Searcharxiv⌕ Search

arXiv subjects

Cong Wang

Publications and source records attributed to Cong Wang.

At least 55 records · Page 3Linked to original sources
arXiv2026

ActBench: Self-Evolving Benchmark of Behavioral Safety in Cowork Agents

Cowork agents may complete benign tasks while disclosing protected data, manipulating unauthorized state, invocate unauthorized API. We define behavioral safety and introduce ActBench, a self-evolving benchmark that evaluates such behavior risk from execution trajectories rather than final responses. Each case pairs a benign task with an adversarial variant that preserves its instruction, configuration, initial state, rating model, and trusted records while injecting a task-reachable payload. ActBench contains 600 cases from 213 scenarios, spanning 15 risk behaviors, six execution spaces, and 48 web-service APIs.To move beyond static payloads, we propose a reward-guided beam search method that jointly optimizes attack effectiveness and task utility, while reflection diagnoses failed execution checkpoint and guides payload revision. Besides, we propose a dual evidence verification mechanism that verifies agent execution safety and utility through log evidence and LLM-based trajectory evidence.We evaluate 15 LLMs and 6 open-source cowork agents over 24,000 trajectories. Under a fixed harness, attack success rates ranges from 10.1% to 94.4% across models, while under a fixed base model, they range from 73.7% to 94.4% across agents.These results show greater variation across models than agent harness, while attacks remain highly successful across all tested harnesses.Our benchmark is released at: https://github.com/zjuicsr/ActBench.

cs.CR↗
arXiv2026

EmbodiedVAE: Disentangled Video VAE for Efficient and Controllable Embodied Manipulation

Latent diffusion models (LDMs) have recently significantly advanced embodied learning in constructing powerful embodied manipulation world models. However, despite the remarkable performance, existing LDMs predominantly rely on Variational Autoencoders (VAEs) optimized for natural scenes while failing to account for the unique characteristics of embodied manipulation scenarios, yielding latent representations that are neither compact nor controllable, thereby hindering efficient training of LDMs and precise robotic control. To solve this problem, we present EmbodiedVAE, a novel video VAE that provides compact yet controllable latent representations tailored for the robotic manipulation world models. Specifically, EmbodiedVAE adopts a dual-encoder, single-decoder architecture with an asymmetric spatio-temporal compression module, which automatically disentangles the robot arm's motion from background environment, resulting in overall compactness while providing explicit embodied latent to support fine-grained action control. To further preserve the temporal consistency of learned robotic motion latent, we introduce an optimal-transport-based consistency module that explicitly enforces motion fidelity and inter-frame coherence. Extensive experiments demonstrate that our proposed EmbodiedVAE achieves superior reconstruction quality with high compression rate, while enabling more precise action control in robotic manipulation scenarios with an average of 2dB PSNR improvement over state-of-the-art video VAEs.

cs.RO↗
arXiv2026

Medical world models in healthcare: foundations, applications, and challenges for trustworthy clinical translation

Medical world models offer a framework for extending medical artificial intelligence beyond static prediction by representing evolving patient states and modelling how they change over time and in response to clinical interventions. This Review defines the conceptual boundaries, technical foundations, application domains, and evidence requirements of the field through a structured narrative synthesis with reproducible evidence mapping. We screened 1,455 unique records and assembled a corpus of 98 sources, including 14 studies that met a strict empirical definition of a medical world model. The field is organised around four capabilities: patient state representation, temporal dynamics modelling, intervention-conditioned simulation, and clinician-supervised planning. Evidence spans medical imaging, longitudinal electronic health records, treatment response modelling, physiological and multimodal state modelling, ultrasound and surgical interaction, and population and health-system simulation; clinical digital twins are treated as a cross-cutting integration framework. Current studies provide early evidence of technical feasibility for trajectory forecasting and comparison of candidate interventions, but most remain retrospective, task-specific, or preclinical. The evidence base is further limited by incomplete longitudinal intervention data, inconsistent action semantics, limited causal identifiability, long-horizon error accumulation, inadequate uncertainty estimation, and limited external validation. Clinical translation will therefore depend on precise intervention representations, robust causal and mechanistic grounding, calibrated trajectory-level uncertainty, safety-constrained planning, and prospective multicentre validation against clinically meaningful endpoints.

cs.CV↗
arXiv2026

CAGE: Certified Authorization under Typed-Return Uncertainty for Tool-Using Agents

