arXiv subjects

Zhaoyang Zhang

Publications and source records attributed to Zhaoyang Zhang.

At least 109 records · Page 6Linked to original sources

arXiv2025

Computational Imaging-Based ISAC Method with Large Pixel Division

One of the key points in designing an integrated sensing and communication (ISAC) system using computational imaging is the division size of imaging pixels. If the size is too small, it leads to a high number of pixels that need processing. On the contrary, it usually causes large processing errors since each pixel is no longer uniformly coherent. In this paper, a novel method is proposed to address such a problem in environment sensing in millimeter-wave wireless cellular networks, which effectively cancels the severe errors caused by large pixel division as in conventional computational imaging algorithms. To this end, a novel computational imaging model in an integral form is introduced, which leverages the continuous characteristics of object surfaces in the environment and takes into account the different phases associated with the different parts of the pixel. The proposed algorithm extends computational imaging to large wireless communication scenarios for the first time. The performance of the proposed method is then analyzed, and extensive numerical results verify its effectiveness.

eess.SP↗

arXiv2025

Cobra: Efficient Line Art COlorization with BRoAder References

The comic production industry requires reference-based line art colorization with high accuracy, efficiency, contextual consistency, and flexible control. A comic page often involves diverse characters, objects, and backgrounds, which complicates the coloring process. Despite advancements in diffusion models for image generation, their application in line art colorization remains limited, facing challenges related to handling extensive reference images, time-consuming inference, and flexible control. We investigate the necessity of extensive contextual image guidance on the quality of line art colorization. To address these challenges, we introduce Cobra, an efficient and versatile method that supports color hints and utilizes over 200 reference images while maintaining low latency. Central to Cobra is a Causal Sparse DiT architecture, which leverages specially designed positional encodings, causal sparse attention, and Key-Value Cache to effectively manage long-context references and ensure color identity consistency. Results demonstrate that Cobra achieves accurate line art colorization through extensive contextual reference, significantly enhancing inference speed and interactivity, thereby meeting critical industrial demands. We release our codes and models on our project page: https://zhuang2002.github.io/Cobra/.

cs.CV↗

arXiv2025

FlexiAct: Towards Flexible Action Control in Heterogeneous Scenarios

Action customization involves generating videos where the subject performs actions dictated by input control signals. Current methods use pose-guided or global motion customization but are limited by strict constraints on spatial structure, such as layout, skeleton, and viewpoint consistency, reducing adaptability across diverse subjects and scenarios. To overcome these limitations, we propose FlexiAct, which transfers actions from a reference video to an arbitrary target image. Unlike existing methods, FlexiAct allows for variations in layout, viewpoint, and skeletal structure between the subject of the reference video and the target image, while maintaining identity consistency. Achieving this requires precise action control, spatial structure adaptation, and consistency preservation. To this end, we introduce RefAdapter, a lightweight image-conditioned adapter that excels in spatial adaptation and consistency preservation, surpassing existing methods in balancing appearance consistency and structural flexibility. Additionally, based on our observations, the denoising process exhibits varying levels of attention to motion (low frequency) and appearance details (high frequency) at different timesteps. So we propose FAE (Frequency-aware Action Extraction), which, unlike existing methods that rely on separate spatial-temporal architectures, directly achieves action extraction during the denoising process. Experiments demonstrate that our method effectively transfers actions to subjects with diverse layouts, skeletons, and viewpoints. We release our code and model weights to support further research at https://shiyi-zh0408.github.io/projectpages/FlexiAct/

cs.CV↗

arXiv2025

BrushEdit: All-In-One Image Inpainting and Editing

Image editing has advanced significantly with the development of diffusion models using both inversion-based and instruction-based methods. However, current inversion-based approaches struggle with big modifications (e.g., adding or removing objects) due to the structured nature of inversion noise, which hinders substantial changes. Meanwhile, instruction-based methods often constrain users to black-box operations, limiting direct interaction for specifying editing regions and intensity. To address these limitations, we propose BrushEdit, a novel inpainting-based instruction-guided image editing paradigm, which leverages multimodal large language models (MLLMs) and image inpainting models to enable autonomous, user-friendly, and interactive free-form instruction editing. Specifically, we devise a system enabling free-form instruction editing by integrating MLLMs and a dual-branch image inpainting model in an agent-cooperative framework to perform editing category classification, main object identification, mask acquisition, and editing area inpainting. Extensive experiments show that our framework effectively combines MLLMs and inpainting models, achieving superior performance across seven metrics including mask region preservation and editing effect coherence.

