arXiv subjects

Shuang Xu

Publications and source records attributed to Shuang Xu.

At least 19 recordsLinked to original sources

arXiv2026

World Models for Policy Refinement in StarCraft II

Large Language Models (LLMs) have recently shown strong reasoning capabilities, motivating their use in complex decision-making environments. StarCraft II (SC2), with its massive state-action space and partial observability, is a challenging testbed. However, existing LLM-based SC2 agents primarily focus on improving the policy itself, leaving the integration of a learnable, action-conditioned dynamics model into the decision loop largely unexplored. In this work, we propose StarWM and StarWM-Agent, and conduct the first systematic study of the learnability and decision utility of player-view, action-conditioned textual world models for SC2. StarWM predicts short-horizon future observations under partial observability. StarWM-Agent integrates StarWM into a lightweight Generate-Simulate-Refine loop for foresight-driven policy refinement. Extensive experiments show that StarWM substantially outperforms zero-shot baselines across multiple dimensions, while StarWM-Agent achieves consistent win-rate gains of 30%, 15%, and 30% against the SC2 built-in AI at Hard (LV5), Harder (LV6), and VeryHard (LV7), respectively. Additional analyses show that our method is complementary to existing history-summarization SC2-agent approaches, with their combination achieving the strongest online performance in our evaluation.

cs.AI↗

arXiv2025

MrCoM: A Meta-Regularized World-Model Generalizing Across Multi-Scenarios

Model-based reinforcement learning (MBRL) is a crucial approach to enhance the generalization capabilities and improve the sample efficiency of RL algorithms. However, current MBRL methods focus primarily on building world models for single tasks and rarely address generalization across different scenarios. Building on the insight that dynamics within the same simulation engine share inherent properties, we attempt to construct a unified world model capable of generalizing across different scenarios, named Meta-Regularized Contextual World-Model (MrCoM). This method first decomposes the latent state space into various components based on the dynamic characteristics, thereby enhancing the accuracy of world-model prediction. Further, MrCoM adopts meta-state regularization to extract unified representation of scenario-relevant information, and meta-value regularization to align world-model optimization with policy learning across diverse scenario objectives. We theoretically analyze the generalization error upper bound of MrCoM in multi-scenario settings. We systematically evaluate our algorithm's generalization ability across diverse scenarios, demonstrating significantly better performance than previous state-of-the-art methods.

cs.LG↗

arXiv2025

DAIL: Beyond Task Ambiguity for Language-Conditioned Reinforcement Learning

Comprehending natural language and following human instructions are critical capabilities for intelligent agents. However, the flexibility of linguistic instructions induces substantial ambiguity across language-conditioned tasks, severely degrading algorithmic performance. To address these limitations, we present a novel method named DAIL (Distributional Aligned Learning), featuring two key components: distributional policy and semantic alignment. Specifically, we provide theoretical results that the value distribution estimation mechanism enhances task differentiability. Meanwhile, the semantic alignment module captures the correspondence between trajectories and linguistic instructions. Extensive experimental results on both structured and visual observation benchmarks demonstrate that DAIL effectively resolves instruction ambiguities, achieving superior performance to baseline methods. Our implementation is available at https://github.com/RunpengXie/Distributional-Aligned-Learning.

cs.AI↗

arXiv2025

Seedream 4.0: Toward Next-generation Multimodal Image Generation

We introduce Seedream 4.0, an efficient and high-performance multimodal image generation system that unifies text-to-image (T2I) synthesis, image editing, and multi-image composition within a single framework. We develop a highly efficient diffusion transformer with a powerful VAE which also can reduce the number of image tokens considerably. This allows for efficient training of our model, and enables it to fast generate native high-resolution images (e.g., 1K-4K). Seedream 4.0 is pretrained on billions of text-image pairs spanning diverse taxonomies and knowledge-centric concepts. Comprehensive data collection across hundreds of vertical scenarios, coupled with optimized strategies, ensures stable and large-scale training, with strong generalization. By incorporating a carefully fine-tuned VLM model, we perform multi-modal post-training for training both T2I and image editing tasks jointly. For inference acceleration, we integrate adversarial distillation, distribution matching, and quantization, as well as speculative decoding. It achieves an inference time of up to 1.8 seconds for generating a 2K image (without a LLM/VLM as PE model). Comprehensive evaluations reveal that Seedream 4.0 can achieve state-of-the-art results on both T2I and multimodal image editing. In particular, it demonstrates exceptional multimodal capabilities in complex tasks, including precise image editing and in-context reasoning, and also allows for multi-image reference, and can generate multiple output images. This extends traditional T2I systems into an more interactive and multidimensional creative tool, pushing the boundary of generative AI for both creativity and professional applications. We further scale our model and data as Seedream 4.5. Seedream 4.0 and Seedream 4.5 are accessible on Volcano Engine https://www.volcengine.com/experience/ark?launch=seedream.

