi-route scheduling, instance-level load balancing, and multi-stage hardware co-location with spatial multiplexing to better support diverse multimodal workloads. Comprehensive experiments on multimodal understanding models demonstrate that, under high-concurrency scenarios, EPD-Serve improves end-to-end throughput by 57.37-69.48% compared to PD-disaggregated deployment, while satisfying strict SLO constraints, including TTFT below 2000 ms and TPOT below 50 ms. These results highlight the effectiveness of stage-level disaggregation for optimizing multimodal large model inference systems.

cs.DC↗

arXiv2025

JDPNet: A Network Based on Joint Degradation Processing for Underwater Image Enhancement

Given the complexity of underwater environments and the variability of water as a medium, underwater images are inevitably subject to various types of degradation. The degradations present nonlinear coupling rather than simple superposition, which renders the effective processing of such coupled degradations particularly challenging. Most existing methods focus on designing specific branches, modules, or strategies for specific degradations, with little attention paid to the potential information embedded in their coupling. Consequently, they struggle to effectively capture and process the nonlinear interactions of multiple degradations from a bottom-up perspective. To address this issue, we propose JDPNet, a joint degradation processing network, that mines and unifies the potential information inherent in coupled degradations within a unified framework. Specifically, we introduce a joint feature-mining module, along with a probabilistic bootstrap distribution strategy, to facilitate effective mining and unified adjustment of coupled degradation features. Furthermore, to balance color, clarity, and contrast, we design a novel AquaBalanceLoss to guide the network in learning from multiple coupled degradation losses. Experiments on six publicly available underwater datasets, as well as two new datasets constructed in this study, show that JDPNet exhibits state-of-the-art performance while offering a better tradeoff between performance, parameter size, and computational cost.

cs.CV↗

arXiv2025

Relativistic Core-Valence-Separated Molecular Mean-Field Exact-Two-Component Equation-of-Motion Coupled Cluster Theory: Applications to L-edge X-ray Absorption Spectroscopy

L-edge X-ray absorption spectra for first-row transition metal complexes are obtained from relativistic equation-of-motion singles and doubles coupled-cluster (EOM-CCSD) calculations that make use of the core-valence separation (CVS) scheme, with scalar and spin--orbit relativistic effects modeled within the molecular mean-field exact two-component (X2C) framework. By incorporating relativistic effects variationally at the Dirac--Coulomb--Breit (DCB) reference level, this method delivers accurate predictions of L-edge features, including energy shifts, intensity ratios, and fine-structure splittings, across a range of molecular systems. Benchmarking against perturbative spin--orbit treatments and relativistic TDDFT highlights the superior performance and robustness of the CVS-DCB-X2C-EOM-CCSD approach, including the reliability of basis set recontraction schemes. While limitations remain in describing high-density spectral regions, our results establish CVS-DCB-X2C-EOM-CCSD as a powerful and broadly applicable tool for relativistic core-excitation spectroscopy.

physics.chem-ph↗

arXiv2025

Text-Guided Channel Perturbation and Pretrained Knowledge Integration for Unified Multi-Modality Image Fusion

Multi-modality image fusion enhances scene perception by combining complementary information. Unified models aim to share parameters across modalities for multi-modality image fusion, but large modality differences often cause gradient conflicts, limiting performance. Some methods introduce modality-specific encoders to enhance feature perception and improve fusion quality. However, this strategy reduces generalisation across different fusion tasks. To overcome this limitation, we propose a unified multi-modality image fusion framework based on channel perturbation and pre-trained knowledge integration (UP-Fusion). To suppress redundant modal information and emphasize key features, we propose the Semantic-Aware Channel Pruning Module (SCPM), which leverages the semantic perception capability of a pre-trained model to filter and enhance multi-modality feature channels. Furthermore, we proposed the Geometric Affine Modulation Module (GAM), which uses original modal features to apply affine transformations on initial fusion features to maintain the feature encoder modal discriminability. Finally, we apply a Text-Guided Channel Perturbation Module (TCPM) during decoding to reshape the channel distribution, reducing the dependence on modality-specific channels. Extensive experiments demonstrate that the proposed algorithm outperforms existing methods on both multi-modality image fusion and downstream tasks.