cs.CV↗

arXiv2025

ColorFlow: Retrieval-Augmented Image Sequence Colorization

Automatic black-and-white image sequence colorization while preserving character and object identity (ID) is a complex task with significant market demand, such as in cartoon or comic series colorization. Despite advancements in visual colorization using large-scale generative models like diffusion models, challenges with controllability and identity consistency persist, making current solutions unsuitable for industrial application.To address this, we propose ColorFlow, a three-stage diffusion-based framework tailored for image sequence colorization in industrial applications. Unlike existing methods that require per-ID finetuning or explicit ID embedding extraction, we propose a novel robust and generalizable Retrieval Augmented Colorization pipeline for colorizing images with relevant color references. Our pipeline also features a dual-branch design: one branch for color identity extraction and the other for colorization, leveraging the strengths of diffusion models. We utilize the self-attention mechanism in diffusion models for strong in-context learning and color identity matching. To evaluate our model, we introduce ColorFlow-Bench, a comprehensive benchmark for reference-based colorization. Results show that ColorFlow outperforms existing models across multiple metrics, setting a new standard in sequential image colorization and potentially benefiting the art industry. We release our codes and models on our project page: https://zhuang2002.github.io/ColorFlow/.

cs.CV↗

arXiv2025

Robust Deep Learning-Based Physical Layer Communications: Strategies and Approaches

Deep learning (DL) has emerged as a transformative technology with immense potential to reshape the sixth-generation (6G) wireless communication network. By utilizing advanced algorithms for feature extraction and pattern recognition, DL provides unprecedented capabilities in optimizing the network efficiency and performance, particularly in physical layer communications. Although DL technologies present the great potential, they also face significant challenges related to the robustness, which are expected to intensify in the complex and demanding 6G environment. Specifically, current DL models typically exhibit substantial performance degradation in dynamic environments with time-varying channels, interference of noise and different scenarios, which affect their effectiveness in diverse real-world applications. This paper provides a comprehensive overview of strategies and approaches for robust DL-based methods in physical layer communications. First we introduce the key challenges that current DL models face. Then we delve into a detailed examination of DL approaches specifically tailored to enhance robustness in 6G, which are classified into data-driven and model-driven strategies. Finally, we verify the effectiveness of these methods by case studies and outline future research directions.

cs.IT↗

arXiv2025

Communication-Efficient Personalized Distributed Learning with Data and Node Heterogeneity

To jointly tackle the challenges of data and node heterogeneity in decentralized learning, we propose a distributed strong lottery ticket hypothesis (DSLTH), based on which a communication-efficient personalized learning algorithm is developed. In the proposed method, each local model is represented as the Hadamard product of global real-valued parameters and a personalized binary mask for pruning. The local model is learned by updating and fusing the personalized binary masks while the real-valued parameters are fixed among different agents. To further reduce the complexity of hardware implementation, we incorporate a group sparse regularization term in the loss function, enabling the learned local model to achieve structured sparsity. Then, a binary mask aggregation algorithm is designed by introducing an intermediate aggregation tensor and adding a personalized fine-tuning step in each iteration, which constrains model updates towards the local data distribution. The proposed method effectively leverages the relativity among agents while meeting personalized requirements in heterogeneous node conditions. We also provide a theoretical proof for the DSLTH, establishing it as the foundation of the proposed method. Numerical simulations confirm the validity of the DSLTH and demonstrate the effectiveness of the proposed algorithm.

cs.LG↗

arXiv2025

VideoPainter: Any-length Video Inpainting and Editing with Plug-and-Play Context Control

