arXiv subjects

Sheng Li

Publications and source records attributed to Sheng Li.

At least 19 recordsLinked to original sources

arXiv2026

Venus-DefakerOne2: A Vision-Forensics Foundation Model with Broader Coverage

Real-world forgery is inherently heterogeneous: it manifests not only as digital manipulation but also as physical spoofing, and is carried not only by images but also by videos. Despite substantial progress, a unified anti-forgery foundation model covering such a broad threat spectrum remains absent. In this paper, we present Venus-DefakerOne2, which substantially extends the forensic capability boundary along both the forgery-form and media-type dimensions. Trained on tens of millions of open-source and in-house samples, it incorporates two key designs. Tide-Style-Tuning alternates LoRA and full-parameter fine-tuning to balance computational efficiency, generalization, and hard-sample fitting while mitigating catastrophic forgetting. Entropy-based Test Adaptation determines the optimal prediction modality by analyzing prediction entropy. Venus-DefakerOne2 achieves new state-of-the-art performance across a comprehensive suite of benchmarks, including DeepFake faces (96.7%), AI-generated images (93.3%), tampered documents (87.6%), manipulated natural images (89.6%), physical spoofs (93.1%), and AI-generated videos (96.0%). It significantly outperforms both specialized expert models and its predecessor, DefakerOne, establishing itself as a powerful and versatile foundation model for binary-decision forensic tasks.

cs.CV↗

arXiv2026

IndustryNav: Evaluating Spatial Reasoning and Safety of Embodied Agents in Dynamic Industrial Navigation

Visual large language models (VLLMs) are increasingly used as embodied agents, yet industrial warehouses demand more than reaching a goal: agents must navigate safely around moving workers and vehicles. To evaluate embodied agents' goal reaching and safety, we introduce IndustryNav, a simulation environment with 24 warehouse scenes for closed-loop navigation to specified destinations. Alongside task success and efficiency, we introduce safety-aware metrics that capture collisions and near-collision warnings along each trajectory. Experiments with eleven leading open- and closed-source VLLMs and three navigation baselines show that agents can reach their goals yet collide with moving obstacles, exposing a gap between success and safe navigation. This gap matters in safety-critical settings, where a single collision can end the task. IndustryNav also supports policy learning through automated oracle trajectories and simulator interaction, with held-out scenes for generalization tests. These experiments show that policies can succeed on new routes in familiar warehouses yet fail to transfer to held-out layouts, making cross-scene generalization an open challenge.

cs.RO↗

arXiv2026

CFBrainAug: Modularity-Induced Counterfactual Brain Network Augmentation for Interpretable fMRI-based Autism Diagnosis

Machine learning on resting-state functional magnetic resonance imaging (rs-fMRI) brain networks is promising for diagnosing Autism Spectrum Disorder (ASD), but existing methods are prone to overfitting on limited training data, and their post-hoc statistical attribution offers limited clinical interpretability. We propose CFBrainAug, a counterfactual brain network learning framework that unifies data augmentation and counterfactual explanation by learning label-altered brain networks. Unlike conventional label-preserving augmentation, CFBrainAug first retrieves a label-opposite yet topologically similar candidate and then refines it under a module-number constraint. Jointly training on factual and counterfactual networks enriches the training distribution around the decision boundary, enabling a lightweight feature-selection-based SVM to perform competitively with recent graph deep learning methods. Meanwhile, the counterfactuals provide intrinsic explanations by identifying the connectivity differences between each subject and its counterfactual. Experiments on the ABIDE I dataset show that CFBrainAug reaches an accuracy of 74.1% on the NYU site, suggesting that counterfactual augmentation can complement more complex models in small-sample brain network analysis.

cs.LG↗

arXiv2026

MAS-Guard: Unified Guardrails from Individual Agents to Multi-Agent LLM Systems

Agentic AI extends large language models (LLMs) from generating responses to planning and executing complex workflows, substantially expanding their utility while exposing applications to unsafe actions, information leakage, and adversarial manipulation. Multi-agent systems further expand these capabilities through specialization and collaboration, but distributed authority, heterogeneous roles, and opaque or potentially compromised participants make safeguarding even more challenging. Existing defenses address important individual threats but provide fragmented coverage across local execution, system-level interactions, and enforcement mechanisms. We introduce MAS-Guard, a unified framework that closes this gap for safeguarding single- and multi-agent AI applications containing both instrumentable and opaque agents. Within instrumentable agents, the proposed local Agent Guardrail (A-Guard) layer applies task-specific experts to assess agentic events. At the system level, developer-owned Agent Charters and a Control Plane combine deterministic structural checks of routes, agent capabilities, and message formats with semantic assessments of roles and message content. These complementary layers enable enforcement through a shared set of interventions: allow, reflection, confirmation, and block. Evaluated across six benchmarks covering diverse single- and multi-agent scenarios with three agent LLMs from different providers, MAS-Guard reduces the overall attack success rate by 71.5-86.3% relative to unguarded agents, while meeting the utility reference criterion on benign tasks.