cs.CV↗

arXiv2025

SC2Arena and StarEvolve: Benchmark and Self-Improvement Framework for LLMs in Complex Decision-Making Tasks

Evaluating large language models (LLMs) in complex decision-making is essential for advancing AI's ability for strategic planning and real-time adaptation. However, existing benchmarks for tasks like StarCraft II fail to capture the game's full complexity, such as its complete game context, diverse action spaces, and all playable races. To address this gap, we present SC2Arena, a benchmark that fully supports all playable races, low-level action spaces, and optimizes text-based observations to tackle spatial reasoning challenges. Complementing this, we introduce StarEvolve, a hierarchical framework that integrates strategic planning with tactical execution, featuring iterative self-correction and continuous improvement via fine-tuning on high-quality gameplay data. Its key components include a Planner-Executor-Verifier structure to break down gameplay, and a scoring system for selecting high-quality training samples. Comprehensive analysis using SC2Arena provides valuable insights into developing generalist agents that were not possible with previous benchmarks. Experimental results also demonstrate that our proposed StarEvolve achieves superior performance in strategic planning. Our code, environment, and algorithms are publicly available.

cs.LG↗

arXiv2025

Beyond Low-rankness: Guaranteed Matrix Recovery via Modified Nuclear Norm

The nuclear norm (NN) has been widely explored in matrix recovery problems, such as Robust PCA and matrix completion, leveraging the inherent global low-rank structure of the data. In this study, we introduce a new modified nuclear norm (MNN) framework, where the MNN family norms are defined by adopting suitable transformations and performing the NN on the transformed matrix. The MNN framework offers two main advantages: (1) it jointly captures both local information and global low-rankness without requiring trade-off parameter tuning; (2) Under mild assumptions on the transformation, we provided exact theoretical recovery guarantees for both Robust PCA and MC tasks-an achievement not shared by existing methods that combine local and global information. Thanks to its general and flexible design, MNN can accommodate various proven transformations, enabling a unified and effective approach to structured low-rank recovery. Extensive experiments demonstrate the effectiveness of our method. Code and supplementary material are available at https://github.com/andrew-pengjj/modified_nuclear_norm.

cs.CV↗

arXiv2025

DPMT: Dual Process Multi-scale Theory of Mind Framework for Real-time Human-AI Collaboration

Real-time human-artificial intelligence (AI) collaboration is crucial yet challenging, especially when AI agents must adapt to diverse and unseen human behaviors in dynamic scenarios. Existing large language model (LLM) agents often fail to accurately model the complex human mental characteristics such as domain intentions, especially in the absence of direct communication. To address this limitation, we propose a novel dual process multi-scale theory of mind (DPMT) framework, drawing inspiration from cognitive science dual process theory. Our DPMT framework incorporates a multi-scale theory of mind (ToM) module to facilitate robust human partner modeling through mental characteristic reasoning. Experimental results demonstrate that DPMT significantly enhances human-AI collaboration, and ablation studies further validate the contributions of our multi-scale ToM in the slow system.

cs.LG↗

arXiv2025

Non-Markovian dynamics with a driven three-level giant atom in a semi-infinite photonic waveguide

The non-Markovian effects of open quantum systems subjected to external environments are deemed to be valuable resources in quantum optics and quantum information processing. In this work, we investigate the non-Markovian dynamics of a three-level giant atom coupling with a semi-infinite photonic waveguide through multiple coupling points and driven by a classical driving field. We derive the analytical expressions for the probability amplitudes of the driven three-level giant atom and obtain two independent conditions. We find two different types of bound states (including the static bound states and the periodic equal-amplitude oscillating bound states) and discuss the physical origins of the bound states formation. Moreover, we discuss the case of the driven three-level giant atom interacting with the infinite photonic waveguide, where there is only one purely imaginary solution (i.e., only one bound state condition exists) for its complex frequency (coming from the absence of mirror at one end of the waveguide) compared to that of a driven three-level giant atom coupling with a semi-infinite photonic waveguide. With this, we also find two different types of bound states, including the static bound state and the periodic equal-amplitude oscillating bound states. Finally, the above results are generalized to a more general model involving a semi-infinite photonic waveguide coupling with an arbitrary number of noninteracting three-level giant atoms driven by the driving fields. The proposed protocol could provide a pathway to precisely elucidate the non-Markovian dynamics of driven, multi-level giant atoms coupled to semi-infinite or infinite photonic waveguides.

