Searcharxiv⌕ Search

arXiv subjects

Wei Li

Publications and source records attributed to Wei Li.

At least 73 records · Page 4Linked to original sources
arXiv2026

AgenticGen: Reward-Guided Agentic Video Generation for Advertising

Advertising video generation is not only a video synthesis task, but also a product-conditioned reasoning problem whose success is measured by online business metrics. Recent video foundation models can generate realistic clips from multimodal conditions, yet they do not optimize how a product should be transformed into an effective advertisement or how future generation should be improved from online business feedback. To close this loop, we propose AgenticGen, a reward-guided agentic framework that decomposes advertising video generation into two trainable reasoning stages, strategy selection and draft generation, thereby exposing optimization targets that online business feedback can supervise. AgenticGen learns a performance-based reward from accumulated online feedback and a complementary rubric-based reward aligned with human quality standards, then uses them to supervise policy optimization. DPO first moves the agentic policies toward online preferences, and GRPO further refines both stages with process and outcome rewards. Offline experiments validate the reward models and successive policy optimization. Online A/B experiments in the TikTok advertising system show that AgenticGen after DPO and GRPO improves CTR by 2.72%, CVR by 2.63%, and Advv by 9.61% over the SFT baseline.

cs.CV↗
arXiv2026

Large VLM-based Vision-Language-Action Models for Robotic Manipulation: A Survey

Robotic manipulation, a key frontier in robotics and embodied AI, requires precise motor control and multimodal understanding, yet traditional rule-based methods fail to scale or generalize in unstructured, novel environments. In recent years, Vision-Language-Action (VLA) models, built upon Large Vision-Language Models (VLMs) pretrained on vast image-text datasets, have emerged as a transformative paradigm. This survey provides the first systematic, taxonomy-oriented review of large VLM-based VLA models for robotic manipulation. We begin by clearly defining large VLM-based VLA models and delineating two principal architectural paradigms: (1) monolithic models, encompassing single-system and dual-system designs with differing levels of integration; and (2) hierarchical models, which explicitly decouple planning from execution via interpretable intermediate representations. Building on this foundation, we present an in-depth examination of large VLM-based VLA models: (1) integration with advanced domains, including reinforcement learning, training-free optimization, learning from human videos, and world model integration; (2) synthesis of distinctive characteristics, consolidating architectural traits, operational strengths, and the datasets and benchmarks that support their development; (3) identification of promising directions, including memory mechanisms, 4D perception, efficient adaptation, multi-agent cooperation, and other emerging capabilities. This survey consolidates recent advances to resolve inconsistencies in existing taxonomies, mitigate research fragmentation, and fill a critical gap through the systematic integration of studies at the intersection of large VLMs and robotic manipulation. We provide a regularly updated project page to document ongoing progress: https://github.com/JiuTian-VL/Large-VLM-based-VLA-for-Robotic-Manipulation

cs.RO↗
arXiv2026

Riemann-1.0: An Embodied World Action Model for Physical AI

We introduce Riemann-1.0, a fully causal autoregressive World Action Model for embodied intelligence. Riemann-1.0 jointly models multi-view visual observations, robot states, and embodiment-specific actions within a unified causal autoregressive sequence, representing robot actions and world evolution as causal state transitions. Unlike existing WAMs based on joint generation, video-first prediction, or decoupled modeling paradigms, Riemann-1.0 unifies online robot policy execution and action-conditioned world simulation within a single model, enabling it to function as both an executable robot policy and a multi-embodiment visual world simulator. To scale embodied experience across heterogeneous data sources, we further develop a progressive embodied pretraining framework that unifies learning from egocentric human videos, handheld-gripper demonstrations, and heterogeneous robot trajectories under a shared World Action Modeling objective. Built upon 200K+ hours of interaction data, Riemann-1.0 progressively transfers large-scale embodied experience into executable robot manipulation capabilities. Riemann-1.0 achieves state-of-the-art performance across both simulation benchmarks and real-world manipulation tasks. It achieves success rates of 94.3% on RoboTwin2.0, 99.0% on LIBERO, and 62.6% on the long-horizon compositional benchmark RoboCasa-365, outperforming the previous best method by 8.4% On long-horizon real-world manipulation tasks, Riemann-1.0 achieves a Success Rate (SR) of 85.0% and a Progress Success Rate (PSR) of 94.4%, exceeding the strongest open-source baseline by 15% in SR. These results demonstrate that unified World Action Modeling together with progressive embodied pretraining effectively transforms large-scale embodied experience into generalizable robot manipulation capabilities.

