Search arXiv⌕ Search

arXiv subjects

Hua Zhang

Publications and source records attributed to Hua Zhang.

At least 19 recordsLinked to original sources
arXiv2026

Derivative Formulae and Gradient Estimates for Stochastic Hamiltonian Systems with Jumps

This paper is concerned with stochastic Hamiltonian systems driven by Lévy processes. By employing the lent particle method developed by Bouleau and Denis in the framework of Malliavin calculus with jumps, we establish an explicit Bismut--Elworthy--Li type derivative formula for the associated Markov semigroup, as well as corresponding gradient estimates. The main novelty lies in extending the known results for Brownian-motion-driven degenerate systems to the nonlocal setting with jumps, where the presence of a pure jump noise and the degeneracy of the coefficients pose substantial difficulties. Our approach provides a systematic treatment of such nonlocal degenerate operators and fills a gap in the literature on derivative formulae for jump-type stochastic Hamiltonian systems.

math.PR↗
arXiv2026

Phantom-Insight: Adaptive Multi-cue Fusion for Video Camouflaged Object Detection with Multimodal LLM

Video camouflaged object detection (VCOD) is challenging due to dynamic environments. Existing methods face two main issues: (1) SAM-based methods struggle to separate camouflaged object edges due to model freezing, and (2) MLLM-based methods suffer from poor object separability as large language models merge foreground and background. To address these issues, we propose a novel VCOD method based on SAM and MLLM, called Phantom-Insight. To enhance the separability of object edge details, we represent video sequences with temporal and spatial clues and perform feature fusion via LLM to increase information density. Next, multiple cues are generated through the dynamic foreground visual token scoring module and the prompt network to adaptively guide and fine-tune the SAM model, enabling it to adapt to subtle textures. To enhance the separability of objects and background, we propose a decoupled foreground-background learning strategy. By generating foreground and background cues separately and performing decoupled training, the visual token can effectively integrate foreground and background information independently, enabling SAM to more accurately segment camouflaged objects in the video. Experiments on the MoCA-Mask dataset show that Phantom-Insight achieves state-of-the-art performance across various metrics. Additionally, its ability to detect unseen camouflaged objects on the CAD2016 dataset highlights its strong generalization ability.

cs.CV↗
arXiv2026

Where Not to Learn: Prior-Aligned Training with Subset-based Attribution Constraints

Reliable models should not only predict correctly, but also base their decisions on acceptable evidence. However, conventional supervised learning typically provides only class-level labels, allowing models to achieve high accuracy by exploiting shortcut correlations rather than intended decision evidence. Human priors, such as bounding boxes or target interface elements, can help constrain such behavior, but aligning model evidence with these priors remains challenging because learned decision evidence often diverges from human perception. In this work, we study attribution-prior alignment with subset-selection-based attribution. Motivated by prior deletion and insertion evaluations showing that subset-selection attribution can identify compact decision-supporting regions, we use it as a training-time signal to expose the model's attributed evidence. When the top-attributed evidence deviates substantially from the prior region, we penalize off-prior attribution and encourage the model to shift its attributed evidence toward the intended regions. This yields a selective prior-constrained objective that avoids uniformly suppressing all non-prior regions. We validate our method on both image classification and click decision tasks in MLLM-based GUI agents. Across discriminative classification and autoregressive decision-making settings, our method improves task accuracy while enhancing attribution-prior alignment.

cs.CV↗
arXiv2026

FedADB: Class Anchor-Driven Dual-Branch Federated Learning for Mitigating Forgetting

Multimodal data collected by heterogeneous devices are used for collaborative training, where federated learning (FL) serves as a key paradigm for effective distributed modeling with data privacy preservation. However, local training suffers from the forgetting of previously learned global knowledge under cross-client data heterogeneity, which leads to significant declines in both performance and convergence speed. Most previous studies rely on global alignment strategies to retain global knowledge, which hinder local optimization and lead to inadequate supervision of missing classes. Some studies introduce proxy datasets to supplement supervision for missing classes. However, it remains a challenge to balance class-wise global consistency and local optimization objectives without proxy datasets. In this work, we propose FedADB, a Class Anchor-Driven Dual-Branch FL framework. Specifically, the server generates class anchors optimized in a differentiable input space, which are shared across clients. These class anchors serve as global references that provide supervision for missing classes during local training. A dual-branch collaborative training mechanism is designed for clients. In this mechanism, the anchor-based global branch focuses on learning with global consistency, achieving global knowledge alignment by class-anchor balanced sampling. The local calibration branch focuses on learning discriminative local features, mitigating the degradation of local representations caused by excessive global alignment. Extensive experiments across multiple medical and natural datasets demonstrate that FedADB achieves significant improvements in both accuracy and convergence speed.

