Searcharxiv⌕ Search

arXiv subjects

Xiang Zhang

Publications and source records attributed to Xiang Zhang.

At least 109 records · Page 6Linked to original sources
arXiv2026

PosterGen: Aesthetic-Aware Multi-Modal Paper-to-Poster Generation via Multi-Agent LLMs

Multi-agent systems built upon large language models (LLMs) have demonstrated remarkable capabilities in tackling complex compositional tasks. In this work, we apply this paradigm to the paper-to-poster generation problem, a practical yet time-consuming process faced by researchers preparing for conferences. While recent approaches have attempted to automate this task, most neglect core design and aesthetic principles, resulting in posters that require substantial manual refinement. To address these design limitations, we propose PosterGen, a multi-agent framework that mirrors the workflow of professional poster designers. It consists of four collaborative specialized agents: (1) Parser and Curator agents extract content from the paper and organize storyboard; (2) Layout agent maps the content into a coherent spatial layout; (3) Stylist agents apply visual design elements such as color and typography; and (4) Renderer composes the final poster. Together, these agents produce posters that are both semantically grounded and visually appealing. To evaluate design quality, we introduce a vision-language model (VLM)-based rubric that measures layout balance, readability, and aesthetic coherence. Experimental results show that PosterGen consistently matches in content fidelity, and significantly outperforms existing methods in visual designs, generating posters that are presentation-ready with minimal human refinements.

cs.AI↗
arXiv2026

Generalizable Deepfake Detection Based on Forgery-aware Layer Masking and Multi-artifact Subspace Decomposition

Deepfake detection remains highly challenging, particularly in cross-dataset scenarios and complex real-world settings. This challenge mainly arises because artifact patterns vary substantially across different forgery methods, whereas adapting pretrained models to such artifacts often overemphasizes forgery-specific cues and disturbs semantic representations, thereby weakening generalization. Existing approaches typically rely on full-parameter fine-tuning or auxiliary supervision to improve discrimination. However, they often struggle to model diverse forgery artifacts without compromising pretrained representations. To address these limitations, we propose FMSD, a deepfake detection framework built upon Forgery-aware Layer Masking and Multi-Artifact Subspace Decomposition. Specifically, Forgery-aware Layer Masking evaluates the bias-variance characteristics of layer-wise gradients to identify forgery-sensitive layers, thereby selectively updating them while reducing unnecessary disturbance to pretrained representations. Building upon this, Multi-Artifact Subspace Decomposition further decomposes the selected layer weights via Singular Value Decomposition (SVD) into a semantic subspace and multiple learnable artifact subspaces. These subspaces are optimized to capture heterogeneous and complementary forgery artifacts, enabling effective modeling of diverse forgery patterns while preserving pretrained semantic representations. Furthermore, orthogonality and spectral consistency constraints are imposed to regularize the artifact subspaces, reducing redundancy across them while preserving the overall spectral structure of pretrained weights.

cs.CV↗
arXiv2026

Semantic Manipulation Localization

Image Manipulation Localization (IML) aims to identify edited regions in an image. However, with the increasing use of modern image editing and generative models, many manipulations no longer exhibit obvious low-level artifacts. Instead, they often involve subtle but meaning-altering edits to an object's attributes, state, or relationships while remaining highly consistent with the surrounding content. This makes conventional IML methods less effective because they mainly rely on artifact detection rather than semantic sensitivity. To address this issue, we introduce Semantic Manipulation Localization (SML), a new task that focuses on localizing subtle semantic edits that significantly change image interpretation. We further construct a dedicated fine-grained benchmark for SML using a semantics-driven manipulation pipeline with pixel-level annotations. Based on this task, we propose TRACE (Targeted Reasoning of Attributed Cognitive Edits), an end-to-end framework that models semantic sensitivity through three progressively coupled components: semantic anchoring, semantic perturbation sensing, and semantic-constrained reasoning. Specifically, TRACE first identifies semantically meaningful regions that support image understanding, then injects perturbation-sensitive frequency cues to capture subtle edits under strong visual consistency, and finally verifies candidate regions through joint reasoning over semantic content and semantic scope. Extensive experiments show that TRACE consistently outperforms existing IML methods on our benchmark and produces more complete, compact, and semantically coherent localization results. These results demonstrate the necessity of moving beyond artifact-based localization and provide a new direction for image forensics in complex semantic editing scenarios.

