arXiv subjects

Dehui Wang

Publications and source records attributed to Dehui Wang.

At least 19 recordsLinked to original sources

arXiv2026

Efficient Poisson Subsampling for the Partially Linear Additive Cox Model

To address the computational and storage challenges often encountered in large-scale survival data analysis, we propose an efficient Poisson subsampling method for the partially linear additive Cox model. This model provides a flexible yet interpretable framework by incorporating linear covariate effects, additive nonparametric components for nonlinear covariates, and a nonparametric baseline hazard function. The proposed method adopts B-spline basis functions to approximate the nonparametric components and employs the decorrelated score technique to construct a Poisson subsampling-based estimation equation, based on which we establish the asymptotic normality of the resulting estimator and derive the optimal subsampling probabilities according to the L-optimality criterion. Furthermore, we design a two-step adaptive algorithm for practical implementation. The proposed approach enables computationally efficient statistical inference for large-scale survival analysis without processing the full dataset. We validate the performance of the proposed method through extensive simulation studies and a real-world application to a lymphoma cancer dataset, demonstrating its efficiency and accuracy in large-scale settings.

stat.ME↗

arXiv2026

Rein3D: Reinforced 3D Indoor Scene Generation with Panoramic Video Diffusion Models

The growing demand for Embodied AI and VR applications has highlighted the need for synthesizing high-quality 3D indoor scenes from sparse inputs. However, existing approaches struggle to infer massive amounts of missing geometry in large unseen areas while maintaining global consistency, often producing locally plausible but globally inconsistent reconstructions. We present Rein3D, a framework that reconstructs full 360-degree indoor environments by coupling explicit 3D Gaussian Splatting (3DGS) with temporally coherent priors from video diffusion models. Our approach follows a "restore-and-refine" paradigm: we employ a radial exploration strategy to render imperfect panoramic videos along trajectories starting from the origin, effectively uncovering occluded regions from a coarse 3DGS initialization. These sequences are restored by a panoramic video-to-video diffusion model and further enhanced via video super-resolution to synthesize high-fidelity geometry and textures. Finally, these refined videos serve as pseudo-ground truths to update the global 3D Gaussian field. To support this task, we construct PanoV2V-15K, a dataset of over 15K paired clean and degraded panoramic videos for diffusion-based scene restoration. Experiments demonstrate that Rein3D produces photorealistic and globally consistent 3D scenes and significantly improves long-range camera exploration compared with existing baselines.

cs.CV↗

arXiv2026

Difference-in-Differences using Double Negative Controls and Graph Neural Networks for Unmeasured Network Confounding

Estimating causal effects from observational network data faces dual challenges of network interference and unmeasured confounding. To address this, we propose a general Difference-in-Differences framework that integrates double negative controls (DNC) and graph neural networks (GNNs). Based on the modified parallel trends assumption and DNC, semiparametric identification of direct and indirect causal effects is established. We then propose doubly robust estimators. Specifically, an approach combining GNNs with the generalized method of moments is developed to estimate the functions of high-dimensional covariates and network structure. Furthermore, we derive the estimator's asymptotic normality under the $\psi$-network dependence and approximate neighborhood interference. Simulations show the finite-sample performance of our estimators. Finally, we apply our method to analyze the impact of China's green credit policy on corporate green innovation.

econ.EM↗

arXiv2025

Expertise need not monopolize: Action-Specialized Mixture of Experts for Vision-Language-Action Learning

Vision-Language-Action (VLA) models are experiencing rapid development and demonstrating promising capabilities in robotic manipulation tasks. However, scaling up VLA models presents several critical challenges: (1) Training new VLA models from scratch demands substantial computational resources and extensive datasets. Given the current scarcity of robot data, it becomes particularly valuable to fully leverage well-pretrained VLA model weights during the scaling process. (2) Real-time control requires carefully balancing model capacity with computational efficiency. To address these challenges, We propose AdaMoE, a Mixture-of-Experts (MoE) architecture that inherits pretrained weights from dense VLA models, and scales up the action expert by substituting the feedforward layers into sparsely activated MoE layers. AdaMoE employs a decoupling technique that decouples expert selection from expert weighting through an independent scale adapter working alongside the traditional router. This enables experts to be selected based on task relevance while contributing with independently controlled weights, allowing collaborative expert utilization rather than winner-takes-all dynamics. Our approach demonstrates that expertise need not monopolize. Instead, through collaborative expert utilization, we can achieve superior performance while maintaining computational efficiency. AdaMoE consistently outperforms the baseline model across key benchmarks, delivering performance gains of 1.8% on LIBERO and 9.3% on RoboTwin. Most importantly, a substantial 21.5% improvement in real-world experiments validates its practical effectiveness for robotic manipulation tasks.