cs.CR↗
arXiv2026

GARDRec: Decision-Level Graph Grounding for Large Language Model Recommendation

Large language models (LLMs) offer new opportunities for recommendation by interpreting item descriptions, user instructions, and external knowledge through natural-language prompts. However, existing graph-augmented LLM recommenders often use knowledge graphs mainly as prompt-level evidence, leaving ranking decisions weakly constrained by structured user-item relations. This is problematic for next-item recommendation, where the model must compare candidates under the same user context while preserving temporal preference, collaborative signals, and attribute matches. To address this issue, we propose \emph{GARDRec}, a Graph-grounded Adaptive Reasoning and Decision-aware Recommendation framework for LLM-based next-item ranking. GARDRec constructs semantic-structural item representations from textual node features and graph propagation, derives personalized graph contexts from temporally weighted histories and first-order neighborhoods, and aligns graph-derived representations with a frozen LLM through continuous multimodal prompts. Explicit interaction and matching features are injected through late-stage decision branches, while inter-candidate attention and restricted generative likelihood support final ranking. Experiments on three public benchmarks with multiple LLM backbones show that GARDRec generally improves candidate-ranking performance over representative baselines. Ablation and diagnostic analyses verify the contributions of graph projection, neighborhood retrieval, explicit decision features, ranking loss, and generative calibration.

cs.IR↗
arXiv2026

Tamed Euler Schemes for Singular SDEs with Multiplicative Levy Noise

We prove strong convergence rates for tamed Euler schemes of multidimensional stochastic differential equations with singular drift, multiplicative Brownian noise, and multiplicative Levy noise. The drift is assumed to satisfy a Ladyzhenskaya-Prodi-Serrin type condition, while the diffusion and jump coefficients have Sobolev-type spatial regularity. The proof is based on a nonlocal Zvonkin transform, Krylov-type estimates, and a stochastic Gronwall argument. The main novelty is an error decomposition that isolates two jump-induced contributions: a compensated jump martingale error and a nonlocal compensator error caused by Euler freezing. This decomposition yields explicit rates and recovers the Brownian-type estimate when the jump coefficient vanishes.

math.PR↗
arXiv2026

PersGuard: Preventing Malicious Personalization in Text-to-Image Diffusion Models via Model Backdoors

Diffusion models (DMs) have advanced text-to-image (T2I) synthesis, yet their personalization capabilities raise serious privacy and copyright concerns. Malicious actors can misuse these models to generate unauthorized portraits or artistic style replicas. Existing proactive defenses primarily rely on applying adversarial perturbations to reference images to disrupt training. However, these approaches face limitations: they assume all training images are pre-perturbed and are prone to failure when datasets contain unperturbed images or undergo minor data transformations. In this paper, we introduce PersGuard, a novel backdoor-based framework designed to prevent unauthorized personalization of pre-trained T2I diffusion models. Unlike perturbation-based methods, we assume protectors can embed protective backdoors into the models before their release. This mechanism ensures that if a downstream user fine-tunes the model on protected images, the model retains the backdoor and generates predefined protective outputs; conversely, for unprotected images, the backdoor is effectively removed during fine-tuning to ensure normal model utility. We formulate the backdoor injection as a unified optimization problem incorporating three objectives: a backdoor behavior loss to activate protection, a prior preservation loss to maintain standard generation capabilities, and a novel backdoor retention loss. The retention loss is specifically designed to mirror personalization loss, ensuring the backdoor remains robust during downstream fine-tuning. Extensive experiments across gray-box and black-box settings, multi-object protection, and facial identity protection demonstrate that PersGuard provides superior privacy protection compared to existing perturbation-based methods.