cs.CV↗
arXiv2026

A H.265/HEVC Fine-Grained ROI Video Encryption Algorithm Based on Coding Unit and Prompt Segmentation

ROI (Region of Interest) video selective encryption based on H.265/HEVC is a technology that protects the sensitive regions of videos by perturbing the syntax elements associated with target areas. However, existing methods typically adopt Tile (with a relatively large size) as the minimum encryption unit, which suffers from problems such as inaccurate encryption regions and low encryption precision. This low-precision encryption makes them difficult to apply in sensitive fields such as medicine, military, and remote sensing. In order to address the aforementioned problem, this paper proposes a fine-grained ROI video selective encryption algorithm based on Coding Units (CUs) and prompt segmentation. First, to achieve a more precise ROI acquisition, we present a novel ROI mapping approach based on prompt segmentation. This approach enables precise mapping of ROIs to small $8\times8$ CU levels, significantly enhancing the precision of encrypted regions. Second, we propose a selective encryption scheme based on multiple syntax elements, which distorts syntax elements within high-precision ROI to effectively safeguard ROI security. Finally, we design a diffusion isolation based on Pulse Code Modulation (PCM) mode and MV restriction, applying PCM mode and MV restriction strategy to the affected CU to address encryption diffusion during prediction. The above three strategies break the inherent mechanism of using Tiles in existing ROI encryption and push the fine-grained level of ROI video encryption to the minimum $8\times8$ CU precision. The experimental results demonstrate that the proposed algorithm can accurately segment ROI regions, effectively perturb pixels within these regions, and eliminate the diffusion artifacts introduced by encryption. The method exhibits great potential for application in medical imaging, military surveillance, and remote areas.

eess.IV↗
arXiv2026

LSGS-Loc: Towards Robust 3DGS-Based Visual Localization for Large-Scale UAV Scenarios

Visual localization in large-scale UAV scenarios is a critical capability for autonomous systems, yet it remains challenging due to geometric complexity and environmental variations. While 3D Gaussian Splatting (3DGS) has emerged as a promising scene representation, existing 3DGS-based visual localization methods struggle with robust pose initialization and sensitivity to rendering artifacts in large-scale settings. To address these limitations, we propose LSGS-Loc, a novel visual localization pipeline tailored for large-scale 3DGS scenes. Specifically, we introduce a scale-aware pose initialization strategy that combines scene-agnostic relative pose estimation with explicit 3DGS scale constraints, enabling geometrically grounded localization without scene-specific training. Furthermore, in the pose refinement, to mitigate the impact of reconstruction artifacts such as blur and floaters, we develop a Laplacian-based reliability masking mechanism that guides photometric refinement toward high-quality regions. Extensive experiments on large-scale UAV benchmarks demonstrate that our method achieves state-of-the-art accuracy and robustness for unordered image queries, significantly outperforming existing 3DGS-based approaches. Code is available at: https://github.com/xzhang-z/LSGS-Loc

cs.CV↗
arXiv2026

A Video Steganography for H.265/HEVC Based on Multiple CU Size and Block Structure Distortion

Video steganography based on block structure, which embeds secret information by modifying Coding Unit (CU) block structure of I-frames, is currently a research hotspot. However, the existing algorithms still suffer from the limitation of poor anti-steganalysis, which results from significantly disrupting the original CU block structure after embedding secret information. To overcome this limitation, this paper proposes a video steganography algorithm based on multiple CU size and block structure distortion. Our algorithm introduces three key innovations: 1) a CU Block Structure Stability Metric (CBSSM) based on CU block structure restoration phenomenon to reveal the reasons for the insufficient anti-steganalysis performance of current algorithms. 2) a novel mapping rule based on multiple CU size to reduce block structure change and enhance embedding capacity. 3) a three-level distortion function based on block structure to better guide the secret information embedding. This triple strategy ensures that the secret information embedding minimizes disruption to the original CU block structure while concealing it primarily in areas where block structure changes occur after recompression, ultimately enhancing the algorithm's anti-steganalysis. Comprehensive experimental results highlight the crucial role of the proposed CBSSM in evaluating anti-steganalysis performance even at a low embedding rate. Meanwhile, compared to State-of-the-Art video steganography algorithms based on block structure, our proposed steganography algorithm exhibits greater anti-steganalysis, as well as further improving visual quality, bitrate increase ratio and embedding capacity.

