Searcharxiv⌕ Search

arXiv subjects

Xiaoyun Wang

Publications and source records attributed to Xiaoyun Wang.

At least 19 recordsLinked to original sources

DiFiT: Distribution-Fidelity-guided Visual Token Pruning for Multimodal Large Language Models

Visual token pruning reduces the inference cost of multimodal large language models, but answer accuracy only partly reflects its effect. Pruning can change the probabilities of alternative response tokens even when greedy answers remain correct and unchanged. We compare full and pruned response distributions under shared reference prefixes. Average predictive drift grows with compression on fixed correctly answered samples. Separate analyses associate larger drift with more confidence-decision disagreements and sampling failures. We propose DiFiT (Distribution-Fidelity-guided Token pruning) to learn discrete selection from reference answers and full-token response distributions. Dual-View Distillation matches probabilities across the vocabulary, while answer supervision supports task quality. A learned scorer and differentiable Top-$K$ learn which visual evidence to retain. A multi-budget mask objective supports different retention ratios through one shared ranking over a frozen model. On Qwen2.5-VL-7B, DiFiT reduces sampling-induced negative flips by 30.3% relative to FastV at 10% retention on four-task greedy-correct controls. It preserves 98.74% of baseline performance on MME at 30% retention.

cs.CV↗

DynaCrys: Crystal Generation with Dynamic Space-Group Diffusion

The search for new crystalline materials spans an enormous compositional and structural space. Generating candidates in this space requires modeling the relationships between elemental composition, crystallographic symmetry, and crystal geometry. The space group specifies the crystal's symmetry and determines the admissible Wyckoff positions and geometric degrees of freedom; fixing it in advance prevents this global symmetry choice from being revised using information from evolving Wyckoff occupations and elemental composition. We introduce DynaCrys, the first framework for complete crystal generation in which the space group co-evolves with Wyckoff occupations and elements. We develop physics-informed space-group diffusion that explicitly incorporates crystallographic group-subgroup relations, and a legality-constrained stochastic decoder that jointly samples compatible space groups, Wyckoff occupations, and elements, followed by symmetry-constrained geometry generation. In large-scale evaluations on standard benchmarks, DynaCrys achieves state-of-the-art performance in symmetry-aware crystal generation while supporting fast sampling.

cond-mat.mtrl-sci↗

Confidential Error-Correcting Codes

We introduce Confidential Error-correcting Codes (CECs), a new cryptographic primitive that enables reliable decoding from perturbed codewords while preserving message confidentiality. We formalize this primitive through three security properties: confidentiality, robustness, and soundness. To realize this goal, we develop a general framework for constructing CECs based on a Randomness-Security Separation (RSS) structure, which first encodes a nonce to introduce randomness and then masks the encoded payload with a pseudorandom function. Based on our RSS framework, we instantiate CECs, called Conc-CEC, using concatenated Reed-Muller and Reed-Solomon codes. The resulting Conc-CEC admits provable guarantees for all three properties. Notably, compared to PRCs (CRYPTO'24), another recent cryptographic primitive with error-correcting capability, CECs achieve significantly stronger concrete security guarantees under practical parameter settings. As a primary application of CECs, we further present CECMark, a watermarking scheme for AI-generated images. Our experiments on real-world generative image models show that CECMark consistently outperforms state-of-the-art watermarking methods, including Gaussian Shading (CVPR'24), PRCMark (ICLR'25), and T2SMark (NeurIPS'25). In particular, CECMark achieves higher robustness across all 56 evaluated perturbations and ensures a false positive rate below $10^{-5}$ without introducing side effects. Taken together, our work lays a novel cryptographic foundation for trustworthy watermarking schemes that, for the first time, unify provable security guarantees with strong practical reliability.

cs.CR↗

An RL View of OPD: Least Square Policy Distillation for Sample-Efficient LLM Reasoning