cs.RO↗
arXiv2026

WiCi: Wireless GPU Computing Infrastructure

LLM inference applications are gaining significant traction. The demand for inference is growing exponentially, and the GPU usage of inference is increasingly surpassing that of training. Due to the mobility penalty, edge-side inference fails to deliver satisfactory performance. Consequently, most inference service providers currently rely on cloud-based inference, which incurs substantial, not sustainable costs for enterprises, and is even increasing in the agentic paradigm. Therefore, our goal is to enable powerful computing capabilities as server-grade GPUs on mobile devices. We propose Wireless GPU Computing Infrastructure (WiCi) in this paper. Through WiCi, mobile devices can wirelessly access server-grade GPUs, running inference tasks on mobile clients but offloading GPU-related computations to a nearby GPU via WiFi. WiCi introduces a series of designs to make sure the infrastructure is scalable with different applications, compatible with different mobile devices, and has comparable performance to running on a physical GPU. We test WiCi from mobile devices and find that WiCi can reduce time to first token by up to 90%, improve the token rate by approximately 39x compared to local inference on mobile devices for the same model, and support much larger models. WiCi also achieves up to nearly 80% of the native performance of the server-grade GPU across different applications.

cs.NI↗
arXiv2026

Rethinking the Transferable Adversarial Attacks and Robust Defense in Federated Learning

The development of federated learning (FL) techniques has helped improve the privacy preservation of users' data and extended the applications of machine learning models. However, the involvement of a large number of users in FL also creates open opportunities for different adversaries, such as poisoning attacks, Byzantine attacks, and adversarial example attacks. Yet, recent research has disclosed that existing poisoning attacks and Byzantine attacks can not achieve satisfactory penetration in realistic FL scenarios caused by strong assumptions, \textit{e.g.,} client selection rate, and the ratio of malicious attackers. In this paper, the transferability of adversarial examples among different client models is analyzed to understand the relation between adversarial examples and clients' data distribution. Moreover, to mitigate the attacks of transferable adversarial examples, we design a defense mechanism stemming from the transferability of model robustness by adversarial training. As a result, through theoretical analysis of transferability, we gain insights into adversarial examples and the vulnerability of federated learning systems. Our proposed adversarial attack and defense methods are evaluated via real-life datasets in various settings to show their performance over the existing state-of-the-art methods.

cs.LG↗
arXiv2026

DART: Dual-Axis Airborne Reachability-Gated Torque-Reaction for Off-Road Vehicle Jumps

Traversing crests, ledges, and ditches at high speed often launches vehicles into the air, and a mishandled landing presents a substantial crash hazard. We show that the airborne phase is barely controllable: on a 1383 kg platform the wheel angular-momentum budget caps the recoverable pitch-rate change at roughly $9$-$13^\circ$/s in the tighter nose-up direction under drive at typical takeoff wheel speeds, and at about twice that in the reverse-inclusive braking direction; driving the wheels to their drivetrain hard limit raises the measured nose-up ceiling to only $16$-$18^\circ$/s. Takeoff pitch-rate disturbances beyond this directional budget are physically unrecoverable in flight, so the decisive leverage lies before takeoff. DART (Dual-Axis Airborne Reachability-Gated Torque-Reaction) back-propagates the landing constraint into a closed-form certified feasible-takeoff set, which supplies a conservative go/no-go condition and a pre-takeoff speed-shaping law. In flight, DART regulates pitch and roll via steer-resolved wheel-reaction torque, governed by a per-flight roll latch derived from the yaw-coupling analysis. In deterministic full-scale simulation in BeamNG.tech, a calibrated pre-takeoff speed regulator reduces touchdown speed by 36% and raises on-target landings from 0/30 to 30/30. Under the same steep-lip approach the airborne law completes 29/30 safe landings under crash-avoidance bounds versus 0/30 for reaction-wheel-style PD (RW-PD) and time-optimal bang-bang (TOBB). On banked run-ups DART holds the median pitch error at or below $2^\circ$ at every cross-slope, with the largest baseline separation at $γ=12^\circ$. Across disturbance regimes, the latch preserves pitch-only allocation on low-disturbance entries and enables dual-axis control when roll becomes binding. All results are from simulation; hardware validation remains open.

