Searcharxiv⌕ Search

arXiv subjects

Yanping Li

Publications and source records attributed to Yanping Li.

12 recordsLinked to original sources
arXiv2026

Spatiotemporally Decoupled Autoregressive Diffusion Model for Human Motion Generation

Text-driven human motion synthesis has made substantial development with two core modules of motion representation and generative architecture. For representation, Vector Quantization (VQ)-based methods compress motion data into discrete tokens while latent-based models operate directly in continuous space. However, both of these representations exhibit significant limitations. VQ-based methods suffer from inherent information loss, which compromises the quality, diversity, and generalization of generated motions, while continuous representation on holistic whole-body motion hinders part-level flexibility. For architecture, diffusion and autoregressive diffusion models have demonstrated their superiority, yet the fine-grained controllability over individual body parts is also limited. Thus, we propose a unified spatiotemporally decoupled framework named DeMoDiff, which jointly redesigns representation and architecture. To enhance representation extraction capabilities and offer greater part-level controllability, we present a spatial-temporal VAE that encodes each body joint rather than compressing the whole-body motion into a single latent space. Then, we incorporate spatial-temporal masking and attention mechanisms into an autoregressive diffusion generator, achieving both generative capability and controllable editability. Extensive experiments on the HumanML3D and KIT-ML datasets demonstrate that our model achieves state-of-the-art reconstruction performance and compelling motion generation results. Moreover, our framework demonstrates strong temporal and spatial editing capabilities, further validating its effectiveness. Our project page: https://rex0191.github.io/DeMoDiff/

cs.CV↗
arXiv2026

MAFIA: Query-Only Memory Attacks via Probing and Factual Injection against Audited LLM Agents

Memory-augmented LLM agents rely on rich context for long-horizon reasoning and acting, yet their memory modules expose a persistent attack surface for malicious records, making the study of memory poisoning threats imperative. However, existing query-only attacks often fail to remain effective in two realistic and prevalent settings: large-scale benign memory pools and active input auditing. Consequently, current approaches fall short when facing the dual challenges of high retrieval competitiveness and rigorous semantic checks. To overcome these limitations, we propose MAFIA, a query-only Memory Attack framework via probing and Factual Injection against Audit, tailored to this extended threat model. Specifically, MAFIA introduces: (1) a placement strategy that ensures retrieval-competitive injection via memory probing, budget allocation, and scheduling; and (2) a payload design that bypasses audits using compact factual cloaks, preserving malicious effects while maintaining high semantic similarity. Extensive evaluations reveal that MAFIA achieves up to a 90.7% attack success rate while suppressing audit detection from a peak of 83.3% to at most 7.4%, exposing critical vulnerabilities across agentic memory systems. Code will be made publicly available at https://github.com/JiamingChen1234/MAFIA.

cs.AI↗
arXiv2026

SFMambaNet: Spectral-Frequency Enhanced Selective State Space Model for Correspondence Pruning

Correspondence pruning aims to identify inliers from an initial set of correspondences. Most existing Graph Neural Network (GNN)-based methods rely on geometric features mapped from coarse Euclidean coordinates, which struggle to capture the subtle geometric consistencies presented by inliers. While Mamba-based methods possess global receptive fields and long sequence modeling capabilities, they tend to accumulate substantial inconsistent features within the hidden state space, making it difficult to distinguish inliers from outliers. In this paper, we integrate frequency domain perception into this task for the first time and propose SFMambaNet, a novel Spectral-Frequency enhanced Mamba-based two-view correspondence pruning network. Our method is collaboratively composed of two components: First, we design a Local Spectral-Geometric Attention (LSGA) block. LSGA incorporates spectral positional encoding into local graph interactions and introduces multi-scale Mamba processing to enhance the capture of subtle geometric consistencies and improve local feature discriminability. Building upon this, we design a Spectral-Integrated Global Mamba (SIGM) block. SIGM embeds a frequency gating mechanism within the state space, utilizing the frequency information provided by LSGA to explicitly suppress high-frequency noise accumulation within hidden states and mitigate the propagation of inconsistent features. This enhances inlier-outlier separability and achieves robust global context modeling capabilities with nearly linear complexity. Extensive experiments demonstrate that SFMambaNet outperforms current state-of-the-art methods on several challenging tasks. The code is available at https://github.com/Kirito14IT/SFMambaNet.

cs.CV↗
arXiv2026

GEAR: Granularity-Adaptive Advantage Reweighting for LLM Agents via Self-Distillation