Tool-using LLM agents act on typed tool returns, records pairing provenance and categorical fields with numerical values. Runtime permission gates generally authorize the observed return and action, leaving the decision unprotected against small errors in how the return was bound to its source. We ask whether a candidate action stays authorized over a declared neighborhood of plausible correctly bound returns: one admissible binding fault plus bounded numerical drift. We prove that certifying the categorical and numerical channels separately does not compose: perturbations that are safe on each channel alone can jointly turn the same action unsafe. CAGE certifies this joint neighborhood directly, enumerating the discrete branches exactly and certifying the continuous perturbation within each branch. Across synthetic, policy-as-code, regulatory, and real-transaction settings, CAGE removes the in-budget false allows that accurate pointwise gates admit, while keeping a useful fraction of decisions autonomous. When the policy is executable, CAGE-Exact certifies the policy itself; otherwise CAGE-Lip and CAGE-RS certify a learned gate under an explicit, measured fidelity assumption.

cs.AI↗
arXiv2026

Multi-Modal Object Re-Identification with Dual Semantic Guidance and Global-Local Mutual Modulation

Multi-modal object Re-Identification (ReID) aims to retrieve target instances by leveraging complementary information across modalities. However, existing methods suffer from two challenges. First, they often fail to exploit well-aligned and reliable semantic priors, making them vulnerable to background clutter and cross-modal misalignment. On the other hand, they typically rely on holistic feature modeling, overlooking the synergy between global and local representations. To overcome these limitations, we propose a robust multi-modal ReID framework with dual semantic guidance and global-local mutual modulation, which mainly consists of three key components, namely the Text-Semantic Injector (TSI), the Masked Global-Local Modulator (MGLM), and the Hierarchical MoE Fusion (HMF). The TSI enhances semantic awareness by integrating clean and coherent textual features into visual tokens. The MGLM enables part-aware cross-modal interaction through joint guidance from soft masks and global context, improving fine-grained feature alignment. Finally, the HMF adaptively aggregates multi-spectral features under local semantic supervision, yielding discriminative and robust representations. Extensive experiments on three multi-modal ReID benchmarks demonstrate the effectiveness of the proposed method. The code will be made publicly available at https://github.com/zw-absin/DSGM upon acceptance.

cs.CV↗
arXiv2026

AgentHOI: Multi-Agent Reasoning for Human-Object-Interaction Video Generation via Implicit Representation Alignment

Recent advances in video diffusion models have spurred interest in human-object interaction (HOI) video generation, which demands fine-grained control over interaction logic beyond single-subject animation. However, existing HOI methods rely heavily on explicit motion control, limiting scalability and generalization across diverse objects and interactions. In this study, we propose AgentHOI, a text-driven HOI video generation following a thinking-before-generation framework that bridges the gap between high-level textual intent and physical execution through multi-agent reasoning over perception, interaction, and motion planning. Building upon the generated interaction plans, we further strengthen text-driven motion understanding. We introduce an implicit text-motion alignment strategy that distills text-to-motion priors into the video diffusion model, enabling robust HOI synthesis without explicit motion inputs at inference. Experiments show that AgentHOI significantly improves interaction naturalness, object appearance preservation, and adherence to complex textual instructions across challenging object-centric scenarios such as wearing and riding. The code is available at https://github.com/bone-11/agenthoi.

cs.CV↗
arXiv2026

Agentic AI in medicine: architectures, applications, evaluation, and challenges for clinical translation

Large language models and multimodal foundation models are enabling medical artificial intelligence (AI) systems to move beyond isolated prediction and undertake multistep clinical tasks that require planning, tool use, memory, iterative correction, and coordination among specialized agents. However, the scope of agentic AI in medicine remains unsettled, and current evaluation practices are not yet aligned with the requirements of clinical use. We conducted a scoping review with systematic evidence mapping across five electronic sources, screened 1,649 exportable records, and provisionally included 557 unique studies that met predefined criteria for goal-directed task execution, tool use, interaction with external resources, feedback-based refinement, or multi-agent collaboration. The included studies describe single agents that use external tools, workflows supported by retrieval and external knowledge, multimodal agents, and multi-agent systems applied to medical question answering, image interpretation, electronic health record analysis, drug safety, and clinical trial prediction. The evidence base remains dominated by public benchmarks, simulated settings, retrospective datasets, and small-scale expert evaluation. Process reliability, evidence traceability, uncertainty, safety, workflow impact, and external validity are evaluated less consistently. Clinical translation will depend on clearer definitions, reproducible evaluation, auditable oversight, interoperable system design, and prospective validation in real-world clinical workflows.