cs.RO↗
arXiv2026

MinerU.Chem: A High-Precision System for Optical Chemical Structure and Reaction Recognition

In organic chemistry papers and patents, molecular structures, reaction schemes, and experimental conditions are often presented as molecular structure depictions, reaction diagrams, and complex tables or figures. Such information is difficult for general-purpose document parsing systems to directly convert into machine-readable data. This limits data production for organic chemistry knowledge base construction and for AI for Chemistry tasks such as reaction prediction, retrosynthesis, condition recommendation, molecular property prediction, and drug molecule design. This report introduces MinerU-Chem, a document parsing system for organic chemistry literature integrated into the MinerU online platform. Built on top of MinerU's general document parsing pipeline, MinerU-Chem adds five chemistry-specific modules: chemistry relevance filtering, molecular structure detection, molecule identifier extraction, molecular structure recognition, and reaction scheme parsing. Together, these modules convert organic-chemistry-related image regions in documents into a Molecule Summary List and a Reaction Summary List. For molecular structure recognition, MinerU-Chem uses CARBON (Complex Atomic Representation and Bonding Object Notation) as its core representation. CARBON enables recognition results to preserve both the visual layout of the original image and complex chemical semantics, while supporting the export of standard downstream formats such as MolFile and SMILES. On the SMILES-evaluable subset of MolRecBench-Wild (N=2,392), MinerU-Chem's molecular structure recognition module achieves a SMILES exact-match accuracy of 93.02%, outperforming the best evaluated comparison system, GPT-5.6-Sol (74.87%), by 18.15 percentage points. The system has been integrated into the MinerU online platform and is available at https://mineru.net/OpenSourceTools/Extractor .

cs.CV↗
arXiv2026

DART-S: Reachability-Audited Active-Suspension Preconditioning for Off-Road Vehicle Jumps

Airborne torque reaction cannot recover takeoff errors beyond the wheel angular-momentum budget. DART-S applies ramp-face suspension preconditioning to change pitch, pitch rate, and wheel spin before liftoff, thereby shifting the queried state and altering the remaining authority budget. To predict how each suspension action reshapes this state-budget pair, DART-S employs a local calibration map. A support-aware selector combines the predicted shift with local outcome evidence and an interval-reachability screen; an exact-pair audit reports residual authority. Across 600 new runs in 72 independent BeamNG sessions, every positive, negative, and boundary query follows its prespecified branch. At the confirmed 40°/13 m/s boundary, DART-S attains 24/24 post-touchdown attitude-criterion successes versus 0/24 for DART (session-level Holm-adjusted p=0.0234). At 11.5 m/s, a 0.35 s timing action attains 23/24 versus 0/24 for the static preset (p=0.0156). The 200 rad/s command guard keeps drivetrain hard-limit exceedance at zero across all 600 runs. The source code will be available at https://github.com/MeridianCAS/DART-S

cs.RO↗
arXiv2026

On two conjectures concerning Kemeny's constant of graphs

Kemeny's constant for a connected graph $G$, denoted by $\mathcal{K}(G)$, is the expected time for a random walk to reach a randomly chosen vertex $u$, regardless of the choice of the initial vertex. Recently, Kim et al. (2026) proposed two conjectures on Kemeny's constant. The first conjecture asserts that if $G$ is a connected graph of order $n$ and diameter 2, then $\mathcal{K}(G) = O(n)$. The second conjecture asserts that if $G$ be a graph of order $n$, then $\min\{\mathcal{K}(G), \mathcal{K}(\overline{G})\} = O(n)$, and if both $G$ and $\overline{G}$ are connected, then $\mathcal{K}(G)\mathcal{K}(\overline{G}) = O(n^4)$, where $\overline{G}$ denotes the complement of $G$. In this paper, we confirm both conjectures. For the first conjecture, we prove that if $G$ is a connected graph of order $n$ and diameter 2, then \[ \mathcal{K}(G) \leq (3 + \sqrt{5})(n - 1). \] For the second conjecture, we prove that for any $n$-vertex graph $G$, \[ \min\{\mathcal{K}(G), \mathcal{K}(\overline{G})\} \leq (8 + 2\sqrt{5})n - (10 + 2\sqrt{5}). \] Moreover, if both $G$ and $\overline{G}$ are connected, then \[ \mathcal{K}(G)\mathcal{K}(\overline{G}) \leq \frac{3 + \sqrt{5}}{2}n^4. \] Our proof relies on effective estimates on resistance distances and spectral gaps of graphs.