Reinforcement learning has become a widely used post-training approach for LLM agents, where training commonly relies on outcome-level rewards that provide only coarse supervision. While finer-grained credit assignment is promising for effective policy updates, obtaining reliable local credit and assigning it to the right parts of the long-horizon trajectory remains an open challenge. In this paper, we propose Granularity-adaptivE Advantage Reweighting (GEAR), an adaptive-granularity credit assignment framework that reshapes the trajectory-level GRPO advantage using token- and segment-level signals derived from self-distillation. GEAR compares an on-policy student with a ground-truth-conditioned teacher to obtain a reference-guided divergence signal for identifying adaptive segment boundaries and modulating local advantage weights. This divergence often spikes at the onset of a semantic deviation, while later tokens in the same autoregressive continuation may return to low divergence. GEAR therefore treats such spikes as anchors for adaptive credit regions: where the student remains aligned with the teacher, token-level resolution is preserved; where it departs, GEAR groups the corresponding continuation into an adaptive segment and uses the divergence at the departure point to modulate the segment' s advantage. Experiments across eight mathematical reasoning and agentic tool-use benchmarks with Qwen3 4B and 8B models show that GEAR consistently outperforms standard GRPO, self-distillation-only baselines, and token- or turn-level credit-assignment methods. The gains are especially strong on benchmarks with lower GRPO baseline accuracy, reaching up to around 20\% over GRPO, suggesting that the proposed adaptive reweighting scheme is especially useful in more challenging long-horizon settings.

cs.LG↗
arXiv2026

Token-level Data Selection for Safe LLM Fine-tuning

Fine-tuning large language models (LLMs) on custom datasets has become a standard approach for adapting these models to specific domains and applications. However, recent studies have shown that such fine-tuning can lead to significant degradation in the model's safety. Existing defense methods operate at the sample level and often suffer from an unsatisfactory trade-off between safety and utility. To address this limitation, we perform a systematic token-level diagnosis of safety degradation during fine-tuning. Based on this, we propose token-level data selection for safe LLM fine-tuning (TOSS), a novel framework that quantifies the safety risk of each token by measuring the loss difference between a safety-degraded model and a utility-oriented model. This token-level granularity enables accurate identification and removal of unsafe tokens, thereby preserving valuable task-specific information. In addition, we introduce a progressive refinement strategy, TOSS-Pro, which iteratively enhances the safety-degraded model's ability to identify unsafe tokens. Extensive experiments demonstrate that our approach robustly safeguards LLMs during fine-tuning while achieving superior downstream task performance, significantly outperforming existing sample-level defense methods. Our code is available at https://github.com/Polly-LYP/TOSS.

cs.CL↗
arXiv2025

RemedyGS: Defend 3D Gaussian Splatting against Computation Cost Attacks

As a mainstream technique for 3D reconstruction, 3D Gaussian splatting (3DGS) has been applied in a wide range of applications and services. Recent studies have revealed critical vulnerabilities in this pipeline and introduced computation cost attacks that lead to malicious resource occupancies and even denial-of-service (DoS) conditions, thereby hindering the reliable deployment of 3DGS. In this paper, we propose the first effective and comprehensive black-box defense framework, named RemedyGS, against such computation cost attacks, safeguarding 3DGS reconstruction systems and services. Our pipeline comprises two key components: a detector to identify the attacked input images with poisoned textures and a purifier to recover the benign images from their attacked counterparts, mitigating the adverse effects of these attacks. Moreover, we incorporate adversarial training into the purifier to enforce distributional alignment between the recovered and original natural images, thereby enhancing the defense efficacy. Experimental results demonstrate that our framework effectively defends against white-box, black-box, and adaptive attacks in 3DGS systems, achieving state-of-the-art performance in both safety and utility.

cs.CV↗
arXiv2025

High-quality metalens enables minimally invasive CFB endoscopy

Metalenses, owing to their ultra-thin planar structures, present a promising solution for reducing endoscopic invasiveness. However, achieving high-quality imaging with minimal invasiveness (short focal length of metalens) remains a critical challenge. This paper presents a deep learning assisted metalens with a 1mm focal length tailored for coherent fiber bundle, constituting the least invasive metalens-CFB system reported to date. To overcome the increased chromatic dispersion and aberrations associated with high NA metalens, we compensate for lateral etching at the base of the nanopillars by adjusting the thickness of a sacrificial hard mask. This approach enables the fabrication of nanopillars with small cross sections and high aspect ratios, featuring nearly vertical sidewalls (~90{\deg}), thereby enhancing the phase accuracy of the metalens. Experimental validation using the metalens-CFB system demonstrates that the metalens achieves an expanded field of view of 48.3{\deg} and a depth of field exceeding 125 mm. This work establishes a new paradigm for ultra-minimally invasive endoscopic imaging.

physics.optics↗
arXiv2023

Learning multi-domain feature relation for visible and Long-wave Infrared image patch matching

Recently, learning-based algorithms have achieved promising performance on cross-spectral image patch matching, which, however, is still far from satisfactory for practical application. On the one hand, a lack of large-scale dataset with diverse scenes haunts its further improvement for learning-based algorithms, whose performances and generalization rely heavily on the dataset size and diversity. On the other hand, more emphasis has been put on feature relation in the spatial domain whereas the scale dependency between features has often been ignored, leading to performance degeneration especially when encountering significant appearance variations for cross-spectral patches. To address these issues, we publish, to be best of our knowledge, the largest visible and Long-wave Infrared (LWIR) image patch matching dataset, termed VL-CMIM, which contains 1300 pairs of strictly aligned visible and LWIR images and over 2 million patch pairs covering diverse scenes such as asteroid, field, country, build, street and water.In addition, a multi-domain feature relation learning network (MD-FRN) is proposed. Input by the features extracted from a four-branch network, both feature relations in spatial and scale domains are learned via a spatial correlation module (SCM) and multi-scale adaptive aggregation module (MSAG), respectively. To further aggregate the multi-domain relations, a deep domain interactive mechanism (DIM) is applied, where the learnt spatial-relation and scale-relation features are exchanged and further input into MSCRM and SCM. This mechanism allows our model to learn interactive cross-domain feature relations, leading to improved robustness to significant appearance changes due to different modality.