Video inpainting, which aims to restore corrupted video content, has experienced substantial progress. Despite these advances, existing methods, whether propagating unmasked region pixels through optical flow and receptive field priors, or extending image-inpainting models temporally, face challenges in generating fully masked objects or balancing the competing objectives of background context preservation and foreground generation in one model, respectively. To address these limitations, we propose a novel dual-stream paradigm VideoPainter that incorporates an efficient context encoder (comprising only 6% of the backbone parameters) to process masked videos and inject backbone-aware background contextual cues to any pre-trained video DiT, producing semantically consistent content in a plug-and-play manner. This architectural separation significantly reduces the model's learning complexity while enabling nuanced integration of crucial background context. We also introduce a novel target region ID resampling technique that enables any-length video inpainting, greatly enhancing our practical applicability. Additionally, we establish a scalable dataset pipeline leveraging current vision understanding models, contributing VPData and VPBench to facilitate segmentation-based inpainting training and assessment, the largest video inpainting dataset and benchmark to date with over 390K diverse clips. Using inpainting as a pipeline basis, we also explore downstream applications including video editing and video editing pair data generation, demonstrating competitive performance and significant practical potential. Extensive experiments demonstrate VideoPainter's superior performance in both any-length video inpainting and editing, across eight key metrics, including video quality, mask region preservation, and textual coherence.

cs.CV↗

arXiv2025

Fair Resource Allocation in UAV-based Semantic Communication System with Fluid Antenna

In this paper, the problem of maximization of the minimum equivalent rate in a unmanned-aerial-vehicle (UAV)-based multi-user semantic communication system is investigated. In the considered model, a multi-antenna UAV employs semantic extraction techniques to compress the data ready to be sent to the users, which are equipped with fluid antennas. Our aim is to jointly optimize the trajectory of the UAV, the transmit beamforming and the semantic compression rate at the UAV, as well as the selection of activated ports in fluid antenna system (FAS), to maximize the minimum equivalent transmission rate among all user. An alternating algorithm is designed to solve the problem. Simulation results validate the effectiveness of the proposed algorithm.

eess.SP↗

arXiv2025

Spin-to-orbital angular momentum conversion in non-Hermitian photonic graphene

Optical beams with orbital angular momentum (OAM) have numerous potential applications, but the means used for their generation often lack crucial on-demand control. In this work, we present a mechanism of converting spin angular momentum (SAM) to OAM in a non-structured beam. The conversion occurs through spin-orbit coupling in a reconfigurable photonic honeycomb lattice with staggering implemented by electromagnetically-induced transparency in an atomic vapor cell. The spin-orbit coupling allows to outcouple the OAM signal from a particular band in a given valley determined by the chirality of light or the lattice staggering, providing a non-zero Berry curvature for generating OAM. The dependence of the output OAM on the chirality of the input beam is the first control knob. The staggering works as a second control knob, flipping the sign of OAM for the fixed chirality. The demonstrated conversion between SAM and OAM is important for optical communications. Our results can be extended to other implementations of paraxial photonic graphene.

cond-mat.mes-hall↗

arXiv2025

DiTCtrl: Exploring Attention Control in Multi-Modal Diffusion Transformer for Tuning-Free Multi-Prompt Longer Video Generation

Sora-like video generation models have achieved remarkable progress with a Multi-Modal Diffusion Transformer MM-DiT architecture. However, the current video generation models predominantly focus on single-prompt, struggling to generate coherent scenes with multiple sequential prompts that better reflect real-world dynamic scenarios. While some pioneering works have explored multi-prompt video generation, they face significant challenges including strict training data requirements, weak prompt following, and unnatural transitions. To address these problems, we propose DiTCtrl, a training-free multi-prompt video generation method under MM-DiT architectures for the first time. Our key idea is to take the multi-prompt video generation task as temporal video editing with smooth transitions. To achieve this goal, we first analyze MM-DiT's attention mechanism, finding that the 3D full attention behaves similarly to that of the cross/self-attention blocks in the UNet-like diffusion models, enabling mask-guided precise semantic control across different prompts with attention sharing for multi-prompt video generation. Based on our careful design, the video generated by DiTCtrl achieves smooth transitions and consistent object motion given multiple sequential prompts without additional training. Besides, we also present MPVBench, a new benchmark specially designed for multi-prompt video generation to evaluate the performance of multi-prompt generation. Extensive experiments demonstrate that our method achieves state-of-the-art performance without additional training.