We study on-policy distillation (OPD) through the lens of reinforcement learning, establishing a connection between the reverse-KL objective in OPD and KL-regularized policy optimization. Building on this connection, we introduce Least-Square Policy Distillation (LSPD), an RL-inspired framework that brings optimistic exploration and off-policy data reuse from value-based RL into policy distillation. LSPD preserves policy diversity through exploration while improving rollout efficiency by repeatedly learning from previously collected trajectories. Our theoretical analysis connects LSPD to optimistic value-based learning and shows that its idealized formulation achieves a sharp $\tilde{\mathcal O}(\log K)$ regret bound under online exploration. Empirically, LSPD consistently outperforms existing distillation baselines across six mathematical reasoning benchmarks and diverse teacher-student settings, with average gains of +1.59 points in Avg@16. Remarkably, through Pass@k evaluations up to k=64, we found that LSPD better preserves policy diversity by achieving stronger performance as k grows. Its fully off-policy variant achieves comparable performance to vanilla OPD using only the first 25% of rollout batches. Together, these results provide an RL perspective on OPD that offers both a principled interpretation and a practical route toward more effective and rollout-efficient language model distillation.

cs.LG↗

GuidedRay: Diversity-Guided Direction Discovery for Targeted Hard-Label Black-Box Attacks

Deep neural networks are vulnerable to adversarial attacks. Among black-box attacks, targeted decision-based attacks are particularly difficult: the attacker observes only the target model's top-1 label and aims to make it predict a prespecified target class under a bounded perturbation. Before perturbation refinement, the attacker must discover a direction that reaches the prescribed target region. This initialization step can incur substantial query cost. We propose GuidedRay, a targeted decision-based attack based on diversity-guided direction discovery. GuidedRay builds on two observations: target-class reference samples provide useful target-conditioned direction priors, and diverse candidates increase the probability of discovering a targeted adversarial direction. GuidedRay generates varied candidates from one or multiple target-class references and uses a one-query Fast Test to screen their induced sign directions. Once a feasible direction is found, GuidedRay applies Ray Search to reduce its decision-boundary radius. Experiments on CIFAR-10, CIFAR-100, and ImageNet demonstrate that GuidedRay consistently outperforms five state-of-the-art decision-based attacks at four evaluated query budgets from 500 to 5,000, with particularly pronounced gains in direction discovery during initialization. Against models protected by adversarial training or TRADES, it likewise achieves the highest attack success rate at all four query budgets.

cs.CR↗

Cryptanalytic Extraction of Neural Networks Without Known Architecture Assumption

Cryptanalytic model extraction aims to reconstruct a functionally equivalent model through black-box interactions with the victim model. Under the fundamental assumption that the network architecture is completely known, existing attacks achieve the goal by recovering the model parameters. In this paper, we explore whether this assumption can be removed practically. Focusing on ReLU fully connected networks, which are widely studied in this field, we propose a guess-and-determine framework that jointly recovers the network architecture (including network depth and hidden-layer dimensions) and the model parameters. This framework is based on a simple yet effective high-level idea: after designing a parameter recovery attack under the known-architecture assumption, we can analyze the architecture-sensitive traces observed during parameter recovery to recover the network architecture. We identify two such traces in differential extraction attacks: (i) a \emph{zero suffix} in the merged weight vectors produced by signature recovery, whose length reveals the hidden layer dimension; and (ii) an \emph{equality pattern} in the preimage-based sign recovery, which occurs only under the true hidden layer dimension. These two signals give rise to two routes for network architecture recovery. For the second-to-last layer, we further propose two methods, one for identifying it, and one for recovering its dimension. Practical end-to-end attacks are implemented on a wide range of ReLU neural networks, including both expansive and non-expansive networks. To the best of our knowledge, this is the first time the feasibility of achieving functionally equivalent extraction on deep neural networks, after removing the known-architecture assumption, has been demonstrated in practice.

cs.CR↗

The convergent laboratory: when AI reasoning, autonomous experiments, high performance and quantum computing reshape chemistry