cs.CV↗
arXiv2023

Probing the Anomalous Hall Transport and Magnetic Reversal of Chiral-Lattice Antiferromagnet Co$_{1/3}$NbS$_2$

Antiferromagnets exhibiting giant anomalous Hall effect (AHE) and anomalous Nernst effect (ANE) have recently aroused broad interest, not only for their potential applications in future electronic devices, but also because of the rich physics arising from the Berry curvature near the Fermi level. $\rm{Co_{1/3}NbS_2}$, by intercalating $\rm{Co^{2+}}$ ions between $\rm{NbS_2}$ layers, is a quasi-two-dimensional layered antiferromagnet with a chiral lattice. A large AHE has been observed in $\rm{Co_{1/3}NbS_2}$, but its origin is under debate. In this letter, we report the large AHE and ANE in exfoliated $\rm{Co_{1/3}NbS_2}$ flakes. By analyzing the thermoelectric data via the Mott relation, we determined that the observed large AHE and ANE primarily result from the intrinsic Berry curvature. We also observed the magnetic domains in $\rm{Co_{1/3}NbS_2}$ by reflective magnetic circular dichroism measurements. Combined with electrical transport measurements, we confirmed that the magnetic reversal in $\rm{Co_{1/3}NbS_2}$ is determined by domain wall motion, and the critical field ($H_c$) exhibits a memory effect of consecutive magnetic sweeps. Our work provides insight into the topological properties of $\rm{Co_{1/3}NbS_2}$ and paves the way to studying the spin configuration and magnetic domain dynamics in this fascinating antiferromagnet.

cond-mat.mes-hall↗
arXiv2022

Ferromagnetic-antiferromagnetic coexisting ground states and exchange bias effects in $\bf{MnBi_4Te_7}$ and $\bf{MnBi_6Te_{10}}$

Natural superlattice structures $\rm{(MnBi_2Te_4)(Bi_2Te_3)}$$_n$ ($n$ = 1, 2,...), in which magnetic $\rm{MnBi_2Te_4}$ layers are separated by nonmagnetic $\rm{Bi_2Te_3}$ layers, hold band topology, magnetism and reduced interlayer coupling, providing a promising platform for the realization of exotic topological quantum states. However, their magnetism in the two-dimensional limit, which is crucial for further exploration of quantum phenomena, remains elusive. Here, complex ferromagnetic (FM)-antiferromagnetic (AFM) coexisting ground states that persist up to the 2-septuple layers (SLs) limit are observed and comprehensively investigated in $\rm{MnBi_4Te_7}$ ($n$ = 1) and $\rm{MnBi_6Te_{10}}$ ($n$ = 2). The ubiquitous Mn-Bi site mixing modifies or even changes the sign of the subtle inter-SL magnetic interactions, yielding a spatially inhomogeneous interlayer coupling. Further, a tunable exchange bias effect is observed in $\rm{(MnBi_2Te_4)(Bi_2Te_3)}$$_n$ ($n$ = 1, 2), arising from the coupling between the FM and AFM components in the ground state. Our work highlights a new approach toward the fine-tuning of magnetism and paves the way for further study of quantum phenomena in $\rm{(MnBi_2Te_4)(Bi_2Te_3)}$$_n$ ($n$ = 1, 2,...) as well as their magnetic applications.

cond-mat.mtrl-sci↗
arXiv2021

Research on Gender-related Fingerprint Features

Fingerprint is an important biological feature of human body, which contains abundant gender information. At present, the academic research of fingerprint gender characteristics is generally at the level of understanding, while the standardization research is quite limited. In this work, we propose a more robust method, Dense Dilated Convolution ResNet (DDC-ResNet) to extract valid gender information from fingerprints. By replacing the normal convolution operations with the atrous convolution in the backbone, prior knowledge is provided to keep the edge details and the global reception field can be extended. We explored the results in 3 ways: 1) The efficiency of the DDC-ResNet. 6 typical methods of automatic feature extraction coupling with 9 mainstream classifiers are evaluated in our dataset with fair implementation details. Experimental results demonstrate that the combination of our approach outperforms other combinations in terms of average accuracy and separate-gender accuracy. It reaches 96.5% for average and 0.9752 (males)/0.9548 (females) for separate-gender accuracy. 2) The effect of fingers. It is found that the best performance of classifying gender with separate fingers is achieved by the right ring finger. 3) The effect of specific features. Based on the observations of the concentrations of fingerprints visualized by our approach, it can be inferred that loops and whorls (level 1), bifurcations (level 2), as well as line shapes (level 3) are connected with gender. Finally, we will open source the dataset that contains 6000 fingerprint images

cs.CV↗