cs.CV↗

arXiv2025

Electromagnetic Channel Modeling and Capacity Analysis for HMIMO Communications

Advancements in emerging technologies, e.g., reconfigurable intelligent surfaces and holographic MIMO (HMIMO), facilitate unprecedented manipulation of electromagnetic (EM) waves, significantly enhancing the performance of wireless communication systems. To accurately characterize the achievable performance limits of these systems, it is crucial to develop a universal EM-compliant channel model. This paper addresses this necessity by proposing a comprehensive EM channel model tailored for realistic multi-path environments, accounting for the combined effects of antenna array configurations and propagation conditions in HMIMO communications. Both polarization phenomena and spatial correlation are incorporated into this probabilistic channel model. Additionally, physical constraints of antenna configurations, such as mutual coupling effects and energy consumption, are integrated into the channel modeling framework. Simulation results validate the effectiveness of the proposed probabilistic channel model, indicating that traditional Rician and Rayleigh fading models cannot accurately depict the channel characteristics and underestimate the channel capacity. More importantly, the proposed channel model outperforms free-space Green's functions in accurately depicting both near-field gain and multi-path effects in radiative near-field regions. These gains are much more evident in tri-polarized systems, highlighting the necessity of polarization interference elimination techniques. Moreover, the theoretical analysis accurately verifies that capacity decreases with expanding communication regions of two-user communications.

cs.IT↗

arXiv2025

VBIM-Net: Variational Born Iterative Network for Inverse Scattering Problems

Recently, studies have shown the potential of integrating field-type iterative methods with deep learning (DL) techniques in solving inverse scattering problems (ISPs). In this article, we propose a novel Variational Born Iterative Network, namely, VBIM-Net, to solve the full-wave ISPs with significantly improved structural rationality and inversion quality. The proposed VBIM-Net emulates the alternating updates of the total electric field and the contrast in the variational Born iterative method (VBIM) by multiple layers of subnetworks. We embed the analytical calculation of the contrast variation into each subnetwork, converting the scattered field residual into an approximate contrast variation and then enhancing it by a U-Net, thus avoiding the requirement of matched measurement dimension and grid resolution as in existing approaches. The total field and contrast of each layer's output is supervised in the loss function of VBIM-Net, imposing soft physical constraints on the variables in the subnetworks, which benefits the model's performance. In addition, we design a training scheme with extra noise to enhance the model's stability. Extensive numerical results on synthetic and experimental data both verify the inversion quality, generalization ability, and robustness of the proposed VBIM-Net. This work may provide some new inspiration for the design of efficient field-type DL schemes.

eess.SP↗

arXiv2025

Mitigating Knowledge Conflicts in Language Model-Driven Question Answering

In the context of knowledge-driven seq-to-seq generation tasks, such as document-based question answering and document summarization systems, two fundamental knowledge sources play crucial roles: the inherent knowledge embedded within model parameters and the external knowledge obtained through context. Recent studies revealed a significant challenge: when there exists a misalignment between the model's inherent knowledge and the ground truth answers in training data, the system may exhibit problematic behaviors during inference, such as ignoring input context, or generating unfaithful content. Our investigation proposes a strategy to minimize hallucination by building explicit connection between source inputs and generated outputs. We specifically target a common hallucination pattern in question answering, examining how the correspondence between entities and their contexts during model training influences the system's performance at inference time.

