arXiv subjects

Jie Zhang

Publications and source records attributed to Jie Zhang.

At least 235 records · Page 13Linked to original sources

arXiv2026

TRACE: Structure-Aware Character Encoding for Robust and Generalizable Document Watermarking

We propose TRACE, a structure-aware framework leveraging diffusion models for localized character encoding to embed data. Unlike existing methods that rely on edge features or pre-defined codebooks, TRACE exploits character structures that provide inherent resistance to noise interference due to their stability and unified representation across diverse characters. Our framework comprises three key components: (1) adaptive diffusion initialization that automatically identifies handle points, target points, and editing regions through specialized algorithms including movement probability estimator (MPE), target point estimation (TPE) and mask drawing model (MDM), (2) guided diffusion encoding for precise movement of selected point, and (3) masked region replacement with a specialized loss function to minimize feature alterations after the diffusion process. Comprehensive experiments demonstrate \name{}'s superior performance over state-of-the-art methods, achieving more than 5 dB improvement in PSNR and 5\% higher extraction accuracy following cross-media transmission. \name{} achieves broad generalizability across multiple languages and fonts, making it particularly suitable for practical document security applications.

cs.CV↗

arXiv2026

INFACT: A Diagnostic Benchmark for Induced Faithfulness and Factuality Hallucinations in Video-LLMs

Despite rapid progress, Video Large Language Models (Video-LLMs) remain unreliable due to hallucinations, which are outputs that contradict either video evidence (faithfulness) or verifiable world knowledge (factuality). Existing benchmarks provide limited coverage of factuality hallucinations and predominantly evaluate models only in clean settings. We introduce \textsc{INFACT}, a diagnostic benchmark comprising 9{,}800 QA instances with fine-grained taxonomies for faithfulness and factuality, spanning real and synthetic videos. \textsc{INFACT} evaluates models in four modes: Base (clean), Visual Degradation, Evidence Corruption, and Temporal Intervention for order-sensitive items. Reliability under induced modes is quantified using Resist Rate (RR) and Temporal Sensitivity Score (TSS). Experiments on 14 representative Video-LLMs reveal that higher Base-mode accuracy does not reliably translate to higher reliability in the induced modes, with evidence corruption reducing stability and temporal intervention yielding the largest degradation. Notably, many open-source baselines exhibit near-zero TSS on factuality, indicating pronounced temporal inertia on order-sensitive questions.

cs.CV↗

arXiv2026

Recover to Predict: Progressive Retrospective Learning for Variable-Length Trajectory Prediction

Trajectory prediction is critical for autonomous driving, enabling safe and efficient planning in dense, dynamic traffic. Most existing methods optimize prediction accuracy under fixed-length observations. However, real-world driving often yields variable-length, incomplete observations, posing a challenge to these methods. A common strategy is to directly map features from incomplete observations to those from complete ones. This one-shot mapping, however, struggles to learn accurate representations for short trajectories due to significant information gaps. To address this issue, we propose a Progressive Retrospective Framework (PRF), which gradually aligns features from incomplete observations with those from complete ones via a cascade of retrospective units. Each unit consists of a Retrospective Distillation Module (RDM) and a Retrospective Prediction Module (RPM), where RDM distills features and RPM recovers previous timesteps using the distilled features. Moreover, we propose a Rolling-Start Training Strategy (RSTS) that enhances data efficiency during PRF training. PRF is plug-and-play with existing methods. Extensive experiments on datasets Argoverse 2 and Argoverse 1 demonstrate the effectiveness of PRF. Code is available at https://github.com/zhouhao94/PRF.

cs.RO↗

arXiv2026

V-Attack: Targeting Disentangled Value Features for Controllable Adversarial Attacks on LVLMs