cs.CV↗
arXiv2026

KernelScript: Cross-Boundary Typed DSL for eBPF Applications

eBPF lets developers extend Linux with custom packet processing, tracing, and scheduling logic, and a verifier proves before execution that the code will not crash the kernel. The programming model, however, is fragmented: a single application spans kernel code, a userspace loader, and shared maps, yet the relationships among these pieces go unchecked. E.g. A map or event type defined differently on each side silently corrupts shared state. We observe that these cross-boundary relationships duplicate information that a type system can unify. We present KernelScript, a DSL that types maps, program handles, and execution domains in one source, then compiles to standard C through the original toolchain. We evaluate KernelScript on 43 eBPF workloads covering XDP, TC, kprobe, tracepoint, and struct_ops. KernelScript rejects cross-boundary bugs at compile time that standard C/libbpf still builds and loads, a unified source shrinks the diffs for cross-boundary changes by 5x, and generated code remains compatible with the existing toolchain.

cs.PL↗
arXiv2026

Multi-Modal Object Re-Identification with Prompt-S6 and Semantic-Aware Knowledge Guidance

Multi-modal object Re-Identification (ReID) aims to retrieve specific objects by integrating complementary information from multiple modalities. However, existing multi-modal ReID methods do not effectively address background interference suppression or achieve tri-modal alignment, instead focusing on pairwise feature fusion. Moreover, many current aggregation approaches suffer from high computational complexity. To address these limitations, we propose PRISM, a novel multi-modal ReID framework built upon Prompt-S6 (PS6) and semantic-aware knowledge guidance. PS6 maintains the linear complexity and strong sequence modeling capability of Mamba while enabling efficient cross-modal interaction. Leveraging these advantages, we design two key components: Semantic-Driven Token Pruning (SDTP) and Progressive Fusion Network (PFN). Parsing semantic priors from the segmentation foundation models, the SDTP then leverages these priors and applies dynamic token pruning to suppress background noise and refine feature representations. The PFN progressively aggregates multi-modal features to achieve tri-modal alignment and fully exploit modality complementarity. With the proposed modules, PRISM generates more robust multi-modal representations under complex scenarios. Extensive experiments on four multi-modal object ReID benchmarks demonstrate the effectiveness and efficiency of our approach. The source code is available at https://github.com/zw-absin/PRISM.

cs.CV↗
arXiv2026

Radiative decays $J/ψ,\,ψ(2S)\rightarrowγη^{(\prime)}$ in perturbative QCD with relativistic corrections

We present the first calculation of the radiative decays $J/ψ,ψ(2S)\rightarrowγη^{(\prime)}$ in perturbative QCD that includes the order-$q^{2}$ relativistic corrections in all three short-distance contributions, namely the quark-antiquark, two-gluon, and QED contributions. The amplitudes are found to be remarkably insensitive to the light-cone distribution amplitude and to the light-quark mass, a robustness that persists through order $q^{2}$ and makes the predictions correspondingly reliable. The relativistic correction enhances the $J/ψ$ branching ratios by roughly a factor of two, narrowing their shortfall from experiment, whereas for the $ψ(2S)$ it is about twice as large as for the $J/ψ$ and the low-order expansion converges poorly. In two representative $η$--$η'$ mixing schemes, the ratio $\mathcal{R}_{1S}=\mathcal{B}(γη')/\mathcal{B}(γη)$ proves sharply sensitive to the mixing angle and favours the smaller of the two. The predicted $ψ(2S)$ rates lie well above the data in both channels, already at leading order, and most severely for the anomalously small $γη$ channel. Such a discrepancy suggests that a mechanism beyond the hard perturbative process is at work. As a physically motivated attempt, we explore the $η_{c}$-mixing contribution, which adds coherently to the perturbative one and is comparable to it in the $γη$ channel, and find that the interference can bring the $ψ(2S)$ rates into agreement with the data, although its extraction is limited by a strong sensitivity to the mixing parameters.

hep-ph↗
arXiv2026

Token-Operations-Oriented Inference Optimization Techniques for Large Models

Large model inference optimization serves as a key foundation for supporting the scalable, low-cost, and highly stable operation of large model services. Centered on token-oriented inference optimization technology, this paper proposes for the first time a four-layer technical architecture consisting of Multi-model Fusion, Model Optimization, Compute-Model Fusion, and Compute-Network-Model Fusion. It systematically reviews the key technologies and current industry status across these four levels and analyzes the application value of related technologies in real-world business scenarios. This paper provides a practical technical path for reducing token production costs, improving token service efficiency, ensuring the stability of token supply, and driving the transition of large model services from being merely callable to being operable.

cs.SE↗
arXiv2026

How Agentic Is Agentic Commerce? A Population-Scale Measurement of x402 Adoption and Authenticity