cs.RO↗

arXiv2025

Vlaser: Vision-Language-Action Model with Synergistic Embodied Reasoning

While significant research has focused on developing embodied reasoning capabilities using Vision-Language Models (VLMs) or integrating advanced VLMs into Vision-Language-Action (VLA) models for end-to-end robot control, few studies directly address the critical gap between upstream VLM-based reasoning and downstream VLA policy learning. In this work, we take an initial step toward bridging embodied reasoning with VLA policy learning by introducing Vlaser - a Vision-Language-Action Model with synergistic embodied reasoning capability, which is a foundational vision-language model designed to integrate high-level reasoning with low-level control for embodied agents. Built upon the high-quality Vlaser-6M dataset, Vlaser achieves state-of-the-art performance across a range of embodied reasoning benchmarks - including spatial reasoning, embodied grounding, embodied QA, and task planning. Furthermore, we systematically examine how different VLM initializations affect supervised VLA fine-tuning, offering novel insights into mitigating the domain shift between internet-scale pre-training data and embodied-specific policy learning data. Based on these insights, our approach achieves state-of-the-art results on the WidowX benchmark and competitive performance on the Google Robot benchmark.

cs.CV↗

arXiv2025

SimpleVLA-RL: Scaling VLA Training via Reinforcement Learning

Vision-Language-Action (VLA) models have recently emerged as a powerful paradigm for robotic manipulation. Despite substantial progress enabled by large-scale pretraining and supervised fine-tuning (SFT), these models face two fundamental challenges: (i) the scarcity and high cost of large-scale human-operated robotic trajectories required for SFT scaling, and (ii) limited generalization to tasks involving distribution shift. Recent breakthroughs in Large Reasoning Models (LRMs) demonstrate that reinforcement learning (RL) can dramatically enhance step-by-step reasoning capabilities, raising a natural question: Can RL similarly improve the long-horizon step-by-step action planning of VLA? In this work, we introduce SimpleVLA-RL, an efficient RL framework tailored for VLA models. Building upon veRL, we introduce VLA-specific trajectory sampling, scalable parallelization, multi-environment rendering, and optimized loss computation. When applied to OpenVLA-OFT, SimpleVLA-RL achieves SoTA performance on LIBERO and even outperforms $\pi_0$ on RoboTwin 1.0\&2.0 with the exploration-enhancing strategies we introduce. SimpleVLA-RL not only reduces dependence on large-scale data and enables robust generalization, but also remarkably surpasses SFT in real-world tasks. Moreover, we identify a novel phenomenon ``pushcut'' during RL training, wherein the policy discovers previously unseen patterns beyond those seen in the previous training process. Github: https://github.com/PRIME-RL/SimpleVLA-RL

cs.RO↗

arXiv2025

Semiparametric model averaging for high-dimensional quantile regression with nonignorable nonresponse

Model averaging has demonstrated superior performance for ensemble forecasting in high-dimensional framework, its extension to incomplete datasets remains a critical but underexplored challenge. Moreover, identifying the parsimonious model through averaging procedure in quantile regression demands urgent methodological innovation. In this paper, we propose a novel model averaging method for high-dimensional quantile regression with nonignorable missingness. The idea is to relax the parametric constraint on the conditional distribution of respondents, which is constructed through the two-phase scheme: (i) a semiparametric likelihood-based estimation for the missing mechanism, and (ii) a semiparametric weighting procedure to combine candidate models. One of pivotal advantages is our SMA estimator can asymptotically concentrate on the optimally correct model when the candidate set involves at least one correct model. Theoretical results show that the estimator achieves asymptotic optimality even under complex missingness conditions. Empirical conclusions illustrate the efficiency of the method.

stat.ME↗

arXiv2025

Benchmarking Generalizable Bimanual Manipulation: RoboTwin Dual-Arm Collaboration Challenge at CVPR 2025 MEIS Workshop

Embodied Artificial Intelligence (Embodied AI) is an emerging frontier in robotics, driven by the need for autonomous systems that can perceive, reason, and act in complex physical environments. While single-arm systems have shown strong task performance, collaborative dual-arm systems are essential for handling more intricate tasks involving rigid, deformable, and tactile-sensitive objects. To advance this goal, we launched the RoboTwin Dual-Arm Collaboration Challenge at the 2nd MEIS Workshop, CVPR 2025. Built on the RoboTwin Simulation platform (1.0 and 2.0) and the AgileX COBOT-Magic Robot platform, the competition consisted of three stages: Simulation Round 1, Simulation Round 2, and a final Real-World Round. Participants totally tackled 17 dual-arm manipulation tasks, covering rigid, deformable, and tactile-based scenarios. The challenge attracted 64 global teams and over 400 participants, producing top-performing solutions like SEM and AnchorDP3 and generating valuable insights into generalizable bimanual policy learning. This report outlines the competition setup, task design, evaluation methodology, key findings and future direction, aiming to support future research on robust and generalizable bimanual manipulation policies. The Challenge Webpage is available at https://robotwin-benchmark.github.io/cvpr-2025-challenge/.