cs.CR↗

arXiv2026

What Language Adds, and at What Cost: Four Controlled Probes of Pathology Vision-Language Models

Vision--language models (VLMs) promise to make computational pathology queryable and able to draft reports, but how much language adds beyond the vision backbone is unclear, because almost every system is evaluated by its developers on test sets of their own choosing. We grade published evidence by how it was produced and test four claims with controlled probes that reuse released checkpoints and public data, varying one input at a time. Report text adds prognostic signal: adding a frozen embedding of the stage-redacted report to a slide embedding raises the pooled C-index by 0.019 for progression-free interval and by 0.016 for overall survival across 1{,}633 TCGA patients, and restoring the stage statement adds nothing. Prompts add no measurable accuracy at inference: with features and pooling fixed, zero-shot classification is statistically indistinguishable from a supervised linear probe for lung and renal cancer subtyping. Slide-level question answering is barely grounded: SlideChat answers 74.82\% of SlideBench-VQA(TCGA) items correctly with the slide and 74.41\% without it, whereas its language backbone alone reaches 43.52\%. Paraphrasing leaves accuracy unchanged but flips 8--23\% of answers. Self-reported gains from agentic navigation and few-shot language priors cannot yet be attributed to language. Language thus adds real but narrow value, and we propose minimum controls that would expose these failure modes in future evaluations.

cs.CV↗

arXiv2026

Evaluating Tibetan Speech Recognition Across Dialects and Adaptation Regimes: A Multi-Granularity Analysis

Tibetan speech recognition is usually evaluated with character or syllable error rates because the writing system does not mark word boundaries. These metrics can rank systems differently and respond differently to adaptation failures. We combine character error rate, syllable error rate, and segmented word error rate in one evaluation framework. We use dictionary-based, neural, and language-model-based segmentation and compare their boundaries with expert annotations. We evaluate five ASR architectures on Lhasa Tibetan, then test zero-shot transfer and target-corpus adaptation on five additional corpus subsets. We also vary LoRA rank and training-data size to examine adaptation failure. The results show that a lower character error does not always lead to a lower syllable or segmented-word error. These ranking differences persist across corpora. Metric correlations also weaken in some unsuccessful adaptation settings. Higher LoRA ranks improve syllable and segmented-word scores but remain worse than full fine-tuning in the tested range. A standardized phonological representation provides another view of spelling errors. These findings support evaluation at several linguistic levels and motivate better segmentation and adaptation methods for low-resource ASR.

cs.CL↗

arXiv2026

What Survives Self-Evolution: Provenance Marks in Copied Agent Harnesses

Agent harnesses, the prompt and tool code around a base model, leak. Claude Code's 512,000 lines were mirrored within hours of a packaging error. A plagiarist can pass a leaked harness through a published self-evolver and ship the result as its own. We define this as a covert rewriting attack. No one targets a provenance mark, yet a large language model (LLM) proposer rewrites the file that carries it and a score keeps the better rewrites. We give the first measurement of marks under this attack, with three evolvers, five proposers and two environments. Published marks fail. KGW in the prompt survives GEPA's rewrites in 0 of 12 copies, AgentMark as released in 5 of 12. The proposer's habits decide survival, by two orders of magnitude across models, and selection protects a mark only when each decision sees enough games. We then build a mark for this attack. Its key is the order of near-tied entries of a table of where objects are found, built from public games, so deleting the table costs success, rewrites that keep its rows keep the order, and a customer reads the key from the agent's actions alone. On held-out games it is detected in 62 of 76 copies whose lineage rewrote its file. A scheme-aware plagiarist reshuffles the ties at no measurable cost, and a plagiarist who asks the proposer for more general code removes the key in 75-100% of rewrites.