This Comment emerges from TPC26 (https://tpc26.org), a conference convening leaders from academia, national laboratories, and industry who are reshaping materials science discovery. The meeting explored how AI, autonomous agents, self-driving labs, higher performance and quantum computing converge to amplify their individual impact on materials science discovery. The perspectives here reflect the firsthand experiences of researchers at these frontiers and capture the essence of this global endeavor. As AI-driven reasoning, autonomous agentic frameworks, self-driving laboratories, and fault-tolerant quantum processors mature simultaneously, we offer this Comment as a reference at what we believe is a tipping point of transformative advances and productive disruption in the chemical sciences.

cs.AI↗

One Risk Down, Another Up: Cross-Risk Interactions Induced by LLM Defenses

Large Language Models (LLMs) are increasingly deployed in high-stakes settings, where they face diverse risks. Numerous defense strategies have been proposed to mitigate these risks, but they are almost always evaluated in isolation. This isolated view leaves a critical question open: does mitigating one risk inadvertently change a model's exposure to others? Beyond the well-studied risk-utility trade-off, we present the first systematic study of cross-risk interactions induced by LLM defenses. We propose CrossRiskEval, an evaluation paradigm that situates a defended model in a multi-dimensional risk space and quantifies how a defense built for one risk shifts the others. Among 166 cross-risk evaluations covering 32 defended models, 77.1% exhibit statistically significant cross-risk interactions. Most of these interactions amplify non-target risks, with increases exceeding 100% in some cases. Beyond behavioral evaluation, we conduct neuron-level analyses in seven selected cases to investigate one possible pathway associated with these interactions. We identify conflict-entangled neurons whose activation interventions produce opposing effects on proxies for the target and non-target risks. In conflict cases, restoring these neurons to their base-model activations partially reduces the corresponding risk increases, providing evidence that defense-induced changes to these neurons may contribute to the observed interactions. Building on this evidence, we propose Conflict-Aware Freezing, a training-time strategy that prevents direct updates to the parameters associated with the identified neurons. Across five conflict cases, it offsets 35%-196% of non-target risk amplification while meeting the original defense criterion.

cs.CR↗

Keep the Needle, Prune the Haystack: Defect-Preserving Token Pruning for Efficient Zero-Shot Anomaly Detection

Zero-shot visual anomaly detection has achieved remarkable progress, with recent vision-only approaches further improving performance while simplifying the inference pipeline. However, existing methods typically perform dense computation over all images and spatial tokens, despite the fact that normal samples dominate real-world scenarios and anomalies usually occupy only small regions. Token pruning offers a promising solution, but introduces an asymmetric pruning risk in anomaly detection: retaining normal tokens mainly incurs redundant computation, whereas removing anomalous tokens may eliminate the only evidence for detection and localization. This risk is particularly severe in early layers, where pruning provides the greatest computational benefit but anomaly semantics remain unreliable. We propose KeepAD, a defect-preserving token pruning framework that formulates token selection as high-recall, anomaly-aware routing. In shallow layers, KeepAD combines coverage-preserving selection over local $2\times2$ patch neighborhoods with deterministic anomaly rescue to reduce the risk of discarding subtle defects. In deeper layers, frozen normal and abnormal prototypes guide pruning under an image-adaptive token budget, aggressively removing low-risk normal tokens while preserving local anomaly evidence. Dense-to-sparse self-distillation further supervises early token routing without introducing additional inference overhead. Experiments on six industrial and seven medical zero-shot anomaly detection benchmarks show that KeepAD reduces the token retention ratio to below $20\%$, while limiting the average degradation in image-level and pixel-level AUROC to within $2.7$ percentage points. At the most aggressive operating point, KeepAD achieves a $7.9\times$ speedup over the strongest CLIP-based baseline.