cs.CV↗
arXiv2026

A Good Initialization is All You Need for Faithful Visual Attribution

Faithful visual attribution identifies which image regions support a model prediction. Search-based perturbation methods lead the insertion--deletion faithfulness frontier by masking regions and measuring score changes, but they usually output a complete ordering of all regions. Many applications, especially MLLM attribution and repair, only need a compact top-\(k\) evidence mask. We study this mask-first attribution problem. An exactly \(k\)-region mask is combinatorial: useful evidence can depend on interactions among fine regions. Coarse grouping can stabilize early search but aggregates redundant content, whereas one-step scoring can miss high-value combinations. We introduce two forward-only methods. \textsc{CoPAIR} uses a PhaseWin--Greedy gap diagnosis to construct coarse singleton/pair candidates that warm-start full-ordering search. \textsc{TRACE} directly searches fixed-cardinality fine-region masks with cross-entropy sampling, elite retention, and distribution updates, with a finite-budget recovery analysis. The resulting evidence set can be returned as a compact attribution mask or used to initialize Greedy or PhaseWin when a complete ranking is required. Across ImageNet classification with CLIP ViT-L/14, CLIP RN101, and ResNet-101, our initialized search methods establish a new state-of-the-art frontier for faithful full-ordering attribution under inclusive forward-call accounting. On POPE and RePOPE with Qwen2.5-VL-3B-Instruct and LLaVA-v1.5-7B, \textsc{TRACE}+Greedy gives the strongest search-based MLLM attribution results. Direct \textsc{TRACE} masks further achieve single-point RePOPE repair rates of \(94.44\%\) and \(96.00\%\), showing that compact evidence masks can be actionable attribution outputs, not merely prefixes of full rankings.

cs.CV↗
arXiv2026

PhaseWin: An Efficient Search Algorithm for Faithful Visual Attribution

Visual attribution is a fundamental tool for interpreting modern vision and vision-language models, particularly when their decisions must be inspected, diagnosed, or audited. Its goal is to explain how a model's decision depends on local regions of the visual input, typically by assigning an importance ordering over candidate image regions. Given an image partitioned into $n$ regions, faithful attribution can be cast as an ordered subset-search problem, in which progressively inserting the selected regions should recover the target model response as early as possible. Exhaustive search over region subsets incurs exponential cost, while the widely used greedy search still requires a quadratic number of model evaluations, because every selection step rescores all remaining candidates. We propose PhaseWin, an efficient subset-search algorithm for faithful visual attribution. PhaseWin reorganizes greedy region selection into a phased window-search procedure: rather than re-evaluating the full candidate set at every step, it alternates between global candidate screening, adaptive pruning, and localized window refinement, while preserving the essential region-ranking behavior of greedy search. We analyze PhaseWin under monotone evidence-accumulation conditions and show that, under feature-level structural assumptions, it attains controllable linear evaluation complexity together with near-greedy faithfulness guarantees. Extensive experiments on image classification, object detection, visual grounding, and image captioning show that, among all compared attribution methods, PhaseWin reaches high faithfulness with the fewest forward passes, empirically realizing the predicted reduction from $O(n^2)$ to $O(n)$. The code is available at https://github.com/Qihuai27/phasewin-va.

cs.CV↗
arXiv2026

Rotational Symmetry based Object Pose Estimation from Point Clouds in the Absence of Known 3D Models

Object pose estimation is crucial to many industrial applications, with one example being automated spray painting using a robot. However, confidentiality concerns often limit access to high-quality 3D models, posing a significant challenge for point-cloud-based pose estimation. In such scenarios, rotational symmetry, a readily accessible characteristic of many industrial objects, can provide valuable prior information to facilitate pose estimation.In this paper, we propose a method that leverages the rotational symmetry commonly found in industrial objects to address the challenge caused by the absence of 3D models. The object pose is jointly estimated with point cloud refinement through an iterative optimization process. This optimization relies on a rotational symmetry constraint loss. To construct this loss, each 3D point is rotated according to the currently estimated pose, and multiple correspondences are identified using nearest-neighbor search by exploiting the rotational symmetry property. These correspondences are then used to compute the rotational symmetry constraint loss, which iteratively refines both the pose and the point cloud.By explicitly incorporating rotational symmetry into the optimization process, the proposed method achieves robust pose estimation and generalizes well across diverse object types. The proposed method is evaluated on a dataset specifically created for point clouds without known 3D models, consisting of four categories of synthetic objects and one real wheel hub collected from a production line. Experimental results demonstrate that the proposed method achieves performance comparable to methods that rely on known 3D models.