cs.MM↗
arXiv2026

Reducing Subpacketization in Device-to-Device Coded Caching via Heterogeneous File Splitting

The packet type (PT)-based framework~\cite{zhang2026taming} provides a systematic and principled approach to designing device-to-device (D2D) coded caching schemes that achieve reduced \sbp while preserving the optimal communication rate. However, existing PT designs rely exclusively on homogeneous \sbp, where all packets have an identical size regardless of their types. This restriction limits the achievable \sbp reduction in certain parameter regimes. In this paper, we extend the PT framework to \emph{heterogeneous} \sbp, allowing packet sizes to vary across types under a refined type classification. The packet sizes, in conjunction with user grouping and multicast transmitter selection, are jointly optimized to minimize the overall \sbp level while preserving the optimal rate. Based on the heterogeneous PT framework, we construct a new class of D2D coded caching schemes for $(K, KM/N)=(2q+1, 2r)$ with $q,r \in \mathbb{N}_+$, where $K,N$ and $M$ denote the number of users, files and cache memory size, respectively. The proposed construction achieves a constant-factor reduction in \sbp compared to the Ji-Caire-Molisch (JCM) caching scheme~\cite{ji2016fundamental} and complements existing PT designs that are not applicable in this parameter regime.

cs.IT↗
arXiv2026

Information-Theoretic Decentralized Secure Aggregation with Passive Collusion Resilience

In decentralized federated learning (FL), multiple clients collaboratively learn a shared machine learning (ML) model by leveraging their privately held datasets distributed across the network, through interactive exchange of the intermediate model updates. To ensure data security, cryptographic techniques are commonly employed to protect model updates during aggregation. Despite growing interest in secure aggregation, existing works predominantly focus on protocol design and computational guarantees, with limited understanding of the fundamental information-theoretic limits of such systems. Moreover, optimal bounds on communication and key usage remain unknown in decentralized settings, where no central aggregator is available. Motivated by these gaps, we study the problem of decentralized secure aggregation (DSA) from an information-theoretic perspective. Specifically, we consider a network of $K$ fully-connected users, each holding a private input -- an abstraction of local training data -- who aim to securely compute the sum of all inputs. The security constraint requires that no user learns anything beyond the input sum, even when colluding with up to $T$ other users. We characterize the optimal rate region, which specifies the minimum achievable communication and secret key rates for DSA. In particular, we show that to securely compute one symbol of the desired input sum, each user must (i) transmit at least one symbol to others, (ii) hold at least one symbol of secret key, and (iii) all users must collectively hold no fewer than $K - 1$ independent key symbols. Our results establish the fundamental performance limits of DSA, providing insights for the design of provably secure and communication-efficient protocols in decentralized learning.

cs.IT↗
arXiv2026

Information-Theoretic Secure Aggregation in Decentralized Networks

Motivated by the increasing demand for data security in decentralized federated learning (FL) and stochastic optimization, we formulate and investigate the problem of information-theoretic \emph{decentralized secure aggregation} (DSA). Specifically, we consider a network of $K$ interconnected users, each holding a private input, representing, for example, local model updates in FL, who aim to simultaneously compute the sum of all inputs while satisfying the security requirement that no user, even when colluding with up to $T$ others, learns anything beyond the intended sum. We characterize the optimal rate region, which specifies the minimum achievable communication and secret key rates for DSA. In particular, we show that to securely compute one bit of the desired input sum, each user must (i) transmit at least one bit to all other users, (ii) hold at least one bit of secret key, and (iii) all users must collectively hold no fewer than $K - 1$ independent key bits. Our result establishes the fundamental performance limits of DSA and offers insights into the design of provably secure and communication-efficient protocols for distributed learning systems.

cs.IT↗
arXiv2026

Beyond-Ten-Hour Coherence in a Decoherence-Free Trapped-Ion Clock Qubit

Quantum systems promise to revolutionize information processing science and technology [1-3]. The preservation of quantum coherence, the defining property of qubits, fundamentally constrains the performance of quantum information processing with quantum memories [4]. While trapped atomic ions theoretically support million-year coherence based on spontaneous emission [5-7], experimental demonstrations have reached far less, only about an hour [8-13]. Here we combine clock-state qubits with decoherence-free subspace (DFS) encoding to achieve coherence exceeding ten hours. Using correlation-based phase tracking in 171Yb+ ion pairs sympathetically cooled by 138Ba+ ion, we demonstrate this without magnetic shielding or enhanced microwave phase stabilization that previously limited coherence times. DFS encoding references the qubit phase to the inter-ion energy difference to reject microwave phase noise and common-mode magnetic fluctuations, while clock states provide environmental insensitivity. Throughout measurements extended to 1600 seconds, we observe minimal coherence decay, with exponential fits yielding a coherence time of (3.77 +/- 1.09) x 10^4 seconds. Our results establish DFS encoding as a form of passive error correction that eliminates technical noise constraints, unlocking the million-year coherence potential of atomic ions for scalable quantum information processing.