cs.CR↗

arXiv2026

MotionCoach: Constructing Multimodal Guidance for Video Professional Motion Editing

Video editing models achieve strong performance in modifying appearance, backgrounds, styles, and motions, empowered by large-scale training data. Beyond general scenarios, professional motions are significant in performance, sports, and martial arts. This paper focuses on professional motion editing, which involves fine-grained motion technique and temporal structure. Given professional motion names but lacking detailed descriptions, existing models often fail to generate professional motions, such as incorrectly lifting the front wheel under a Stoppie instruction. To address this challenge, we introduce MotionCoach, a trained multimodal agent that uses multi-turn tools and performs self-verification, producing a grounded prompt and a reference clip as reliable and sufficient multimodal guidance to inject external professional motion knowledge into video editing models. To supply training data, we build a data pipeline to construct trajectories. We cold-start the agent on MotionCoach-SFT-10k and apply Group Relative Policy Optimization (GRPO) on MotionCoach-RL-2k with rewards on the motion guidance prompt, the motion outcome video, and the format. For evaluation, we introduce VPMEBench, comprising 314 professional motion editing tasks with fine-grained references and annotations. Extensive experiments show that MotionCoach improves the overall score from 55.1 to 65.6 on Seedance 2.0 and from 51.2 to 64.5 on MiniMax-H3. Learning to construct multimodal guidance enables precise professional motion editing without retraining video editing models.

cs.CV↗

arXiv2026

Beyond Speech Captions: Speech-Rewarded Style Planning for Conversational Text-to-Speech

Natural-language style descriptions provide an interpretable interface between large language models (LLMs) and controllable text-to-speech (TTS). However, using descriptions as pseudo-labels compresses target acoustics into text, and descriptive fidelity need not imply effective control of a particular synthesizer. We empirically show that speech-text alignment only weakly predicts downstream acoustic similarity among candidate instructions for the same utterance. We therefore propose Speech-Rewarded Style Planning (SRSP), which trains a text-based style planner through a frozen downstream TTS model. Given dialogue history and response text, the planner generates candidate instructions and is optimized with group-relative policy optimization (GRPO), using the teacher-forced likelihood of target speech tokens as the reward. On an English subset of the ISCSLP 2026 CoT-TTS corpus, SRSP achieves higher speech-style and emotion similarity to target speech and lower mel-cepstral distortion than the Base LLM and target-audio-informed captioning baselines. LLM-based expressive speech evaluation further shows gains over all baselines in contextual appropriateness and reference consistency.

cs.SD↗

arXiv2026

PixelDense: Dense Prediction as Representation Alignment for Pixel Diffusion

Representation alignment (REPA) accelerates diffusion transformer training, but its alignment targets are almost exclusively semantic encoders such as DINOv2 and CLIP. Recent analysis points to spatial structure, not global semantics, as the carrier of the alignment effect, yet dense-prediction foundation models trained to predict that structure remain overlooked as REPA targets. In pixel-space diffusion, SAM2, Depth Anything v2, and Metric3D v2 each outperform the DINOv2-only GenEval baseline, with the two geometric teachers leading the segmentation teacher. A flat sum of all four teachers, however, lands below the best single geometric teacher, as semantic and geometric gradients compete for one denoiser projection. We introduce PixelDense, which routes DINOv2 and SAM2 through a semantic projection stream, routes Depth Anything v2 and Metric3D v2 through a geometric projection stream, and adds a weight-space orthogonality penalty that keeps the two streams in disjoint subspaces. All four teachers are frozen during training and dropped at inference. Applied to PixelGen and DeCo with a single recipe, PixelDense improves GenEval, DPG-Bench, and HPS v2.1, raises PixelGen-XXL's GenEval Overall from 0.7927 to 0.8093, and beats every single-teacher and unfactored multi-teacher variant. In partial-noise reconstruction, independent panoptic, depth, and surface-normal probes show up to 53.1% PQ gain and 36.0% depth AbsRel reduction at $τ=0.5$ across COCO and Flickr30K. From random initialization, PixelDense also reaches the baseline's peak GenEval 1.23x faster. In SDEdit editing on PIE-Bench, PixelDense keeps more of the source background and layout at every edit strength, raising background PSNR by up to 2.2 dB.