cs.CV↗
arXiv2026

Integration by Parts Formulas of Mckean-Vlasov SDEs with Jumps and Some Applications

In this article, we establish integration by parts formulas for the solutions of McKean-Vlasov stochastic differential equations with jumps under elliptic coefficients. The derived formulas accommodate both derivatives with respect to real-valued variables and measure-valued variables, interpreted through the Lions' derivative. As applications, we obtain estimates for the derivatives of the density functions of the McKean-Vlasov SDEs, and relying on the integration by parts formulas, we subsequently prove the existence and uniqueness of classical solutions to the associated PDEs with irregular terminal conditions.

math.PR↗
arXiv2026

Less is More: Efficient Black-box Attribution via Minimal Interpretable Subset Selection

To develop a trustworthy AI system, which aim to identify the input regions that most influence the models decisions. The primary task of existing attribution methods lies in efficiently and accurately identifying the relationships among input-prediction interactions. Particularly when the input data is discrete, such as images, analyzing the relationship between inputs and outputs poses a significant challenge due to the combinatorial explosion. In this paper, we propose a novel and efficient black-box attribution mechanism, LiMA (Less input is More faithful for Attribution), which reformulates the attribution of important regions as an optimization problem for submodular subset selection. First, to accurately assess interactions, we design a submodular function that quantifies subset importance and effectively captures their impact on decision outcomes. Then, efficiently ranking input sub-regions by their importance for attribution, we improve optimization efficiency through a novel bidirectional greedy search algorithm. LiMA identifies both the most and least important samples while ensuring an optimal attribution boundary that minimizes errors. Extensive experiments on eight foundation models demonstrate that our method provides faithful interpretations with fewer regions and exhibits strong generalization, shows an average improvement of 36.3% in Insertion and 39.6% in Deletion. Our method also outperforms the naive greedy search in attribution efficiency, being 1.6 times faster. Furthermore, when explaining the reasons behind model prediction errors, the average highest confidence achieved by our method is, on average, 86.1% higher than that of state-of-the-art attribution algorithms. The code is available at https://github.com/RuoyuChen10/LIMA.

cs.LG↗
arXiv2026

Wave-number-dependent closure condition for fluid moment equations

Fluid models offer crucial computational efficiency for plasma simulations, yet accurately capturing kinetic effects like Landau damping remains a fundamental challenge. While conventional closures (e.g., Hammett-Perkins and Hunana) are widely used, their fidelity relative to exact kinetic response degrades significantly depending on the perturbation wave number. Here, we propose a novel wave-number-dependent closure condition for the three-moment fluid equations that explicitly preserves the primary dispersion relation. By mapping Padé approximant coefficients directly to the kinetic roots of the collisionless Vlasov-Poisson system, we derive an analytical closure that rigorously embeds exact kinetic scaling across all spatial scales. We further demonstrate that this framework readily extends to collisional plasmas via the BGK model. This deterministic approach precisely captures the long-term macroscopic evolution of fluid moments and field energy, offering a rigorous foundation for high-fidelity fluid modeling.

physics.plasm-ph↗
arXiv2026

Cs$_4$Cr$_7$Te$_{10}$: Interwoven Reconstructed Archimedean and Kagome Lattices with a Possible Phase Transition near 130 K

Chromium-based materials with complex lattice geometries provide an important platform for investigating correlated electronic and magnetic states. However, Cr-based compounds with unusual crystal geometries are still rarely reported. Here, we report a new Cr-based compound, Cs$_4$Cr$_7$Te$_{10}$, featuring interwoven Cr and Te sublattices that can be viewed as reconstructed networks derived from Archimedean (3.4.6.4) tiling and the kagome lattice, respectively. Transport measurements reveal the semiconducting nature in Cs$_4$Cr$_7$Te$_{10}$. Magnetization measurements show a weak anisotropy between H//b and H//ac planes, and uncover an anomaly near 130 K that is insensitive to the applied magnetic fields. Specific-heat measurements further confirm this transition, indicating its bulk thermodynamic nature. The associated entropy change is as small as 0.41 J mol^-1 K^-1, ruling out a structural phase transition and pointing to a possible electronic and/or magnetic phase transition. These results provide a new route for designing complex crystal geometries and exploring their associated emergent phenomena.