cs.RO↗

arXiv2025

EmbodiedGen: Towards a Generative 3D World Engine for Embodied Intelligence

Constructing a physically realistic and accurately scaled simulated 3D world is crucial for the training and evaluation of embodied intelligence tasks. The diversity, realism, low cost accessibility and affordability of 3D data assets are critical for achieving generalization and scalability in embodied AI. However, most current embodied intelligence tasks still rely heavily on traditional 3D computer graphics assets manually created and annotated, which suffer from high production costs and limited realism. These limitations significantly hinder the scalability of data driven approaches. We present EmbodiedGen, a foundational platform for interactive 3D world generation. It enables the scalable generation of high-quality, controllable and photorealistic 3D assets with accurate physical properties and real-world scale in the Unified Robotics Description Format (URDF) at low cost. These assets can be directly imported into various physics simulation engines for fine-grained physical control, supporting downstream tasks in training and evaluation. EmbodiedGen is an easy-to-use, full-featured toolkit composed of six key modules: Image-to-3D, Text-to-3D, Texture Generation, Articulated Object Generation, Scene Generation and Layout Generation. EmbodiedGen generates diverse and interactive 3D worlds composed of generative 3D assets, leveraging generative AI to address the challenges of generalization and evaluation to the needs of embodied intelligence related research. Code is available at https://horizonrobotics.github.io/robot_lab/embodied_gen/index.html.

cs.RO↗

arXiv2025

A Novel Framework for Modeling Quarantinable Disease Transmission

The COVID-19 pandemic has significantly challenged traditional epidemiological models due to factors such as delayed diagnosis, asymptomatic transmission, isolation-induced contact changes, and underreported mortality. In response to these complexities, this paper introduces a novel CURNDS model prioritizing compartments and transmissions based on contact levels, rather than merely on symptomatic severity or hospitalization status. The framework surpasses conventional uniform mixing and static rate assumptions by incorporating adaptive power laws, dynamic transmission rates, and spline-based smoothing techniques. The CURNDS model provides accurate estimates of undetected infections and undocumented deaths from COVID-19 data, uncovering the pandemic's true impact. Our analysis challenges the assumption of homogeneous mixing between infected and non-infected individuals in traditional epidemiological models. By capturing the nuanced transmission dynamics of infection and confirmation, our model offers new insights into the spread of different COVID-19 strains. Overall, CURNDS provides a robust framework for understanding the complex transmission patterns of highly contagious, quarantinable diseases.

stat.AP↗

arXiv2025

CondiQuant: Condition Number Based Low-Bit Quantization for Image Super-Resolution

Low-bit model quantization for image super-resolution (SR) is a longstanding task that is renowned for its surprising compression and acceleration ability. However, accuracy degradation is inevitable when compressing the full-precision (FP) model to ultra-low bit widths (2~4 bits). Experimentally, we observe that the degradation of quantization is mainly attributed to the quantization of activation instead of model weights. In numerical analysis, the condition number of weights could measure how much the output value can change for a small change in the input argument, inherently reflecting the quantization error. Therefore, we propose CondiQuant, a condition number based low-bit post-training quantization for image super-resolution. Specifically, we formulate the quantization error as the condition number of weight metrics. By decoupling the representation ability and the quantization sensitivity, we design an efficient proximal gradient descent algorithm to iteratively minimize the condition number and maintain the output still. With comprehensive experiments, we demonstrate that CondiQuant outperforms existing state-of-the-art post-training quantization methods in accuracy without computation overhead and gains the theoretically optimal compression ratio in model parameters. Our code and model are released at https://github.com/Kai-Liu001/CondiQuant.