physics.optics↗

arXiv2025

Dressed bound states and non-Markovian dynamics with a whispering-gallery-mode microcavity coupled to a two-level atom and a semi-infinite photonic waveguide

We investigate the dressed bound states (DBS) in an open cavity with a whispering-gallery-mode microring coupled to a two-level atom and a waveguide with a mirror at the right end. We demonstrate that the non-Hermiticity of an open cavity facilitates the formation of the DBS, which consists of the vacancy-like DBS and Friedrich-Wintgen DBS. By deriving analytical conditions for these DBS, we show that when a two-level atom couples to the standing-wave mode that corresponds to a node of the photonic wave function the vacancy-like DBS occur, which are characterized by null spectral density at cavity resonance. Conversely, Friedrich-Wintgen DBS can be realized by continuously adjusting system parameters and indicated by the disappearance of the Rabi peak in the emission spectrum, which is a distinctive feature in the strong-coupling regime. Moreover, we extend our analysis to the non-Markovian regime and find that our results are consistent with those obtained under the Markovian approximation in the wideband limit. In the non-Markovian regime, we analyze DBS for both zero and non-zero accumulated phase factors. For zero accumulated phase factors, the non-Markovian regime exhibits higher peak values and longer relaxation times for vacancy-like DBS compared to the Markovian regime, where the Friedrich-Wintgen DBS are absent in the non-Markovian case. Finally, we establish the correspondence between the energy spectrum and bound state conditions for non-zero accumulated phase factors and analyze the influence of various parameters on non-Markovian bound states. Our work exhibits bound state manipulations through non-Markovian open quantum system, which holds great potential for building high-performance quantum devices for applications such as sensing, photon storage, and nonclassical light generation.

physics.optics↗

arXiv2025

LVC: A Lightweight Compression Framework for Enhancing VLMs in Long Video Understanding

Long video understanding is a complex task that requires both spatial detail and temporal awareness. While Vision-Language Models (VLMs) obtain frame-level understanding capabilities through multi-frame input, they suffer from information loss due to the sparse sampling strategy. In contrast, Video Large Language Models (Video-LLMs) capture temporal relationships within visual features but are limited by the scarcity of high-quality video-text datasets. To transfer long video understanding capabilities to VLMs with minimal data and computational cost, we propose Lightweight Video Compression (LVC), a novel method featuring the Query-Attention Video Compression mechanism, which effectively tackles the sparse sampling problem in VLMs. By training only the alignment layer with 10k short video-text pairs, LVC significantly enhances the temporal reasoning abilities of VLMs. Extensive experiments show that LVC provides consistent performance improvements across various models, including the InternVL2 series and Phi-3.5-Vision. Notably, the InternVL2-40B-LVC achieves scores of 68.2 and 65.9 on the long video understanding benchmarks MLVU and Video-MME, respectively, with relative improvements of 14.6% and 7.7%. The enhanced models and code will be publicly available soon.

cs.CV↗

arXiv2025

UniViTAR: Unified Vision Transformer with Native Resolution

Conventional Vision Transformer simplifies visual modeling by standardizing input resolutions, often disregarding the variability of natural visual data and compromising spatial-contextual fidelity. While preliminary explorations have superficially investigated native resolution modeling, existing approaches still lack systematic analysis from a visual representation perspective. To bridge this gap, we introduce UniViTAR, a family of homogeneous vision foundation models tailored for unified visual modality and native resolution scenario in the era of multimodal. Our framework first conducts architectural upgrades to the vanilla paradigm by integrating multiple advanced components. Building upon these improvements, a progressive training paradigm is introduced, which strategically combines two core mechanisms: (1) resolution curriculum learning, transitioning from fixed-resolution pretraining to native resolution tuning, thereby leveraging ViT's inherent adaptability to variable-length sequences, and (2) visual modality adaptation via inter-batch image-video switching, which balances computational efficiency with enhanced temporal reasoning. In parallel, a hybrid training framework further synergizes sigmoid-based contrastive loss with feature distillation from a frozen teacher model, thereby accelerating early-stage convergence. Finally, trained exclusively on public datasets, externsive experiments across multiple model scales from 0.3B to 1B demonstrate its effectiveness.