quant-ph↗
arXiv2026

On the Fundamental Limits of Hierarchical Secure Aggregation with Dropout and Collusion Resilience

We study the fundamental communication limits of information-theoretic secure aggregation in a hierarchical network consisting of a server, multiple relays, and multiple users per relay. Communication proceeds over two rounds and two hops, and the system is subject to arbitrary user and relay dropouts. Up to $T$ users may collude with either the server or any single relay. The server aims to recover the sum of the inputs of all users that survive the first round, while learning no additional information beyond the aggregate sum and the inputs of the colluding users. Each relay, however, must learn nothing about the users' inputs except for the information revealed by the inputs of the colluding users under the same collusion model. We introduce a four-dimensional rate tuple that captures the communication cost across rounds and hops. Under a delayed message availability model, we establish necessary and sufficient conditions for feasibility and fully characterize the optimal first-round communication rates. For the second round, we characterize the optimal user-to-relay rate and derive lower and upper bounds on the relay-to-server rate. While these bounds do not coincide in general, they are tight in certain regimes of interest. Our results reveal a sharp threshold phenomenon: secure aggregation is feasible if and only if the total number of surviving users across surviving relays exceeds the collusion threshold. Achievability is established via a vector linear coding scheme with carefully structured correlated randomness exhibiting MDS-like properties, ensuring correctness and information-theoretic security under all possible dropout patterns. Entropic converse bounds are also derived.

cs.IT↗
arXiv2026

CyCLeGen: Cycle-Consistent Layout Prediction and Image Generation in Vision Foundation Models

We present CyCLeGen, a unified vision-language foundation model capable of both image understanding and image generation within a single autoregressive framework. Unlike existing vision models that depend on separate modules for perception and synthesis, CyCLeGen adopts a fully integrated architecture that enforces cycle-consistent learning through image->layout->image and layout->image->layout generation loops. This unified formulation introduces two key advantages: introspection, enabling the model to reason about its own generations, and data efficiency, allowing self-improvement via synthetic supervision under a reinforcement learning objective guided by cycle consistency. Extensive experiments show that CyCLeGen achieves significant gains across diverse image understanding and generation benchmarks, highlighting the potential of unified vision-language foundation models.

cs.CV↗
arXiv2026

Position: Agentic Evolution is the Path to Evolving LLMs

As Large Language Models (LLMs) move from curated training sets into open-ended real-world environments, a fundamental limitation emerges: static training cannot keep pace with continual deployment environment change. Scaling training-time and inference-time compute improves static capability but does not close this train-deploy gap. We argue that addressing this limitation requires a new scaling axis-evolution. Existing deployment-time adaptation methods, whether parametric fine-tuning or heuristic memory accumulation, lack the strategic agency needed to diagnose failures and produce durable improvements. Our position is that agentic evolution represents the inevitable future of LLM adaptation, elevating evolution itself from a fixed pipeline to an autonomous evolver agent. We instantiate this vision in a general framework, A-Evolve, which treats deployment-time improvement as a deliberate, goal-directed optimization process over persistent system state. We further propose the evolution-scaling hypothesis: the capacity for adaptation scales with the compute allocated to evolution, positioning agentic evolution as a scalable path toward sustained, open-ended adaptation in the real world. Our code is publicly available at https://github.com/A-EVO-Lab/a-evolve.

cs.AI↗
arXiv2026

Scalar-induced gravitational waves and primordial black holes from a localized bump or dip feature in a single-field inflationary potential

We study the production of scalar-induced gravitational waves and primordial black holes in a single-field inflation model with a localized bump or dip feature in the potential. Introducing such a localized feature temporarily decelerates the slow-roll inflaton, amplifying the primordial curvature power spectrum into a sharp peak. Consequently, this enhancement sources a significant stochastic background of gravitational waves and leads to abundant formation of primordial black holes. Through eight benchmark cases, we show that the predicted abundances of primordial black holes can remain compatible with current observational limits, while the corresponding gravitational wave spectra peaking across a wide range of frequencies are accessible to future gravitational wave experiments in multiple observational bands.