cs.CV↗

arXiv2025

UMCFuse: A Unified Multiple Complex Scenes Infrared and Visible Image Fusion Framework

Infrared and visible image fusion has emerged as a prominent research area in computer vision. However, little attention has been paid to the fusion task in complex scenes, leading to sub-optimal results under interference. To fill this gap, we propose a unified framework for infrared and visible images fusion in complex scenes, termed UMCFuse. Specifically, we classify the pixels of visible images from the degree of scattering of light transmission, allowing us to separate fine details from overall intensity. Maintaining a balance between interference removal and detail preservation is essential for the generalization capacity of the proposed method. Therefore, we propose an adaptive denoising strategy for the fusion of detail layers. Meanwhile, we fuse the energy features from different modalities by analyzing them from multiple directions. Extensive fusion experiments on real and synthetic complex scenes datasets cover adverse weather conditions, noise, blur, overexposure, fire, as well as downstream tasks including semantic segmentation, object detection, salient object detection, and depth estimation, consistently indicate the superiority of the proposed method compared with the recent representative methods. Our code is available at https://github.com/ixilai/UMCFuse.

cs.CV↗

arXiv2025

Numerically Exact Configuration Interaction at Quadrillion-Determinant Scale

The combinatorial scaling of configuration interaction (CI) has long restricted its applicability to only the simplest molecular systems. Here, we report the first numerically exact CI calculation exceeding one quadrillion ($10^{15}$) determinants, enabled by categorical compression within the small-tensor-product distributed active space (STP-DAS) framework. As a demonstration, we converged the relativistic complete active space CI (CASCI) ground state of HBrTe involving over $10^{15}$ complex-valued 2-spinor determinants in under 34.5 hours (time-to-completion) using 1000 nodes, representing the largest CASCI calculation reported to date. Additionally, we achieved $\boldsymbolσ$-build times of just 5 minutes for systems with approximately 150 billion complex-valued 2-spinor determinants using only a few compute nodes. Extensive benchmarks confirm that the method retains numerical exactness with drastically reduced resource demands. Compared to previous state-of-the-art CI calculations, this work represents a 3-orders-of-magnitude increase in CI space, a 6-orders-of-magnitude increase in FLOP count, and a 6-orders-of-magnitude improvement in computational speed. By introducing a numerically exact, categorically compressed representation of the CI expansion vectors and reformulating the $\boldsymbolσ$-build accordingly, we eliminate memory bottlenecks associated with storing excitation lists and CI vectors while significantly reducing computational cost. A compression-compatible preconditioner further enhances performance by generating compressed CI expansion vectors throughout Davidson iterations. This work establishes a new computational frontier for numerically exact CI methods, enabling chemically and physically accurate simulations of strongly correlated, spin-orbit coupled systems previously thought to be beyond reach.

physics.chem-ph↗

arXiv2025

NDLPNet: A Location-Aware Nighttime Deraining Network and a Real-World Benchmark Dataset

Visual degradation caused by rain streak artifacts in low-light conditions significantly hampers the performance of nighttime surveillance and autonomous navigation. Existing image deraining techniques are primarily designed for daytime conditions and perform poorly under nighttime illumination due to the spatial heterogeneity of rain distribution and the impact of light-dependent stripe visibility. In this paper, we propose a novel Nighttime Deraining Location-enhanced Perceptual Network(NDLPNet) that effectively captures the spatial positional information and density distribution of rain streaks in low-light environments. Specifically, we introduce a Position Perception Module (PPM) to capture and leverage spatial contextual information from input data, enhancing the model's capability to identify and recalibrate the importance of different feature channels. The proposed nighttime deraining network can effectively remove the rain streaks as well as preserve the crucial background information. Furthermore, We construct a night scene rainy (NSR) dataset comprising 900 image pairs, all based on real-world nighttime scenes, providing a new benchmark for nighttime deraining task research. Extensive qualitative and quantitative experimental evaluations on both existing datasets and the NSR dataset consistently demonstrate our method outperform the state-of-the-art (SOTA) methods in nighttime deraining tasks. The source code and dataset is available at https://github.com/Feecuin/NDLPNet.