math.CO↗
arXiv2026

Reprojection-Guided 3D Gaussian Splatting Diffusion for Weakly Supervised Single-Image Normal Estimation

We propose CLONE, a Continuous Latent Optimization framework for Normal Estimation via 3D Gaussian splatting. The core idea is to construct an image-geometry-image consistency strategy that unifies explicit geometric representation with differentiable rendering, thereby enabling weakly supervised learning without normal ground truth. Specifically, CLONE comprises four components. First, by introducing a differentiable light interaction model with a learnable modulation kernel, we perform a unified reparameterization of the 3DGS parameter space. Second, the conditional single-step deterministic refinement network integrates denoising architectures with differentiable reprojection constraints to refine the initial normals, thereby adaptively recovering the high-frequency details erased by the inherently smooth Gaussian primitives. Third, the cross-domain gating fusion mechanism adaptively combines the two complementary normal estimates, reconciling the geometrically consistent yet over-smooth 3DGS estimate with the detailed yet potentially geometry-inconsistent refinement. Finally, all components are jointly optimized under a unified photometric reprojection objective with geometric consistency regularizations in a fully differentiable pathway, achieving an end-to-end optimization closed loop without relying on external normal labels.

cs.CV↗
arXiv2026

Discovering Physically Interpretable Mathematical Expression for Predicting CO2 Adsorption in Metal-Organic Frameworks via Machine Learning-Symbolic Regression

This work presents a machine learning-symbolic regression (ML-SR) strategy to develop a physically interpretable formula for predicting low pressure CO2 adsorption capacity in hypothetical metal-organic frameworks (hMOFs). Four ML models were trained on a small dataset of 1,000 samples, and five key descriptors-largest cavity diameter, pore limiting diameter, void fraction, gravimetric surface area, and number of hydrogen atoms-were identified through SHAP and feature importance analyses. Symbolic regression was then employed to derive a concise adsorption formula, Q=aA, where a represents an adsorption baseline (mmol/g) and A is a dimensionless adsorption number incorporating four structural descriptors. We interpret A as the ratio between an adsorption binding force and a diffusion driving force, revealing how pore topology and surface chemistry jointly influence adsorption. Validation against a comprehensive dataset of 137,652 hMOFs demonstrates that this formula achieves over 70% prediction accuracy for 62,448 structures, confirming strong applicability within defined structural and operational ranges. Unlike conventional black box ML models, the proposed physics-guided expression enables efficient prediction and provides clearer insight into adsorption mechanisms.

cond-mat.mtrl-sci↗
arXiv2026

Continual Learning in Transition

Classical continual learning (CL) has primarily focused on enabling models to update and retain knowledge through parameter-centric mechanisms, e.g., training strategies, architectural designs, and weight adaptation. However, emerging paradigms are reshaping the scope of CL beyond this traditional model adaptation view. For instance, on-policy learning broadens the space of update mechanisms; test-time training extends CL from the training phase to inference; and external harness components such as memory, skill libraries, and interaction protocols extend the evolutionary boundaries of model capabilities far beyond the static parameter space. Collectively, these developments indicate a transition from parameter-centric learning toward system-level adaptation. To characterize this transition, we examine the evolution of continual learning through three dimensions: When, How, and Where learning occurs. The How dimension encompasses off-policy, on-policy, and beyond-gradient optimization mechanics. The When dimension captures evolution across pre-training, post-training, and inference-time stages. The Where dimension delineates updates occurring within internal parameters versus external structural constraints. Anchored by this tri-axial framework, we systematically survey representative methods, trace the ongoing transition of continual learning, and discuss the key challenges, broader implications, and future directions arising from this paradigm shift.