cs.CV↗

Cryptanalysis of LDPC-Based Pseudorandom Error-Correcting Codes

Pseudorandom error-correcting codes (PRCs), a novel cryptographic primitive recently proposed at CRYPTO 2024, are primarily applied in undetectable watermarking schemes for large generative models. However, the security of PRCs has not yet been systematically analyzed. To fill this gap, we present the first cryptanalysis of PRCs. Specifically, focusing on LDPC-PRC, the only known practical instantiation of PRCs, we propose three novel attacks that challenge its undetectability and robustness. To rigorously demonstrate the practical threat, we analyze the concrete attack complexity under realistic parameters and validate the attack effectiveness on both real-world large language models and generative image models, including DeepSeek and Stable Diffusion. Our analysis shows that the claimed security guarantees of LDPC-PRC are undermined across all practically feasible regimes. For example, our attacks can detect the presence of a watermark with overwhelming probability at a cost of $2^{22}$ operations. Beyond attacks, we further propose three defenses: parameter recommendation, implementation suggestion, and a revised key generation function. However, PRC-based watermarking schemes still fail to achieve 128-bit security due to inherent constraints of large generative models, such as the maximum output length of large language models. Overall, our work clarifies the concrete security limits of PRCs in real-world watermarking applications.

cs.CR↗

Robust Watermarks Meet Backdoored Models: Evading Diffusion Semantic Watermarks via Stealthy Backdoor

Although semantic watermarking is considered a promising safeguard for images generated by Latent Diffusion Models (LDMs), the reliance of the watermark detection pipeline on neural networks introduces a critical yet underexplored backdoor attack surface. To systematically study this vulnerability, we propose GhostVAE to plant a stealthy backdoor into the encoder of Variational Autoencoder (VAE), enabling reliable evasion of watermark detection. GhostVAE operates in two stages: it first constructs a universal trigger via power spectrum regularization to improve the trigger robustness, and then trains a backdoored VAE encoder with a parameter-aligned objective. Through extensive evaluations across three state-of-the-art semantic watermarking schemes and three widely adopted LDMs, we show that GhostVAE preserves watermark detection performance on benign images (achieving an average true positive rate of 94.4%), while simultaneously enabling highly effective evasion under trigger activation (achieving an average attack success rate of 94.6%). Moreover, we comprehensively analyze seventeen representative defenses and demonstrate that GhostVAE remains stealthy across the input space, parameter space, and latent space. Our work fundamentally undermines the trustworthiness of semantic watermarking systems and highlights that secure deployment of semantic watermarks requires end-to-end security considerations, particularly for neural network components.

cs.CR↗

Unified Generalization for Frequency-Domain Channel Extrapolation Across Near-Field and Far-Field Scenarios

As antenna arrays grow, near-field effects become non-negligible in large-scale MIMO, making accurate low-overhead channel acquisition crucial in both far-field and near-field regimes. Deep-learning-based frequency-domain channel extrapolation can reduce pilot overhead, but existing extrapolators generalize poorly to unseen distances and environments, especially across near-field and far-field channels. We propose a physically interpretable framework to unify generalization across both regimes. Our key insight is that angular profiles are regime-dependent, while delay profiles share a sparsity structure that can be aligned. Based on this, we develop a physics-guided disentanglement and alignment pipeline with multi-cluster decoupling, angle-delay feature disentanglement, and delay-domain alignment, enabling the model to learn distribution-stable delay features while reusing heterogeneous angular features. We further design a unified near/far-field DL extrapolator (UNiFi-DLE) and detail its dataset preparation, training, and inference. Simulations and sim-to-real experiments show that UNiFi-DLE generalizes robustly to unseen near-field and far-field scenarios and consistently outperforms state-of-the-art methods.

eess.SP↗

EPTS: Elastic Post-Training Sparsity for Efficient Large Language Model Compression