AI agents are said to be forming an economy in which they pay, on their own, for the data, APIs, and compute they consume. x402, which settles a stablecoin payment on-chain for each purchase, is the most widely deployed protocol for this, and its hundreds of millions of settlements are read as proof that the economy has arrived. We show the count cannot be read as adoption: it is the one metric an interested party can manufacture almost for free, since the facilitator sponsors the gas and nothing on-chain marks who controls a payment. We give the first population-scale measurement of x402 on Base, supplemented with a coarser Solana census. Identifying settlements from their on-chain event and resolving the true payer through the meta-transaction layer, we sort each by what its trace can prove via a payment graph. Over a 280-day window Base carries 136{,}708{,}672 settlements worth \$44{,}121{,}383.81, concentrated on every axis we measure (payer, recipient, and value Gini all above 0.98), yet 21.20\% are fictitious and 63.78\% internal settlement within a linked cluster. What is genuinely independent is bounded: it lies between the \$187{,}861.35 that demonstrably reaches a nameable service and the \$20{,}258{,}746.09 (45.92\% of value) not provably manufactured. Finally, we resolve the count's manufacturable component, a coherent operator-driven economy, star-shaped, machine-timed, and gas-subsidized. Settlement count measures manufacturability, not adoption.

cs.CR↗
arXiv2026

BUS: Brain-Inspired Unsupervised Self-Reflection via Backward Prediction for Multimodal Reasoning

Current Vision-Language Models (VLMs) often struggle to handle complex visual tasks that require consistent and fine-grained reasoning. Recent methods aim to train models to facilitate self-reflective reasoning, i.e., reviewing and improving the generated reasoning. However, they require large volumes of annotated data and lack explicit reflective behavior during test time. By contrast, humans perform explicit and efficient self-reflection through mechanisms such as backward prediction, i.e., predicting which current states are likely to precede a given future state. Inspired by neuroscience, this work proposes a novel solution to address these challenges. We first observe and investigate the phenomenon that mainstream VLMs can perform backward prediction, similar to the human brain. A label-free training framework named Brain-inspired Unsupervised Self-reflection (BUS) is proposed to leverage and exploit backward prediction capability to enhance reflective reasoning in complex visual tasks. BUS enables self-verification of reflective reasoning based on backward prediction, providing explicit learning signals under unsupervised conditions. In this way, BUS eliminates reliance on annotated data while improving reasoning performance. Designed as a model-agnostic plug-in, our framework is compatible with popular fine-tuning methods, such as Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL). Initialized from Qwen3-VL-8B, it improves HR-Bench-8K (+8.0%), HR-Bench-4K (+7.7%), V* Bench (+6.3%), and MME-RealWorld-Lite (+5.8%), proving backward prediction is key to advancing reflective reasoning.

cs.CV↗
arXiv2026

MambaLIE: Scene Light Intensity-Boosted Low-Light Image Enhancement with State Space Model

Images captured by consumer electronic devices, such as mobile phones and digital cameras, often suffer from low-light degradation due to sensor limitations and imaging pipelines, which degrades visual quality and affects downstream vision tasks. Existing methods based on Convolutional Neural Networks (CNNs) and Transformers have dominated current low-light image enhancement (LIE) due to their excellent ability to model hierarchical features. However, CNNs operate in local receptive fields that cannot model long-range dependencies, while Transformers overcome this problem but incur substantial computational costs. To address these challenges, we propose MambaLIE, a Scene Light Intensity-Boosted Low-Light Image Enhancement method based on a State Space Model (SSM). We first introduce scene light intensity to improve the structural distribution of illumination, which is then gated with the low-light input to guide enhancement. To better model the illumination while maintaining computational efficiency, we propose the Locally Enhanced State Space Model (LESSM) for efficient light enhancement. Our LESSM contains two branches: an SSM branch and a Local Enhanced branch, where the former is used to model the long-range dependencies with linear time complexity, while the latter is used to enhance local feature representations. Extensive experiments demonstrate that MambaLIE outperforms state-of-the-art CNN-based and Transformer-based LIE methods on four widely used synthetic benchmarks and five publicly available real-world benchmarks in terms of accuracy, speed, and model size, making it suitable for practical deployment on resource-constrained devices.

cs.CV↗
arXiv2026

Resonant-impurity scanning tunneling spectroscopy in altermagnets: dual Fano resonance and Landau-quantization-induced nodal spin contrast