Adversarial attacks have evolved from simply disrupting predictions on conventional task-specific models to the more complex goal of manipulating image semantics on Large Vision-Language Models (LVLMs). However, existing methods struggle with controllability and fail to precisely manipulate the semantics of specific concepts in the image. We attribute this limitation to semantic entanglement in the patch-token representations on which adversarial attacks typically operate: global context aggregated by self-attention in the vision encoder dominates individual patch features, making them unreliable handles for precise local semantic manipulation. Our systematic investigation reveals a key insight: value features (V) computed within the transformer attention block serve as much more precise handles for manipulation. We show that V suppresses global-context channels, allowing it to retain high-entropy, disentangled local semantic information. Building on this discovery, we propose V-Attack, a novel method designed for precise local semantic attacks. V-Attack targets the value features and introduces two core components: (1) a Self-Value Enhancement module to refine V's intrinsic semantic richness, and (2) a Text-Guided Value Manipulation module that leverages text prompts to locate source concept and optimize it toward a target concept. By bypassing the entangled patch features, V-Attack achieves highly effective semantic control. Extensive experiments across diverse LVLMs, including LLaVA, InternVL, DeepseekVL and GPT-4o, show that V-Attack improves the attack success rate by an average of 36% over state-of-the-art methods, exposing critical vulnerabilities in modern visual-language understanding. Our code and data are available https://github.com/Summu77/V-Attack.

cs.CV↗

arXiv2026

Three-stage melting of a macroscopic continuous spacetime crystal

A spacetime crystal is a phase of matter that spontaneously develops periodic order in both space and time. Spacetime crystals have been experimentally observed in microscopic quantum many-body systems and, very recently, in a mesoscopic nematic liquid crystal. However, the melting process of a spacetime crystal and its underlying physical mechanisms have not yet been experimentally reported. Here, we present a direct observation of a classical continuous spacetime crystal melting in a table-top experiment with macroscopic active granular disks in 2+1 spacetime dimensions. The spacetime crystal is characterized by the spontaneous formation of a coherent, rigid-body rotation of a 2D triangular lattice that persists for almost a day and remains remarkably robust to noise. By tuning the disk packing fraction, we observe a complex three-stage melting process involving a spatially hexatic phase and multiple coexistence regions. Importantly, we show that spatial and temporal crystalline orders melt separately through distinct mechanisms: spatial order is destroyed by the proliferation of topological defects, while temporal order is lost through the decay of directional persistence caused by the progressive weakening of many-body interactions. Our results demonstrate that the spontaneous breaking of spatial and temporal translational symmetries can be decoupled, leading to the emergence of exotic out-of-equilibrium classical phases of matter.

cond-mat.soft↗

arXiv2026

Dynamic Stability Assessment of Grid-Connected Data Centers Powered by Small Modular Reactors

The accelerating growth of computational demand in modern data centers has further heightened the need for power infrastructures that are highly reliable, environmentally sustainable, and capable of supporting grid stability. Small Modular Reactors (SMRs) as a clean source of energy are particularly attractive for next-generation hyperscale data centers with significant electrical and cooling demands. This paper presents a comprehensive dynamic modeling and stability analysis of a grid-connected Integrated Energy System (IES) designed for data center applications. The proposed IES integrates an SMR and a battery energy storage system to jointly supply electricity for computational and cooling load while providing stability support to the main grid. A coupled computational-thermal load model is developed to capture the real-time power demand of the data center, incorporating CPU utilization, cooling efficiency, and ambient temperature effects. The integrated SMR-powered data center model is implemented in PSSE and tested on the IEEE 118-bus system under various fault scenarios. Simulation results demonstrate that the IES substantially enhances voltage and frequency stability compared to a conventionally grid-connected data center, minimizing disturbance-induced deviations and improving post-fault recovery.

eess.SY↗

arXiv2026

Torque Hyperuniformity in Frictional Granular Matter - Theory and Experiments

A question of some fundamental importance is whether a given assembly of frictional granules (say sand or powder) will exhibit stress autocorrelations with long-range anisotropic decay as determined by the elastic Green's function. In Hamiltonian systems with central forces, mechanical balance and material isotropy demand the stress auto-correlation matrix to be fully determined by the pressure auto-correlation only. If the local pressure fluctuations are normal, it follows that stress autocorrelations decay at long distance like the elastic Green's function. With friction, Hamiltonian symmetry is lost, and one may expect more constraints. Indeed, it was shown recently that for frictional amorphous solids mechanical balance and material isotropy demand the stress auto-correlation matrix to be fully determined by two spatially isotropic functions: the pressure and torque auto-correlations. Elastic-like decay of the stress autocorrelations follows from normal fluctuations of the pressure and from the torque fluctuations being hyperuniform. The theoretical discovery of these conditions required experimental confirmation, to test whether these conditions are generically obeyed in actual frictional amorphous solids. Recently the confirmation was announced for 2-dimensional amorphous assemblies of frictional disks under isotropic load, in which torque is caused by tangential forces only. In this paper we review that case and report confirmation of the theoretical predictions in 2-dimensional systems of disks under shear and in isotropically loaded frictional ellipses, where contributions to torque come also from normal forces. The paper ends with physical explanations of the hyperuniformity of the torque fluctuations and predictions for how the results are expected to extend to d-dimensions.