cs.CV↗

arXiv2025

Nonreciprocal quantum router with non-Markovian environments

Quantum routers are essential elements of quantum networks, enabling coherent information transfer between distant nodes. While their behavior has been extensively studied under Markovian approximations, investigations in non-Markovian regimes remain limited. In this paper, we study a nonreciprocal quantum router embedded in non-Markovian environments, enabling directional control of single photons, which allows transmission from one side while blocking it from the other. The cascade system under study consists of two quantum nodes: one comprising two coupled coplanar-waveguide resonators and the other featuring a superconducting ring resonator. Each node is respectively coupled to a single Yttrium iron garnet (YIG) disk, with nonreciprocity arising from the selective coupling between magnons and microwave photons in our model. We analytically derive the transmission and reflection spectra of the system when a photon is input respectively from the left and right sides of the transmission line in the non-Markovian regimes. Our results demonstrate that, with appropriate parameters, a single photon can be routed from a given input port to either of the two output ports, while being fully absorbed when incident from the opposite side. We further compare the scattering behavior in non-Markovian and Markovian regimes through numerical simulations. In the non-Markovian case, the transmission spectrum exhibits two unity peaks (two valleys with a minimum value of zero), whereas in the Markovian case, high transmission appears only within a narrow window near zero detuning when the photon is injected from the left. As the environmental bandwidth increases, non-Markovian results converge to the Markovian limit. This formalism may enable new applications in quantum information and communication exploiting non-Markovianity.

physics.optics↗

arXiv2025

Retinex-MEF: Retinex-based Glare Effects Aware Unsupervised Multi-Exposure Image Fusion

Multi-exposure image fusion (MEF) synthesizes multiple, differently exposed images of the same scene into a single, well-exposed composite. Retinex theory, which separates image illumination from scene reflectance, provides a natural framework to ensure consistent scene representation and effective information fusion across varied exposure levels. However, the conventional pixel-wise multiplication of illumination and reflectance inadequately models the glare effect induced by overexposure. To address this limitation, we introduce an unsupervised and controllable method termed Retinex-MEF. Specifically, our method decomposes multi-exposure images into separate illumination components with a shared reflectance component, and effectively models the glare induced by overexposure. The shared reflectance is learned via a bidirectional loss, which enables our approach to effectively mitigate the glare effect. Furthermore, we introduce a controllable exposure fusion criterion, enabling global exposure adjustments while preserving contrast, thus overcoming the constraints of a fixed exposure level. Extensive experiments on diverse datasets, including underexposure-overexposure fusion, exposure controlled fusion, and homogeneous extreme exposure fusion, demonstrate the effective decomposition and flexible fusion capability of our model. The code is available at https://github.com/HaowenBai/Retinex-MEF

cs.CV↗

arXiv2025

Deep Unfolding Multi-modal Image Fusion Network via Attribution Analysis

Multi-modal image fusion synthesizes information from multiple sources into a single image, facilitating downstream tasks such as semantic segmentation. Current approaches primarily focus on acquiring informative fusion images at the visual display stratum through intricate mappings. Although some approaches attempt to jointly optimize image fusion and downstream tasks, these efforts often lack direct guidance or interaction, serving only to assist with a predefined fusion loss. To address this, we propose an ``Unfolding Attribution Analysis Fusion network'' (UAAFusion), using attribution analysis to tailor fused images more effectively for semantic segmentation, enhancing the interaction between the fusion and segmentation. Specifically, we utilize attribution analysis techniques to explore the contributions of semantic regions in the source images to task discrimination. At the same time, our fusion algorithm incorporates more beneficial features from the source images, thereby allowing the segmentation to guide the fusion process. Our method constructs a model-driven unfolding network that uses optimization objectives derived from attribution analysis, with an attribution fusion loss calculated from the current state of the segmentation network. We also develop a new pathway function for attribution analysis, specifically tailored to the fusion tasks in our unfolding network. An attribution attention mechanism is integrated at each network stage, allowing the fusion network to prioritize areas and pixels crucial for high-level recognition tasks. Additionally, to mitigate the information loss in traditional unfolding networks, a memory augmentation module is incorporated into our network to improve the information flow across various network layers. Extensive experiments demonstrate our method's superiority in image fusion and applicability to semantic segmentation.

cs.CV↗

arXiv2024

Hipandas: Hyperspectral Image Joint Denoising and Super-Resolution by Image Fusion with the Panchromatic Image