cond-mat.mtrl-sci↗
arXiv2026

Where MLLMs Attend and What They Rely On: Explaining Autoregressive Token Generation

Multimodal large language models (MLLMs) have demonstrated remarkable capabilities in aligning visual inputs with natural language outputs. Yet, the extent to which generated tokens depend on visual modalities remains poorly understood, limiting interpretability and reliability. In this work, we present EAGLE, a lightweight black-box framework for explaining autoregressive token generation in MLLMs. EAGLE attributes any selected tokens to compact perceptual regions while quantifying the relative influence of language priors and perceptual evidence. The framework introduces an objective function that unifies sufficiency (insight score) and indispensability (necessity score), optimized via greedy search over sparsified image regions for faithful and efficient attribution. Beyond spatial attribution, EAGLE performs modality-aware analysis that disentangles what tokens rely on, providing fine-grained interpretability of model decisions. Extensive experiments across open-source MLLMs show that EAGLE consistently outperforms existing methods in faithfulness, localization, and hallucination diagnosis, while requiring substantially less GPU memory. These results highlight its effectiveness and practicality for advancing the interpretability of MLLMs.

cs.CV↗
arXiv2026

Deconstructing Positional Information: From Attention Logits to Training Biases

Positional encodings enable Transformers to incorporate sequential information, yet their theoretical understanding remains limited to two properties: distance attenuation and translation invariance. Because natural language lacks purely positional data, the interplay between positional and semantic information is still underexplored. We address this gap by deconstructing the attention-logit computation and providing a structured analysis of positional encodings, categorizing them into additive and multiplicative forms. The differing properties of these forms lead to distinct mechanisms for capturing positional information. To probe this difference, we design a synthetic task that explicitly requires strong integration of positional and semantic cues. As predicted, multiplicative encodings achieve a clear performance advantage on this task. Moreover, our evaluation reveals a hidden training bias: an information aggregation effect in shallow layers that we term the single-head deposit pattern. Through ablation studies and theoretical analysis, we proved that this phenomenon is inherent in multiplicative encodings. These findings deepen the understanding of positional encodings and call for further study of their training dynamics.

cs.LG↗
arXiv2025

Bismut-Elworthy-Li Formulae for Forward-Backward SDEs with Jumps and Applications

Under nondegeneracy assumptions on the diffusion coefficients, we establish the derivative formulae of Bismut-Elworthy-Li's type for forward-backward stochastic differential equations with respect to Poisson random measure using the lent particle method created by Bouleau and Denis, which is not given before. Applying this formula, the existence and uniqueness of a solution of nonlocal quasi-linear integral partial differential equations, which are differentiable with respect to the space variable, are obtained, even if the initial datum and coefficients of this equation are not.

math.PR↗
arXiv2025

CoDS: Collaborative Perception via Digital Semantic Communication

Semantic communication has been introduced into collaborative perception systems for autonomous driving, offering a promising approach to enhancing data transmission efficiency and robustness. Despite its potential, existing semantic communication approaches predominantly rely on analog transmission models, rendering these systems fundamentally incompatible with the digital architecture of modern vehicle-to-everything (V2X) networks and posing a significant barrier to real-world deployment. To bridge this critical gap, we propose CoDS, a novel collaborative perception framework based on digital semantic communication, designed to realize semantic-level transmission efficiency within practical digital communication systems. Specifically, we develop a semantic compression codec that extracts and compresses task-oriented semantic features while preserving downstream perception accuracy. Building on this, we propose a novel semantic analog-to-digital converter that converts these continuous semantic features into a discrete bitstream, ensuring integration with existing digital communication pipelines. Furthermore, we develop an uncertainty-aware network (UAN) that assesses the reliability of each received feature and discards those corrupted by decoding failures, thereby mitigating the cliff effect of conventional channel coding schemes under low signal-to-noise ratio (SNR) conditions. Extensive experiments demonstrate that CoDS significantly outperforms existing semantic communication and traditional digital communication schemes, achieving state-of-the-art perception performance while ensuring compatibility with practical digital V2X systems.

eess.SP↗