Post-Training Sparsity (PTS) has emerged as a crucial paradigm for compressing Large Language Models to facilitate efficient deployment on resource-constrained devices. However, existing PTS methodologies are typically confined to Single-Sparsity optimization, necessitating a separate, time-consuming optimization session for each specific sparsity level. This rigid paradigm significantly hinders flexible deployment across diverse hardware scenarios, as adapting to a new sparsity requirement mandates a complete re-optimization process. To address these limitations, we propose Elastic Post-Training Sparsity (EPTS), a unified Multi-Sparsity framework that produces a single elastic model capable of maintaining robust performance across diverse sparsity configurations through a one-shot optimization process. Specifically, we design a Multi-Sparsity Hierarchy LoRA (MS-HiLoRA) mechanism that facilitates knowledge inheritance from low- to high-sparsity groups, effectively mitigating the competition for parameter reconstruction. Furthermore, we introduce a Multi-Sparsity Feature Mixer (MSFM), which significantly enhances the model's adaptability to pruning perturbations by dynamically fusing feature representations of varying sparsity granularities. Extensive experiments on LLaMA and OPT families demonstrate that EPTS achieves competitive performance compared to state-of-the-art methods like SparseGPT and Wanda, while offering significant efficiency gains by enabling multi-scenario deployment from a single optimization. our source code is available at https://github.com/xuke225/EPTS.

cs.LG↗

Physics-Informed Path-Parametric Learning for Efficient and Lightweight CSI Feedback

Channel State Information (CSI) feedback is vital for high spectral efficiency in wireless systems, yet high-dimensional CSI introduce significant feedback overhead. Recent deep learning (DL) approaches alleviate this issue by treating CSI as a visual image, but such "black-box" designs often lack interpretability, producing CSI that is not consistent with multipath propagation principles. To address these limitations, this paper proposes HS-PINNnet, a Hierarchical Sensing mechanism assisted Physics-Informed Neural Network for CSI Feedback. Unlike vision-inspired methods, HS-PINNnet integrates a multipath channel model into the network, reformulating high-dimensional CSI reconstruction as low-dimensional multipath parameter estimation (e.g., amplitude, angle). HS-PINNnet features a hierarchical sensing encoder to produce a compact multipath representation, and a heterogeneous decoder for parameter-specific CSI reconstruction, with dedicated branches to estimate different parameters. Moreover, a PCD module adaptively estimates the number of dominant paths in each CSI sample to enhance generalization across diverse environments. A subchannel-wise shared encoding and parallel decoding strategy is further designed to decompose high-dimensional CSI processing into low-dimensional subchannel tasks, reducing training difficulty and improving scalability of HS-PINNnet for future extremely large-scale multiple-input multiple-output (XL-MIMO) systems. Simulation results show that HS-PINNnet outperforms the state-of-the-art under different configurations, achieving a 92.8% reduction in FLOPs and exhibiting two orders of magnitude lower FPGA simulation latency.

eess.SP↗

Fuzzy PSI from Symmetric Primitives with Exact Logarithmic Dependence on Distance Threshold

Previous FPSI works have demonstrated a linear scaling with the distance threshold $δ$, while some recent works have achieved a poly-logarithmic dependence on $δ$. However, these protocols either support only the $L_\infty$ distance, or they support general $L_{p\in[1,\infty]}$ distances but rely on expensive additive homomorphic encryption (AHE). Achieving exact logarithmic dependence on $δ$ for general $L_{p\in[1,\infty]}$ distances without relying on costly AHE would constitute a theoretical breakthrough in optimal threshold scaling and a practical advance toward scalable FPSI applications. In this work, we present new FPSI protocols for $L_{p\in[1,\infty]}$ distances that are entirely built from oblivious transfer (OT) and symmetric-key primitives. We propose FPSI protocols based on both the apart and the separate assumptions, which are applicable to low- and high-dimensional settings, respectively. Our constructions achieve strictly logarithmic complexity in $δ$, which is optimal in the sense that distinguishing all values in an interval of length $O(δ)$ necessarily requires $Ω(\log δ)$ bits of information. Our core idea is to perform fuzzy matching via prefix representation and interactively determine the correct prefix using equality conditions. To this end, we propose a suite of new components that can be implemented efficiently using only OT and symmetric-key operations. We implement our FPSI protocols and compare them with the state-of-the-art FPSI protocols for $L_{p\in[1,\infty]}$ distance. Experiments show that our protocols outperform the prior state-of-the-art by up to $43.7\times$ in runtime and $31.3\times$ in communication.