cond-mat.soft↗

arXiv2026

Bell Plesset Effects on Rayleigh Taylor Instability of Three Dimensional Spherical Geometry

We develop a weakly nonlinear, multi-mode theory for the Rayleigh-Taylor instability (RTI) on a time-varying spherical interface, fully incorporating mode couplings and the Bell-Plesset (BP) effects arising from interface convergence. Our model extends prior analyses, which have been largely restricted to static backgrounds, 2D cylindrical geometries, or single-mode initial condition. We present a framework capable of evolving arbitrary, fully three-dimensional initial perturbations on a dynamic background. At the first order, mode amplitudes respond to the time-varying interface acceleration with an exponential-like growth, in qualitative agreement with classic static results. At second order, nonlinear mode coupling reveals a powerful selection rule: energy is preferentially channeled into axisymmetric (m=0) modes. We find that the BP effects dramatically amplify the instability growth by a few orders of magnitude, with this amplification being even more significant for second order couplings. Despite this strong channeling, the second order amplitudes remain small relative to the first order, validating the perturbative approach. These findings offer new physical insights into time-dependent interface instabilities relevant to applications such as astrophysical shell collapse and inertial confinement fusion, highlighting the uniquely dominant role of axisymmetric modes in BP-driven convergent flows.

physics.plasm-ph↗

arXiv2026

A pair of oblate bubbles rising in-line: a linear stability analysis

The stability of two bubbles rising initially in-line through a viscous liquid is revisited using a global linear stability analysis formulated within an Arbitrary Lagrangian-Eulerian framework, complemented by fully resolved Embedded Boundary Method simulations. Whereas previous studies attributed the promoted in-line stability of oblate bubbles to a deformation-enhanced wake entrainment, the present analysis demonstrates that the dominant stabilizing mechanism arises instead from an inclination-induced rotational feedback generated as the trailing bubble experiences the asymmetric shear of the leading bubble's wake. This inclination-shear coupling, rather than deformation itself, governs the recovery of stability with increasing aspect ratio. Furthermore, the results reveal that the unstable drafting-kissing-tumbling mode originates from short-range, two-way coupling between the bubbles, whereas the asymmetric side-escape mode corresponds to a long-range, one-way interaction dominated by the trailing bubble response. In addition, a previously unreported oscillatory global mode emerges from the unsteady recirculation linking the two bubbles, acting as a hydrodynamic spring whose effective stiffness and damping govern the oscillation frequency and growth rate. Together, these findings identify inclination-induced lift as the primary mechanism controlling the stability of rising bubble pairs and provide a unified framework for interpreting their stationary and oscillatory transitions across a broad range of inertial and deformable regimes.

physics.flu-dyn↗

arXiv2026

A 3D sharp and conservative VOF method for modeling the contact line dynamics with hysteresis on complex boundaries

We propose a sharp and conservative 3D numerical method for simulating moving contact lines on complex geometries, developed within a coupled geometric Volume-of-Fluid (VOF) and embedded boundary framework. The first major contribution is a modified VOF advection and reconstruction scheme specifically designed for mixed cells containing liquid, gas, and solid phases. This formulation ensures strict local mass conservation in the presence of arbitrarily shaped embedded boundaries. To overcome the severe time-step limitation caused by small cut cells, a redistribution advection strategy is introduced, which completely removes the CFL constraint while preserving both local and global volume conservation. The second key contribution is a novel 3D contact angle imposition technique built upon the height function framework. By incorporating a pre-fitting paraboloid procedure, the method achieves robust curvature estimation and accurate enforcement of contact angle conditions on irregular solid surfaces. In addition, contact angle hysteresis is modeled to capture more realistic wetting dynamics. A series of challenging benchmark tests have been conducted to demonstrate the accuracy, robustness, and superiority of the proposed method compared with existing sharp-interface approaches. This study, for the first time, establishes a fully geometric and conservative VOF-based scheme capable of accurately resolving contact line dynamics on arbitrarily complex 3D surfaces.

math.NA↗

arXiv2026

Complementarity-Supervised Spectral-Band Routing for Multimodal Emotion Recognition