Using a Green's-function formalism, we study the spin-resolved local spectral function of a resonant impurity coupled to a two-dimensional $d$% -wave altermagnetic substrate. It is found that the interplay between direct tunneling from the impurity to the scanning tunneling microscopy (STM) tip and altermagnet-mediated tunneling gives rise to a dual Fano resonance in the absence of an external magnetic field. Moreover, the anisotropic spin-dependent oscillations of the local density of states and the corresponding Fano factors provide information on the altermagnetic splitting strength from complementary local and global perspectives. In addition, spin-selective tunneling can be achieved by tuning the Fermi energy and the tip position. In the presence of a strong magnetic field with Landau-level quantization, the dominant scanning tunneling spectroscopy (STS) signature appears as a spin-dependent nodal structure in real space: the nodal mismatch between opposite spin channels produces a large local spin contrast. These results establish resonant-impurity STM/STS as a phase-sensitive local probe of altermagnetic band anisotropy.

cond-mat.mes-hall↗
arXiv2026

HART: High-Resolution Annotation-Free Reasoning Technique through a Closed-loop Framework

Current Large Multimodal Models (LMMs) struggle with high-resolution visual inputs during the reasoning process, as the number of image tokens increases quadratically with resolution, introducing substantial redundancy and irrelevant information. A common practice is to identify key image regions and refer to their high-resolution counterparts during reasoning, typically trained with external visual supervision. However, such visual supervision cues require costly grounding labels from human annotators. Meanwhile, it remains an open question how to enhance a model's grounding abilities to support reasoning without relying on additional annotations. In this paper, we propose High-resolution Annotation-free Reasoning Technique (HART), a closed-loop framework that enables LMMs to focus on and self-verify key regions of high-resolution visual inputs. HART incorporates a post-training paradigm in which we design Advantage Preference Group Relative Policy Optimization (AP-GRPO) to encourage accurate localization of key regions without external visual annotations. Notably, HART provides explainable reasoning pathways and enables efficient optimization of localization. Extensive experiments on MME-RealWorld-Lite, TreeBench, V* Bench, HR-Bench-4K/8K, and MMStar demonstrate that HART improves performance across a wide range of high-resolution visual tasks, consistently outperforming strong baselines.

cs.CV↗
arXiv2026

Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment

Subject-driven and multi-element video generation are central to controllable video synthesis, but existing methods still struggle to preserve identity consistency and model complex relationships among multiple subjects. In this paper, we propose Aura, a unified framework for high-fidelity and identity-consistent video generation. To better capture scene dynamics and subject interactions, we introduce AI director-level captions that provide dense and structured descriptions of video content. We further leverage a vision-language model (VLM) with learnable queries to extract multimodal semantic features from textual and visual references, covering both global semantics and fine-grained visual cues. To bridge the representational gap between the VLM and the Diffusion Transformer (DiT), we design a two-stage alignment strategy that progressively maps VLM features into the DiT feature space. For visual conditioning, we adopt token concatenation to inject reference information directly into the generation process. To distinguish heterogeneous subject types and reduce common copy-paste artifacts, we develop a subject-aware RoPE-Shift mechanism. To further differentiate reference images of different categories, we introduce subject-aware learnable tokens. In addition, we introduce Memory Tokens to balance the training signal across examples with different numbers of reference subjects. During inference, Progressive-APG (Adaptive Prompt Guidance) further alleviates oversaturation and improves semantic alignment with user prompts. Finally, we build a high-quality video-subject image dataset through a dedicated data construction pipeline. Extensive experiments show that our method achieves state-of-the-art performance on both single-subject generation and more challenging multi-element scenarios.

cs.CV↗
arXiv2026

Goku: A Million-Scale Universal Dataset and Benchmark for Instruction-Based Video Editing

Existing instruction-based video editing datasets commonly focus on single-task appearance editing, failing to meet the complex creative demands of real-world scenarios. To bridge this gap, we present Goku, a large-scale dataset featuring 2 million high-quality, instruction-aligned video editing pairs, which is the first to extend task boundaries from basic appearance editing to multi-task and structural manipulations(e.g., precise control of subject movement). To tackle the data synthesis challenges inherent in these complex tasks, we design an efficient data synthesis pipeline that decomposes complex edits into controllable sub-problems and introduce a progressive filtering system for data reliability throughout the whole process. Furthermore, we explore the optimal network structures on Goku, and propose Goku-Edit. To deeply comprehend complex editing instructions, Goku-Edit leverages an MLLM as its text encoder and adopts a decoupled dual-branch design: a dedicated mask branch handles structural control, freeing the main branch for appearance rendering. A comprehensive video editing benchmark, Goku-Bench, is also proposed with 1,000 human-verified test cases and 7 novel editing-specific metrics. Evaluated on Goku-Bench, Goku-Edit obtains up to +8% improvement on other open-source models in terms of instruction following.

cs.CV↗