cs.CV↗

arXiv2025

Chirality-Driven Magnetization Emerges from Relativistic Four-Current Dynamics

Chirality-induced spin selectivity (CISS) is a striking quantum phenomenon in which electron transport through chiral molecules leads to spin polarization -- even in the absence of external magnetic fields or magnetic components. Although observed in systems such as DNA, helicenes, proteins, and polymers, the fundamental physical origin of CISS remains unresolved. Here, we introduce a time-dependent relativistic four-current framework, in which charge and current densities evolve according to the time-dependent variational principle. Real-time relativistic four-current simulations enable direct analysis of helical currents and induced magnetization dynamics. Applied to helicenes -- axially chiral molecules lacking stereocenters -- our simulations reveal curvature-induced helical electron currents that generate spontaneous magnetic fields aligned along the molecular axis. These fields are handedness-dependent and reach magnitudes of $10^{-1}$ Tesla per single helicene strand. Our results suggest that CISS may arise from intrinsic, relativistic curvature-induced helical currents and the associated magnetic fields within chiral molecules. This four-current mechanism offers a self-contained explanation for the driving force underlying spin selectivity, independent of interfacial effects or unphysically enhanced spin-orbit coupling. Furthermore, our results provide a new perspective that offers a unifying framework with the potential to reconcile many existing hypotheses and theoretical models, while also suggesting several testable predictions that can be examined experimentally.

cond-mat.mes-hall↗

arXiv2025

FlexiD-Fuse: Flexible number of inputs multi-modal medical image fusion based on diffusion model

Different modalities of medical images provide unique physiological and anatomical information for diseases. Multi-modal medical image fusion integrates useful information from different complementary medical images with different modalities, producing a fused image that comprehensively and objectively reflects lesion characteristics to assist doctors in clinical diagnosis. However, existing fusion methods can only handle a fixed number of modality inputs, such as accepting only two-modal or tri-modal inputs, and cannot directly process varying input quantities, which hinders their application in clinical settings. To tackle this issue, we introduce FlexiD-Fuse, a diffusion-based image fusion network designed to accommodate flexible quantities of input modalities. It can end-to-end process two-modal and tri-modal medical image fusion under the same weight. FlexiD-Fuse transforms the diffusion fusion problem, which supports only fixed-condition inputs, into a maximum likelihood estimation problem based on the diffusion process and hierarchical Bayesian modeling. By incorporating the Expectation-Maximization algorithm into the diffusion sampling iteration process, FlexiD-Fuse can generate high-quality fused images with cross-modal information from source images, independently of the number of input images. We compared the latest two and tri-modal medical image fusion methods, tested them on Harvard datasets, and evaluated them using nine popular metrics. The experimental results show that our method achieves the best performance in medical image fusion with varying inputs. Meanwhile, we conducted extensive extension experiments on infrared-visible, multi-exposure, and multi-focus image fusion tasks with arbitrary numbers, and compared them with the perspective SOTA methods. The results of the extension experiments consistently demonstrate the effectiveness and superiority of our method.

cs.CV↗

arXiv2025

FS-Diff: Semantic guidance and clarity-aware simultaneous multimodal image fusion and super-resolution

As an influential information fusion and low-level vision technique, image fusion integrates complementary information from source images to yield an informative fused image. A few attempts have been made in recent years to jointly realize image fusion and super-resolution. However, in real-world applications such as military reconnaissance and long-range detection missions, the target and background structures in multimodal images are easily corrupted, with low resolution and weak semantic information, which leads to suboptimal results in current fusion techniques. In response, we propose FS-Diff, a semantic guidance and clarity-aware joint image fusion and super-resolution method. FS-Diff unifies image fusion and super-resolution as a conditional generation problem. It leverages semantic guidance from the proposed clarity sensing mechanism for adaptive low-resolution perception and cross-modal feature extraction. Specifically, we initialize the desired fused result as pure Gaussian noise and introduce the bidirectional feature Mamba to extract the global features of the multimodal images. Moreover, utilizing the source images and semantics as conditions, we implement a random iterative denoising process via a modified U-Net network. This network istrained for denoising at multiple noise levels to produce high-resolution fusion results with cross-modal features and abundant semantic information. We also construct a powerful aerial view multiscene (AVMS) benchmark covering 600 pairs of images. Extensive joint image fusion and super-resolution experiments on six public and our AVMS datasets demonstrated that FS-Diff outperforms the state-of-the-art methods at multiple magnifications and can recover richer details and semantics in the fused images. The code is available at https://github.com/XylonXu01/FS-Diff.