astro-ph.CO↗
arXiv2026

HarmonyCell: Automating Single-Cell Perturbation Modeling under Semantic and Distribution Shifts

Single-cell perturbation studies face dual heterogeneity bottlenecks: (i) semantic heterogeneity--identical biological concepts encoded under incompatible metadata schemas across datasets; and (ii) statistical heterogeneity--distribution shifts from biological variation demanding dataset-specific inductive biases. We propose HarmonyCell, an end-to-end agent framework resolving each challenge through a dedicated mechanism: an LLM-driven Semantic Unifier autonomously maps disparate metadata into a canonical interface without manual intervention; and an adaptive Monte Carlo Tree Search engine operates over a hierarchical action space to synthesize architectures with optimal statistical inductive biases for distribution shifts. Evaluated across diverse perturbation tasks under both semantic and distribution shifts, HarmonyCell achieves a 95% valid execution rate on heterogeneous input datasets (versus 0% for general agents) while matching or even exceeding expert-designed baselines in rigorous out-of-distribution evaluations. This dual-track orchestration enables scalable automatic virtual cell modeling without dataset-specific engineering.

cs.AI↗
arXiv2026

PixARMesh: Autoregressive Mesh-Native Single-View Scene Reconstruction

We introduce PixARMesh, a method to autoregressively reconstruct complete 3D indoor scene meshes directly from a single RGB image. Unlike prior methods that rely on implicit signed distance fields and post-hoc layout optimization, PixARMesh jointly predicts object layout and geometry within a unified model, producing coherent and artist-ready meshes in a single forward pass. Building on recent advances in mesh generative models, we augment a point-cloud encoder with pixel-aligned image features and global scene context via cross-attention, enabling accurate spatial reasoning from a single image. Scenes are generated autoregressively from a unified token stream containing context, pose, and mesh, yielding compact meshes with high-fidelity geometry. Experiments on synthetic and real-world datasets show that PixARMesh achieves state-of-the-art reconstruction quality while producing lightweight, high-quality meshes ready for downstream applications.

cs.CV↗
arXiv2026

You Only Need One Stage: Novel-View Synthesis From A Single Blind Face Image

We propose a novel one-stage method, NVB-Face, for generating consistent Novel-View images directly from a single Blind Face image. Existing approaches to novel-view synthesis for objects or faces typically require a high-resolution RGB image as input. When dealing with degraded images, the conventional pipeline follows a two-stage process: first restoring the image to high resolution, then synthesizing novel views from the restored result. However, this approach is highly dependent on the quality of the restored image, often leading to inaccuracies and inconsistencies in the final output. To address this limitation, we extract single-view features directly from the blind face image and introduce a feature manipulator that transforms these features into 3D-aware, multi-view latent representations. Leveraging the powerful generative capacity of a diffusion model, our framework synthesizes high-quality, consistent novel-view face images. Experimental results show that our method significantly outperforms traditional two-stage approaches in both consistency and fidelity.

cs.CV↗
arXiv2026

Image Corruption-Inspired Membership Inference Attacks against Large Vision-Language Models

Large vision-language models (LVLMs) have demonstrated outstanding performance in many downstream tasks. However, LVLMs are trained on large-scale datasets, which can pose privacy risks if training images contain sensitive information. Therefore, it is important to detect whether an image is used to train the LVLM. Recent studies have investigated membership inference attacks (MIAs) against LVLMs, including detecting image-text pairs and single-modality content. In this work, we focus on detecting whether a target image is used to train the target LVLM. We design simple yet effective Image Corruption-Inspired Membership Inference Attacks (ICIMIA) against LVLMs, which are inspired by LVLM's different sensitivity to image corruption for member and non-member images. We first perform an MIA method under the white-box setting, where we can obtain the embeddings of the image through the vision part of the target LVLM. The attacks are based on the embedding similarity between the image and its corrupted version. We further explore a more practical scenario where we have no knowledge about target LVLMs and we can only query the target LVLMs with an image and a textual instruction. We then conduct the attack by utilizing the output text embeddings' similarity. Experiments on existing datasets validate the effectiveness of our proposed methods under those two different settings.

cs.CV↗