cs.CR↗

OGLS-SD: On-Policy Self-Distillation with Outcome-Guided Logit Steering for LLM Reasoning

We study on-policy self-distillation (OPSD), where a language model improves its reasoning ability by distilling privileged teacher distributions along its own on-policy trajectories. Despite its promise, OPSD can suffer from training instability due to a pattern mismatch between teacher and student responses. Self-reflected teacher responses may introduce reflection-induced biases and response templates that miscalibrate token-level supervision, ultimately harming the student's reasoning ability. To mitigate this issue, we propose OGLS-SD, an outcome-guided logit-steering framework that leverages verifiable outcome rewards to calibrate privileged teacher logits. Specifically, OGLS-SD contrasts teacher logits induced by successful and failed on-policy trajectories, constructing an outcome-discriminative steering direction for token-level guidance. Experiments on mathematical reasoning benchmarks show that OGLS-SD stabilizes self-distillation and improves performance over standard OPSD and other variants.

cs.LG↗

Local Spatiotemporal Convolutional Network for Robust Gait Recognition

Gait recognition, as a promising biometric technology, identifies individuals through their unique walking patterns and offers distinctive advantages including non-invasiveness, long-range applicability, and resistance to deliberate disguise. Despite these merits, capturing the intrinsic motion patterns concealed within consecutive video frames remains challenging due to the complexity of video data and the interference of external covariates such as viewpoint changes, clothing variations, and carrying conditions. Existing approaches predominantly rely on either static appearance features extracted from individual silhouette frames or employ complex sequential models (\eg, LSTM, 3D convolutions) that demand substantial computational resources and sophisticated training strategies. To address these limitations, we propose a Local Spatiotemporal Convolutional Network (LSTCN), a structurally simple yet highly effective dual-branch architecture that endows standard two-dimensional convolutional networks with the capacity to extract temporal information. Specifically, we introduce a Global Bidirectional Spatial Pooling (GBSP) mechanism that reduces the dimensionality of gait tensors by decomposing spatial features into horizontal and vertical strip-based local representations, enabling the temporal dimension to participate in standard 2D convolution operations. Building upon this, we design a Local Spatiotemporal Convolutional (LSTC) layer that jointly processes temporal and spatial dimensions, allowing the network to adaptively learn strip-based gait motion patterns. We further extend this formulation with asymmetric convolution kernels that independently attend to the temporal, spatial, and joint spatiotemporal domains, thereby enriching the extracted feature representations.

cs.CV↗

Gait Recognition via Deep Residual Networks and Multi-Branch Feature Fusion

Gait recognition has emerged as a compelling biometric modality for surveillance and security applications, offering inherent advantages such as non-intrusiveness, resistance to disguise, and long-range identification capability. However, prevailing approaches struggle to comprehensively capture and exploit the rich biometric cues embedded in human locomotion, particularly under covariate interference including viewpoint variation, clothing change, and carrying conditions. In this paper, we present a high-precision gait recognition framework that deeply extracts and synergistically fuses gait dynamics with body shape characteristics through a multi-branch architecture grounded in deep residual learning. Specifically, we first employ the High-Resolution Network (HRNet) to perform robust skeletal keypoint estimation, preserving fine-grained spatial information even under low-resolution inputs. We then construct three complementary feature branches -- body proportion, gait velocity, and skeletal motion -- from the extracted pose sequences. A 50-layer Residual Network (ResNet-50) backbone is leveraged within a deep feature extraction module to capture hierarchically rich and discriminative representations. To effectively integrate heterogeneous feature streams, we design a Multi-Branch Feature Fusion (MFF) module inspired by channel-wise attention mechanisms, which dynamically allocates contribution weights across branches through learned activation parameters. Extensive experiments on the cross-view multi-condition CASIA-B benchmark demonstrate that our method achieves a Rank-1 accuracy of 94.52\% under normal walking, with the best recognition performance among skeleton-based methods for the coat-wearing condition.

cs.CV↗