Hyperspectral images (HSIs) are frequently noisy and of low resolution due to the constraints of imaging devices. Recently launched satellites can concurrently acquire HSIs and panchromatic (PAN) images, enabling the restoration of HSIs to generate clean and high-resolution imagery through fusing PAN images for denoising and super-resolution. However, previous studies treated these two tasks as independent processes, resulting in accumulated errors. This paper introduces \textbf{H}yperspectral \textbf{I}mage Joint \textbf{Pand}enoising \textbf{a}nd Pan\textbf{s}harpening (Hipandas), a novel learning paradigm that reconstructs HRHS images from noisy low-resolution HSIs (LRHS) and high-resolution PAN images. The proposed zero-shot Hipandas framework consists of a guided denoising network, a guided super-resolution network, and a PAN reconstruction network, utilizing an HSI low-rank prior and a newly introduced detail-oriented low-rank prior. The interconnection of these networks complicates the training process, necessitating a two-stage training strategy to ensure effective training. Experimental results on both simulated and real-world datasets indicate that the proposed method surpasses state-of-the-art algorithms, yielding more accurate and visually pleasing HRHS images.

cs.CV↗

arXiv2024

Task-driven Image Fusion with Learnable Fusion Loss

Multi-modal image fusion aggregates information from multiple sensor sources, achieving superior visual quality and perceptual features compared to single-source images, often improving downstream tasks. However, current fusion methods for downstream tasks still use predefined fusion objectives that potentially mismatch the downstream tasks, limiting adaptive guidance and reducing model flexibility. To address this, we propose Task-driven Image Fusion (TDFusion), a fusion framework incorporating a learnable fusion loss guided by task loss. Specifically, our fusion loss includes learnable parameters modeled by a neural network called the loss generation module. This module is supervised by the downstream task loss in a meta-learning manner. The learning objective is to minimize the task loss of fused images after optimizing the fusion module with the fusion loss. Iterative updates between the fusion module and the loss module ensure that the fusion network evolves toward minimizing task loss, guiding the fusion process toward the task objectives. TDFusion's training relies entirely on the downstream task loss, making it adaptable to any specific task. It can be applied to any architecture of fusion and task networks. Experiments demonstrate TDFusion's performance through fusion experiments conducted on four different datasets, in addition to evaluations on semantic segmentation and object detection tasks.

cs.CV↗

arXiv2024

HSIGene: A Foundation Model For Hyperspectral Image Generation

Hyperspectral image (HSI) plays a vital role in various fields such as agriculture and environmental monitoring. However, due to the expensive acquisition cost, the number of hyperspectral images is limited, degenerating the performance of downstream tasks. Although some recent studies have attempted to employ diffusion models to synthesize HSIs, they still struggle with the scarcity of HSIs, affecting the reliability and diversity of the generated images. Some studies propose to incorporate multi-modal data to enhance spatial diversity, but the spectral fidelity cannot be ensured. In addition, existing HSI synthesis models are typically uncontrollable or only support single-condition control, limiting their ability to generate accurate and reliable HSIs. To alleviate these issues, we propose HSIGene, a novel HSI generation foundation model which is based on latent diffusion and supports multi-condition control, allowing for more precise and reliable HSI generation. To enhance the spatial diversity of the training data while preserving spectral fidelity, we propose a new data augmentation method based on spatial super-resolution, in which HSIs are upscaled first, and thus abundant training patches could be obtained by cropping the high-resolution HSIs. In addition, to improve the perceptual quality of the augmented data, we introduce a novel two-stage HSI super-resolution framework, which first applies RGB bands super-resolution and then utilizes our proposed Rectangular Guided Attention Network (RGAN) for guided HSI super-resolution. Experiments demonstrate that the proposed model is capable of generating a vast quantity of realistic HSIs for downstream tasks such as denoising and super-resolution. The code and models are available at https://github.com/LiPang/HSIGene.

cs.CV↗

arXiv2024

Haar Nuclear Norms with Applications to Remote Sensing Imagery Restoration

Remote sensing image restoration aims to reconstruct missing or corrupted areas within images. To date, low-rank based models have garnered significant interest in this field. This paper proposes a novel low-rank regularization term, named the Haar nuclear norm (HNN), for efficient and effective remote sensing image restoration. It leverages the low-rank properties of wavelet coefficients derived from the 2-D frontal slice-wise Haar discrete wavelet transform, effectively modeling the low-rank prior for separated coarse-grained structure and fine-grained textures in the image. Experimental evaluations conducted on hyperspectral image inpainting, multi-temporal image cloud removal, and hyperspectral image denoising have revealed the HNN's potential. Typically, HNN achieves a performance improvement of 1-4 dB and a speedup of 10-28x compared to some state-of-the-art methods (e.g., tensor correlated total variation, and fully-connected tensor network) for inpainting tasks.

eess.IV↗