Multimodal emotion recognition fuses cues such as text, video, and audio to understand individual emotional states. Prior methods face two main limitations: mechanically relying on independent unimodal performance, thereby missing genuine complementary contributions, and coarse-grained fusion conflicting with the fine-grained representations required by emotion tasks. As inconsistent information density across heterogeneous modalities hinders inter-modal feature mining, we propose the Complementarity-Supervised Multi-Band Expert Network, named Atsuko, to model fine-grained complementary features via multi-scale band decomposition and expert collaboration. Specifically, we orthogonally decompose each modality's features into high, mid, and low-frequency components. Building upon this band-level routing, we design a modality-level router with a dual-path mechanism for fine-grained cross-band selection and cross-modal fusion. To mitigate shortcut learning from dominant modalities, we propose the Marginal Complementarity Module (MCM) to quantify performance loss when removing each modality via bi-modal comparison. The resulting complementarity distribution provides soft supervision, guiding the router to focus on modalities contributing unique information gains. Extensive experiments show our method achieves superior performance on the CMU-MOSI, CMU-MOSEI, CH-SIMS, CH-SIMSv2, and MIntRec benchmarks.

cs.CV↗

arXiv2026

A SISA-based Machine Unlearning Framework for Power Transformer Inter-Turn Short-Circuit Fault Localization

In practical data-driven applications on electrical equipment fault diagnosis, training data can be poisoned by sensor failures, which can severely degrade the performance of machine learning (ML) models. However, once the ML model has been trained, removing the influence of such harmful data is challenging, as full retraining is both computationally intensive and time-consuming. To address this challenge, this paper proposes a SISA (Sharded, Isolated, Sliced, and Aggregated)-based machine unlearning (MU) framework for power transformer inter-turn short-circuit fault (ITSCF) localization. The SISA method partitions the training data into shards and slices, ensuring that the influence of each data point is isolated within specific constituent models through independent training. When poisoned data are detected, only the affected shards are retrained, avoiding retraining the entire model from scratch. Experiments on simulated ITSCF conditions demonstrate that the proposed framework achieves almost identical diagnostic accuracy to full retraining, while reducing retraining time significantly.

eess.SY↗

arXiv2026

Topology-Aware Reinforcement Learning over Graphs for Resilient Power Distribution Networks

Extreme weather events and cyberattacks can cause component failures and disrupt the operation of power distribution networks (DNs), during which reconfiguration and load shedding are often adopted for resilience enhancement. This study introduces a topology-aware graph reinforcement learning (RL) framework for outage management that embeds higher-order topological features of the DN into a graph-based RL model, enabling reconfiguration and load shedding to maximize energy supply while maintaining operational stability. Results on the modified IEEE 123-bus feeder across 300 diverse outage scenarios demonstrate that incorporating the topological data analysis (TDA) tool, persistence homology (PH), yields 9-18% higher cumulative rewards, up to 6% increase in power delivery, and 6-8% fewer voltage violations compared to a baseline graph-RL model. These findings highlight the potential of integrating RL with TDA to enable self-healing in DNs, facilitating fast, adaptive, and automated restoration.

eess.SY↗

arXiv2026

Reinforcement Learning for Vehicle-to-Grid Voltage Regulation: Single-Hub to Multi-Hub Coordination with Battery-Aware Constraints

This paper presents a Vehicle-to-Grid (V2G) coordination framework using reinforcement learning (RL). {An intelligent control strategy based on the soft actor-critic algorithm is developed for voltage regulation through single and multi-hub charging systems while respecting realistic fleet constraints. A two-phase training approach integrates stability-focused learning with battery-aware deployment to ensure practical feasibility. Simulation studies on the IEEE 34-bus system validate the framework against a standard Volt-Var/Volt-Watt droop controller. Results indicate that the RL agent achieves performance comparable to the baseline control strategy in nominal scenarios. Under aggressive overloading, it provides robust voltage recovery (within 10% of the baseline) while prioritizing fleet availability and state-of-charge preservation, demonstrating the viability of constraint-aware learning for critical grid services.}