cs.CV↗

arXiv2026

ReVR: Dual-Path Concept Reasoning for Multimodal Fake News Detection

Vision-language models (VLMs) support multimodal fake news detection (FND) by producing explicit analyses. Recent methods further improve interpretability by organizing verification knowledge into explicit concepts. However, two questions remain: how to improve the reliability and applicability of verification concepts, and how to effectively apply reusable concepts to verify unseen news. We propose \textbf{ReVR}, a dual-path reasoning framework that constructs and applies reusable verification concepts for multimodal fake news detection. An agentic workflow grounds and consolidates candidate concepts, while statistical profiles characterize their historical behavior. During inference, a coverage-oriented path aggregates evidence from the complete concept library using a trainable encoder, while a query-focused path prompts a frozen VLM to reason over selected concepts and their observations. A learned conflict resolver selects between the two predictions when they disagree. Experiments on fake news benchmarks demonstrate the effectiveness of the method regarding detection performance and generalizability.

cs.MM↗

arXiv2026

Contamination, Prior, or Evidence? Decomposing and Training Evidence Use in Whole-Slide Vision-Language Models

Pathology vision-language models (VLMs) are conventionally evaluated by accuracy, but accuracy alone does not measure evidence use: it may conflate dataset contamination, prior knowledge, and image evidence. In a motivating study of lymph-node metastasis prediction, we found that most public pathology VLMs showed minimal differences when changing from feeding the models with whole-slide images, an annotated lesion, or no image at all. To better understand the specific features leveraged by these models, this paper presents two contributions aimed at disentangling these factors. First, we present CleanSlide, a TCGA-based VQA benchmark designed to eliminate image- and question-side contamination. It contains 149K audited multiple-choice questions over 9,985 slides, with patient- and tissue-source-disjoint splits. Every question is audited for option shortcuts, stem leakage, cross-split duplication, and blind solvability. Second, we propose Pair-DPO, a preference loss over counterfactual slide pairs from the same question and source. By controlling for shared confounding factors, Pair-DPO cancels out the question-attributable signal and leaves image evidence as the source of preference. Specifically, each pair consists of two real slides with opposite, verified findings, introducing neither editing artifacts nor unverified labels for diffuse or graded features such as invasion, necrosis, and tumor grade. Experiments show that our method gains 15.29% from image evidence on the CleanSlide, compared with 2.81% for the best published model. On the external CPTAC and BCNB cohorts, our method achieves accuracies of 57.6% and 59.0%, outperforming all other evaluated models by 9.7% and 3.4%, respectively. We will release the benchmark and code.

cs.CV↗

arXiv2026

Retain-Free Machine Unlearning for Speech Emotion Recognition

Speech Emotion Recognition (SER) infers a speaker's emotional state from speech and is increasingly deployed in human-computer interaction, education, and healthcare. Because speech also carries sensitive personal information, speakers may ask that some of their recordings be deleted, which requires removing the influence of those samples from an already trained SER model. Most machine unlearning methods can meet this request only with access to the remaining training data alongside the samples to be forgotten; this is impractical when the remaining data cannot be redistributed or has itself been deleted, and it adds storage and computation as the data grows. To this end, we propose a retain-free unlearning method that updates a pre-trained SER model using only the forget set. Our key idea is to synthesise adversarial samples from the forget set as a surrogate for the unavailable remaining data, and to constrain each parameter update by its estimated importance so that forgetting does not erase general knowledge. The experiments over several emotional-speech corpora and self-supervised backbones show that our method drives forget-set performance down to near chance while retaining much of the model's utility on the remaining and unseen-speaker data, narrowing the gap to methods that rely on the remaining set.

cs.SD↗

arXiv2026

Ultra-fast Neural Inference for Stochastic Gaussian Splatting Denoising

Stochastic rendering eliminates the sorting and alpha blending process in Gaussian splatting, at the cost of introducing spatial noise. Formulating temporal denoising over the pixel stream shared by view-consistent stochastic splatting renderers, we propose a temporal neural denoiser validated on stochastic 2D Gaussian Splatting rendering, combining dual-path exponential moving average accumulation, per-pixel learned trust prediction for history validation, a fixed anisotropic spatial filter and a variance-gated composition with stabilization. The denoiser suppresses the noise, achieving temporally stable, visually compelling outputs during free camera navigation, all while retaining the sort-free, blend-free rasterization performance. The combined pipeline retains a PSNR gap to sorted alpha-blending renderers, but the denoiser's overhead stays below the time saved by removing sorting and blending.