cs.CV↗

arXiv2025

QuGStep: Refining Step Size Selection in Gradient Estimation for Variational Quantum Algorithms

Variational quantum algorithms (VQAs) offer a promising approach to solving computationally demanding problems by combining parameterized quantum circuits with classical optimization. Estimating probabilistic outcomes on quantum hardware requires repeated measurements (shots). However, in practice, the limited shot budget introduces significant noise in the evaluation of the objective function. Gradient estimation in VQAs often relies on the finite-difference, which evaluates the noisy objective function at perturbed circuit parameter values. The accuracy of this estimation is highly dependent on the choice of step size for these perturbations. An inappropriate step size can exacerbate the impact of noise, causing inaccurate gradient estimates and hindering the classical optimization in VQAs. This paper proposes QuGStep, an algorithm that addresses the challenge of determining the appropriate step size for finite-difference gradient estimation under a shot budget. QuGStep is grounded in a theorem that proves the optimal step size, which accounts for the shot budget, minimizes the error bound in gradient estimation using finite differences. Numerical experiments approximating the ground state energy of several molecules demonstrate that QuGStep can identify the appropriate step size for the given shot budget to obtain effective gradient estimation. Notably, the step size identified by QuGStep achieved convergence to the ground state energy with over 94% fewer shots compared to using a default step size (i.e., step size of 0.01). These findings highlight the potential of QuGStep to improve the practical deployment and scalability of quantum computing technologies.

quant-ph↗

arXiv2025

AWM-Fuse: Multi-Modality Image Fusion for Adverse Weather via Global and Local Text Perception

Multi-modality image fusion (MMIF) in adverse weather aims to address the loss of visual information caused by weather-related degradations, providing clearer scene representations. Although less studies have attempted to incorporate textual information to improve semantic perception, they often lack effective categorization and thorough analysis of textual content. In response, we propose AWM-Fuse, a novel fusion method for adverse weather conditions, designed to handle multiple degradations through global and local text perception within a unified, shared weight architecture. In particular, a global feature perception module leverages BLIP-produced captions to extract overall scene features and identify primary degradation types, thus promoting generalization across various adverse weather conditions. Complementing this, the local module employs detailed scene descriptions produced by ChatGPT to concentrate on specific degradation effects through concrete textual cues, thereby capturing finer details. Furthermore, textual descriptions are used to constrain the generation of fusion images, effectively steering the network learning process toward better alignment with real semantic labels, thereby promoting the learning of more meaningful visual features. Extensive experiments demonstrate that AWM-Fuse outperforms current state-of-the-art methods in complex weather conditions and downstream tasks. Our code is available at https://github.com/Feecuin/AWM-Fuse.

cs.CV↗

arXiv2025

MambaTrans: Multimodal Fusion Image Translation via Large Language Model Priors for Downstream Visual Tasks

The goal of multimodal image fusion is to integrate complementary information from infrared and visible images, generating multimodal fused images for downstream tasks. Existing downstream pre-training models are typically trained on visible images. However, the significant pixel distribution differences between visible and multimodal fusion images can degrade downstream task performance, sometimes even below that of using only visible images. This paper explores adapting multimodal fused images with significant modality differences to object detection and semantic segmentation models trained on visible images. To address this, we propose MambaTrans, a novel multimodal fusion image modality translator. MambaTrans uses descriptions from a multimodal large language model and masks from semantic segmentation models as input. Its core component, the Multi-Model State Space Block, combines mask-image-text cross-attention and a 3D-Selective Scan Module, enhancing pure visual capabilities. By leveraging object detection prior knowledge, MambaTrans minimizes detection loss during training and captures long-term dependencies among text, masks, and images. This enables favorable results in pre-trained models without adjusting their parameters. Experiments on public datasets show that MambaTrans effectively improves multimodal image performance in downstream tasks.

cs.CV↗