eess.SY↗

arXiv2026

Cross-Scale Pansharpening via ScaleFormer and the PanScale Benchmark

Pansharpening aims to generate high-resolution multi-spectral images by fusing the spatial detail of panchromatic images with the spectral richness of low-resolution MS data. However, most existing methods are evaluated under limited, low-resolution settings, limiting their generalization to real-world, high-resolution scenarios. To bridge this gap, we systematically investigate the data, algorithmic, and computational challenges of cross-scale pansharpening. We first introduce PanScale, the first large-scale, cross-scale pansharpening dataset, accompanied by PanScale-Bench, a comprehensive benchmark for evaluating generalization across varying resolutions and scales. To realize scale generalization, we propose ScaleFormer, a novel architecture designed for multi-scale pansharpening. ScaleFormer reframes generalization across image resolutions as generalization across sequence lengths: it tokenizes images into patch sequences of the same resolution but variable length proportional to image scale. A Scale-Aware Patchify module enables training for such variations from fixed-size crops. ScaleFormer then decouples intra-patch spatial feature learning from inter-patch sequential dependency modeling, incorporating Rotary Positional Encoding to enhance extrapolation to unseen scales. Extensive experiments show that our approach outperforms SOTA methods in fusion quality and cross-scale generalization. The datasets and source code are available at https://github.com/caoke-963/ScaleFormer.

cs.CV↗

arXiv2026

SURE: Semi-dense Uncertainty-REfined Feature Matching

Establishing reliable image correspondences is essential for many robotic vision problems. However, existing methods often struggle in challenging scenarios with large viewpoint changes or textureless regions, where incorrect cor- respondences may still receive high similarity scores. This is mainly because conventional models rely solely on fea- ture similarity, lacking an explicit mechanism to estimate the reliability of predicted matches, leading to overconfident errors. To address this issue, we propose SURE, a Semi- dense Uncertainty-REfined matching framework that jointly predicts correspondences and their confidence by modeling both aleatoric and epistemic uncertainties. Our approach in- troduces a novel evidential head for trustworthy coordinate regression, along with a lightweight spatial fusion module that enhances local feature precision with minimal overhead. We evaluated our method on multiple standard benchmarks, where it consistently outperforms existing state-of-the-art semi-dense matching models in both accuracy and efficiency. our code will be available on https://github.com/LSC-ALAN/SURE.

cs.CV↗

arXiv2026

When Memory Becomes a Vulnerability: Towards Multi-turn Jailbreak Attacks against Text-to-Image Generation Systems

Modern text-to-image (T2I) generation systems (e.g., DALL$\cdot$E 3) exploit the memory mechanism, which captures key information in multi-turn interactions for faithful generation. Despite its practicality, the security analyses of this mechanism have fallen far behind. In this paper, we reveal that it can exacerbate the risk of jailbreak attacks. Previous attacks fuse the unsafe target prompt into one ultimate adversarial prompt, which can be easily detected or lead to the generation of non-unsafe images due to under- or over-detoxification. In contrast, we propose embedding the malice at the inception of the chat session in memory, addressing the above limitations. Specifically, we propose Inception, the first multi-turn jailbreak attack against real-world text-to-image generation systems that explicitly exploits their memory mechanisms. Inception is composed of two key modules: segmentation and recursion. We introduce Segmentation, a semantic-preserving method that generates multi-round prompts. By leveraging NLP analysis techniques, we design policies to decompose a prompt, together with its malicious intent, according to sentence structure, thereby evading safety filters. Recursion further addresses the challenge posed by unsafe sub-prompts that cannot be separated through simple segmentation. It firstly expands the sub-prompt, then invokes segmentation recursively. To facilitate multi-turn adversarial prompts crafting, we build VisionFlow, an emulation T2I system that integrates two-stage safety filters and industrial-grade memory mechanisms. The experiment results show that Inception successfully allures unsafe image generation, surpassing the SOTA by a 20.0\% margin in attack success rate. We also conduct experiments on the real-world commercial T2I generation platforms, further validating the threats of Inception in practice.

cs.CV↗

arXiv2026

Towards Generalized Multimodal Homography Estimation

Supervised and unsupervised homography estimation methods depend on image pairs tailored to specific modalities to achieve high accuracy. However, their performance deteriorates substantially when applied to unseen modalities. To address this issue, we propose a training data synthesis method that generates unaligned image pairs with ground-truth offsets from a single input image. Our approach renders the image pairs with diverse textures and colors while preserving their structural information. These synthetic data empower the trained model to achieve greater robustness and improved generalization across various domains. Additionally, we design a network to fully leverage cross-scale information and decouple color information from feature representations, thus improving estimation accuracy. Extensive experiments show that our training data synthesis method improves generalization performance. The results also confirm the effectiveness of the proposed network.

cs.CV↗