cs.CV↗

arXiv2026

Attention-Steered Vision-Language Models for Sign Language Translation

Vision-language models (VLMs) have emerged as a powerful framework for multimodal video understanding. However, they remain limited in the sign language translation task, where we identify a key failure mode of existing VLMbased translators: poor spatial-temporal visual grounding. In particular, we find that standard next-token cross-entropy does not directly provide signal for where and when the model should attend, causing models to overlook sign-relevant regions and frames. To address this challenge, we propose AttnSign, a VLM-based spatial-temporal attention steering framework for sign language translation. AttnSign first introduces spatial attention supervision for sign-relevant regions, such as face and hands, in each frame; then develops an RL-based motion-cadence steering method that encourages the model to explore and focus on sign-level keyframes. Experimental results on How2Sign and OpenASL benchmarks show that our proposed AttnSign consistently outperforms existing methods.

cs.CV↗

arXiv2026

Collaborative Optimization of Multiclass Imbalanced Learning: Density-Aware and Region-Guided Boosting

Numerous studies on Boosting attempt to mitigate classification bias caused by class imbalance. However, existing studies have yet to explore the collaborative optimization of imbalanced learning and model training. This constraint hinders further performance improvements. To bridge this gap, this study proposes a collaborative optimization Boosting model of multiclass imbalanced learning. By integrating the density factor and the confidence factor, this model implements a noise-resistant weight update mechanism alongside a dynamic sampling strategy. Rather than functioning as independent components, these modules are tightly integrated to orchestrate weight updates, sample region partitioning, and region-guided sampling. Thus, this study proposes the collaborative optimization of imbalanced learning and model training. Extensive experiments on 40 public imbalanced datasets demonstrate that the proposed model significantly outperforms seven state-of-the-art baselines. The code and datasets for this paper are available at: https://github.com/ChuantaoLi/DARG.

cs.LG↗

arXiv2026

VLX-VR: An Agentic-Aware Video Reasoning Model

Real-world video understanding requires integrating visual, audio, textual, and temporal evidence distributed across a video. Yet many pipelines use a fixed video context and single-pass inference, limiting adaptive evidence acquisition when observations are incomplete, ambiguous, or conflicting. We present VLX-VR, an agentic-aware video reasoning model trained within a video reasoning framework defined by a Think--Memory--Observation loop. At each step, VLX-VR determines the needed evidence, invokes read_memory or write_memory, incorporates the returned Observation, and decides whether to continue or produce the task output. We train VLX-VR with multimodal data, including videos and agent trajectories, using reinforcement learning to learn evidence acquisition, memory use, and termination. On MINERVA, VLX-VR achieves state-of-the-art performance among the models included in our comparison, with 78.79% accuracy. Under the original three duration groups, its accuracies are 76.70%, 78.73%, and 80.92%, with a cross-duration accuracy variance of 2.97~$\mathrm{pp}^2$. On correctly answered samples, 96.20% of VLX-VR's reasoning traces are consistent with the MINERVA reference reasoning traces and the evidence described by them, while approximately 75.80% of all evaluated samples satisfy both answer correctness and this evidence-grounded trace criterion. These results show strong performance and broadly stable behavior across durations, while counting, state changes, causal reasoning, and spatial perception remain challenging.

cs.CL↗

arXiv2026

PTIR-GS: Path-Traced Inverse Rendering with Global Illumination in 3D Gaussian Fields

Ray tracing enables 3D Gaussian fields to serve as a representation for physically based light transport. Faithful inverse rendering requires forward rendering and backward optimization to be defined within a consistent light-transport pipeline. Existing Gaussian inverse-rendering methods typically rely on splatting-derived G-buffers and screen-space optimization. The local affine approximation of perspective projection can introduce inconsistencies with path tracing, while simplified rendering formulations often neglect or approximate indirect illumination and visibility. Therefore, we propose a splatting-free path-traced inverse-rendering framework for 3D Gaussian fields that unifies forward rendering and backward optimization within the same recursive multi-bounce light-transport pipeline. We formulate light transport over overlapping Gaussian primitives directly in path space, enabling Monte Carlo path tracing and pathwise gradient replay over the same recursively sampled transport paths. The framework jointly optimizes materials and environment light under the full rendering equation, with ray-traced visibility and global illumination explicitly evaluated. Extensive experiments demonstrate competitive material estimation and improved path-traced rendering quality, producing more plausible shadows, reflections, and relighting under global illumination.

cs.GR↗