cs.CL↗

arXiv2024

Observation of nonreciprocal transverse localization of light

Magnetic-free nonreciprocal optical devices that can prevent backscattering of signals are essential for integrated optical information processing. The achieved nonreciprocal behaviors mostly rely on various dispersive effects in optical media, which give rise to dispersive modulations of the transverse beam profile, such as spatial broadening and discretization, of the incident signals. Such deformation inevitably reduces the matching with subsequent components for information processing. Here we experimentally demonstrate the nonreciprocal transverse localization of light in a moiré photonic lattice induced in atomic vapors. When the probe field is set to co- or counter-propagate with the coupling field formed by superposing two identical honeycomb beams in a certain rotation angle, the output pattern can exhibit localized or dispersive behavior. The localization in the forward case is derived from the moiré structure, and the nonreciprocal behaviors (in both beam size and transmitted intensity) are introduced by the thermal motion of atoms. The thermal-motion-induced Doppler effect can destroy the coherent condition for electromagnetically induced transparency in the backward case, because of which the probe beam becomes immune to the modulation of the coupling field. The current work provides an approach to control the transverse beam profile in one-way transmission.

cond-mat.mes-hall↗

arXiv2024

Generative AI Empowered Semantic Feature Multiple Access (SFMA) Over Wireless Networks

This paper investigates a novel generative artificial intelligence (GAI) empowered multi-user semantic communication system called semantic feature multiple access (SFMA) for video transmission, which comprises a base station (BS) and paired users. The BS generates and combines semantic information of several frames simultaneously requested by paired users into a single signal. Users recover their frames from this combined signal and input the recovered frames into a GAI-based video frame interpolation model to generate the intermediate frame. To optimize transmission rates and temporal gaps between simultaneously transmitted frames, we formulate an optimization problem to maximize the system sum rate while minimizing temporal gaps. Since the standard signal-to-interference-plus-noise ratio (SINR) equation does not accurately capture the performance of our semantic communication system, we introduce a weight parameter into the SINR equation to better represent the system's performance. Due to its dependence on transmit power, we propose a three-step solution. First, we develop a user pairing algorithm that pairs two users with the highest preference value, a weighted combination of semantic transmission rate and temporal gap. Second, we optimize inter-group power allocation by formulating an optimization problem that allocates proper transmit power across all user groups to maximize system sum rates while satisfying each user's minimum rate requirement. Third, we address intra-group power allocation to enhance each user's performance. Simulation results demonstrate that our method improves transmission rates by up to 24.8%, 45.8%, and 66.1% compared to fixed-power non-orthogonal multiple access (F-NOMA), orthogonal joint source-channel coding (O-JSCC), and orthogonal frequency division multiple access (OFDMA), respectively.

eess.SP↗

arXiv2024

Research on Key Technologies for Cross-Cloud Federated Training of Large Language Models

With the rapid development of natural language processing technology, large language models have demonstrated exceptional performance in various application scenarios. However, training these models requires significant computational resources and data processing capabilities. Cross-cloud federated training offers a new approach to addressing the resource bottlenecks of a single cloud platform, allowing the computational resources of multiple clouds to collaboratively complete the training tasks of large models. This study analyzes the key technologies of cross-cloud federated training, including data partitioning and distribution, communication optimization, model aggregation algorithms, and the compatibility of heterogeneous cloud platforms. Additionally, the study examines data security and privacy protection strategies in cross-cloud training, particularly the application of data encryption and differential privacy techniques. Through experimental validation, the proposed technical framework demonstrates enhanced training efficiency, ensured data security, and reduced training costs, highlighting the broad application prospects of cross-cloud federated training.

cs.LG↗

arXiv2024

Analysis of Financial Risk Behavior Prediction Using Deep Learning and Big Data Algorithms

As the complexity and dynamism of financial markets continue to grow, traditional financial risk prediction methods increasingly struggle to handle large datasets and intricate behavior patterns. This paper explores the feasibility and effectiveness of using deep learning and big data algorithms for financial risk behavior prediction. First, the application and advantages of deep learning and big data algorithms in the financial field are analyzed. Then, a deep learning-based big data risk prediction framework is designed and experimentally validated on actual financial datasets. The experimental results show that this method significantly improves the accuracy of financial risk behavior prediction and provides valuable support for risk management in financial institutions. Challenges in the application of deep learning are also discussed, along with potential directions for future research.

cs.LG↗