cs.LG↗
arXiv2026

Beyond Reconstruction: Full-Context Generative DiT for Music Generation

Hybrid music generators combine the long-range planning of an autoregressive language model with the fidelity of a diffusion- or flow-based acoustic renderer. Yet renderers are trained with clean, target-derived codec tokens but deployed with imperfect language-model predictions, creating codecinterface exposure bias. Rather than treating rendering as a simple reconstruction task,we formulate it as full-context generation from an imperfect discrete plan. We introduce FullDiT, a conditional DiT that fuses eight frame-aligned RVQ streams with independently encoded captions and lyrics and uses non-causal self-attention over the complete acoustic latent sequence. During training, Error-Matched Distractor Conditioning (EMDC) matches per-codebook replacement rates to teacher-forced top-1 error rates and samples near-miss tokens from cosine-KNN neighborhoods without changing the acoustic target. At inference, four-way classifier-free guidance (4-CFG) independently scales codec, lyric, and caption guidance increments. Matched ablations show that EMDC improves ViSQOL by 0.77 under synthetic corruption and is clearly preferred in non-tied comparisons with fixed languagemodel tokens. Further ablations show gains from full-song context and renderer-side text conditioning. The complete system outperforms five commercial systems on 15 of 18 automatic metrics and ranks among the top three on the Artificial Analysis Music with Vocals Leaderboard. The demo page is available at https://selinacloudl.github.io/fulldit-demo/.

eess.AS↗
arXiv2026

Plasmon modes in quadratic and cubic nodal line semimetals

Nodal line semimetals (NLSMs) host distinctive topological properties and unconventional collective excitations. While plasmons in linear NLSMs are well studied, those of higher-order dispersive NLSMs remain poorly understood. We investigate plasmon modes in 3D quadratic and cubic NLSMs within the random phase approximation (RPA), valid for $r_s\ll1$ but breaking down in the small-doping cubic regime where the density of states diverges. Deriving the one-loop polarizability and evaluating its coefficients numerically from the full 3D Lindhard integral, we find distinct carrier-density scaling laws: quadratic NLSMs exhibit $ω_p\sim n^{1/2}$, while cubic NLSMs show $ω_p\sim n^{2/3}$ at large doping crossing over to $ω_p\sim n^{3/4}$ at small doping. The $n^{3/4}$ law is not quantitatively reliable: the diverging density of states drives $r_s>1$ already at $n\sim10^{18}$ cm$^{-3}$, limiting the observable window to at most $10^{18}\lesssim n\lesssim10^{19}$ cm$^{-3}$; below this, beyond-RPA correlations are essential. These scalings originate from the power-law density of states and intraband (Drude) response, and are analogous in phase-space power counting to bilayer and trilayer graphene, respectively. The long-wavelength intraband polarization dominates the plasmon frequency, while interband contributions are subleading (two to three orders of magnitude smaller for $Ω\llμ$). Both systems share a plasmon anisotropy $Ω_p^z/Ω_p^\perp\to\sqrt{2}$ in the thin-ring limit, a generic consequence of the torus Fermi-surface geometry. This $\sqrt{2}$ doublet, though observable by HREELS, cannot distinguish quadratic from cubic dispersion; the density scaling exponent is the true distinguishing signature. Our results provide a framework for the collective dynamics of higher-order dispersive NLSMs and suggest experimentally testable signatures accessible by HREELS.

cond-mat.str-el↗
arXiv2026

ALTER: Modeling Longitudinal Changes via Regional Differencing for 3D CT Report Generation

Computed tomography (CT) is widely used for clinical diagnosis and longitudinal follow-up, yet automatically generating accurate and complete radiology reports from three-dimensional (3D) CT remains challenging. Existing methods improve fine-grained correspondence between images and text by modeling anatomical regions, but remain centered on the current examination. Consequently, patient-specific longitudinal changes within individual regions remain insufficiently modeled. Meanwhile, interval changes are often distributed across multiple anatomical regions, complicating a coherent assessment of the overall longitudinal state. We propose Anatomically Localized Temporal Evidence Representation (ALTER) to address these limitations. Global Prior Integration (GPI) incorporates the prior CT and report to establish historical context for the current examination. Regional Proxy Differencing (RPD) enables each current anatomical region to retrieve a historical proxy from a single shared encoding of the prior volume and to derive localized interval evidence. Interval Change Fusion (ICF) further combines current abnormality states with region-distributed differences, converting their joint representation into change-aware soft prompts that guide report generation. ALTER achieves state-of-the-art results on most evaluation metrics across the RadGenome-ChestCT validation and CTRG-Chest-548K test sets. Code and data preprocessing details are available at https://github.com/peytonkarlie/ALTER/tree/main.

cs.CV↗
arXiv2026

GST-Bench: Can VLMs Develop Global Spatial Awareness from Video?

Spatial intelligence is fundamental to embodied agents, yet existing benchmarks focus on local spatial perception from single or few viewpoints, overlooking global spatial awareness over continuous, long-horizon visual streams. To address this limitation, we introduce the Global-Spatial-Temporal Benchmark (GST-Bench), a VQA benchmark for global spatial intelligence in video understanding, comprising human-verified questions derived from 6,790 minutes of synthetically generated video. It requires models to perform accurate spatial inference from novel viewpoints unseen in the input video and to map egocentric observations onto global top-down images. A comprehensive evaluation of 22 state-of-the-art VLMs exposes a striking gap between models and humans: the strongest zero-shot model attains only 42.68, far below the human score of 79.08. To probe the cause of this gap, we construct GST-Bench-Local and find that models, despite strong local spatial understanding under the same task formulation, still fail to consolidate long-horizon observations into a globally consistent scene representation. We further provide GST-Train, a dataset for global spatial reasoning, as a complementary resource to facilitate future research on this challenge.

cs.CV↗
arXiv2026

IcFuzz: Fuzzing Isaac Sim with Semantic Stage Guidance and Multi-level Mutation

Robotics simulators serve as a foundational infrastructure for embodied AI, facilitating safe and scalable robotic system development. NVIDIA Isaac Sim has emerged as one of the most popular simulators, distinguished by its GPU-accelerated physics engine and photorealistic rendering, which enable high-fidelity modeling of complex environments. However, its inherent complexity inevitably introduces software bugs that can compromise simulation reliability. Existing fuzzing approaches struggle to test Isaac Sim effectively due to challenges of context-aware object semantics, hierarchical simulation control, and a vast simulation state space. In this paper, we propose IcFuzz, the first fuzzing approach for Isaac Sim. IcFuzz first performs an LLM-based semantic stage segmentation, decomposing simulation programs into structured stages that capture context-aware object semantics. Guided by this information, IcFuzz designs multi-level mutation operators to systematically exercise the simulator across hierarchical granularities. To efficiently navigate the vast simulation state space, IcFuzz employs a multi-armed bandit algorithm to adaptively schedule mutation operators. Experimental results show that IcFuzz outperforms the baselines in terms of both code coverage and bug detection. Specifically, IcFuzz achieves approximately 190\%--205\% of the code coverage of the baselines and detects an average of 3.7 unique crashes over three rounds of 12-hour tests, while no crashes are detected by the baselines. Moreover, IcFuzz has uncovered 11 bugs over approximately four months, 9 of which have been confirmed or fixed by the developers.

cs.RO↗
arXiv2026

Improved covariant analysis of $B_c^+ \to χ_{c1}(nP)$ decays and implications for the nature of $χ_{c1}(3872)$

We study the weak decays $B_c^+\toχ_{c1}(nP)\ell^+ ν_{\ell}$ and $B_c^+\toχ_{c1}(nP)X$ (n=1,2,3) within the Bethe-Salpeter formalism, treating $χ_{c1}(3872)$ as the conventional $χ_{c1}(2P)$ charmonium. We upgrade the covariant hadronic transition amplitude to consistently evaluate the final-state wave function in its rest frame, enabling a reliable description of large-recoil processes and sizable relativistic corrections pertinent to highly excited charmonium. Our results provide a novel platform to probe the internal structure of $χ_{c1}(3872)$ via $B_c$ decays. While the LHCb search for $B_c^+\toχ_{c1}(3872)π^+$ yielded only an upper limit, we demonstrate that this is primarily due to insufficient luminosity, approximately 20 times the current $B_c$ data sample would be required for a definitive observation. In contrast, we identify the semileptonic mode $B_c^+\toχ_{c1}(3872)μ^+ν_μ$ as a far more promising channel, which could become accessible with merely twice the existing data set. Our predictions offer concrete guidance for upcoming LHCb analyses and complement ongoing efforts to resolve the nature of $χ_{c1}(3872)$.

hep-ph↗