cs.CV↗

arXiv2024

Programmable Jumping-Droplet Condensation

Self-propelled droplet jumping during condensation has attractive prospects for energy harvesting, water collection and thermal management, but its real-life applications are greatly limited to the challenge of enabling a sustainable control on the entire droplet lifecycle. Herein, we propose a programmable jumping-droplet condensation that evolves along an artificially designed pathway without external stimulations, where the droplets can uniformly form at specific sites, spontaneously migrate and coalesce with their neighboring droplets, and jump off effectively to continuously refresh surface, significantly enhancing the heat transfer performance and durability of condensation. The programmable jumping-droplet condensation is achieved using a wedge-walled rhombus lattice structure surface inspired from the structures and functions of Namib desert beetle skin, shorebird beak and setaria viridis leaf vein. This surface integrates wetting contrast patterns with dual-gradient hierarchical structures, providing persistent and multidimensional droplet rectifications and thus realizing a sustainable control on the entire droplet lifecycle. Furthermore, we systematically investigate the morphology and behavior evolutions of droplets throughout their entire lifecycle, and fully elucidate the programmable control mechanisms of the lattice structure determined by its topology and wettability features. This work not only serves as theoretical foundations and reference framework to realize a durable jumping-droplet condensation and achieve its performance ceiling in a controlled manner, but also promotes the design and fabrication of functional structured surfaces for droplet manipulation and delivery, self-cleaning and anti-fogging/icing.

physics.flu-dyn↗

arXiv2024

Bivariate first-order random coefficient integer-valued autoregressive processes based on modified negative binomial operator

In this paper, a new bivariate random coefficient integer-valued autoregressive process based on modified negative binomial operator with dependent innovations is proposed. Basic probabilistic and statistical properties of this model are derived. To estimate unknown parameters, Yule-Walker, conditional least squares and conditional maximum likelihood methods are considered and evaluated by Monte Carlo simulations. Asymptotic properties of the estimators are derived. Moreover, coherent forecasting and possible extension of the proposed model is provided. Finally, the proposed model is applied to the monthly crime datasets and compared with other models.

math.ST↗

arXiv2024

Inference for multiple change-points in generalized integer-valued autoregressive model

In this paper, we propose a computationally valid and theoretically justified methods, the likelihood ratio scan method (LRSM), for estimating multiple change-points in a piecewise stationary generalized conditional integer-valued autoregressive process. LRSM with the usual window parameter $h$ is more satisfied to be used in long-time series with few and even change-points vs. LRSM with the multiple window parameter $h_{mix}$ performs well in short-time series with large and dense change-points. The computational complexity of LRSM can be efficiently performed with order $O((\log n)^3 n)$. Moreover, two bootstrap procedures, namely parametric and block bootstrap, are developed for constructing confidence intervals (CIs) for each of the change-points. Simulation experiments and real data analysis show that the LRSM and bootstrap procedures have excellent performance and are consistent with the theoretical analysis.

stat.ME↗

arXiv2023

A new First-Order mixture integer-valued threshold autoregressive process based on binomial thinning and negative binomial thinning

In this paper, we introduce a new first-order mixture integer-valued threshold autoregressive process, based on the binomial and negative binomial thinning operators. Basic probabilistic and statistical properties of this model are discussed. Conditional least squares (CLS) and conditional maximum likelihood (CML) estimators are derived and the asymptotic properties of the estimators are established. The inference for the threshold parameter is obtained based on the CLS and CML score functions. Moreover, the Wald test is applied to detect the existence of the piecewise structure. Simulation studies are considered, along with an application: the number of criminal mischief incidents in the Pittsburgh dataset.

stat.AP↗

arXiv2023

Change-points analysis for generalized integer-valued autoregressive model via minimum description length principle

This article considers the problem of modeling a class of nonstationary count time series using multiple change-points generalized integer-valued autoregressive (MCP-GINAR) processes. The minimum description length principle (MDL) is applied to study the statistical inference for the MCP-GINAR model, and the consistency results of the MDL model selection procedure are established respectively under the condition of known and unknown number of change-points. To find the ``best" combination of the number of change-points, the locations of change-points, the order of each segment and its parameters, a genetic algorithm with simulated annealing is implemented to solve this difficult optimization problem. In particular, the simulated annealing process makes up for the precocious problem of the traditional genetic algorithm. Numerical results from simulation experiments and three examples of real data analyses show that the procedure has excellent empirical properties.

stat.AP↗

arXiv2023

Inference for a New Signed Integer Valued Autoregressive Model Based on Pegram's Operator

In the current study, a brand-new SINARS(1) model is proposed for stationary discrete time series defined on $\boldsymbol{Z}$, based on extended binomial distribution and the Pegram's operator. The model effectively characterizes the series of positive and negative integer values generated after differencing some non-stationary time series. The model's attributes are addressed. For the parameter estimation of the model, the conditional maximum likelihood method and Yule-Walker method are taken into consideration. And we prove the asymptotic normality of CML method. By using these two methods, we simulate our model comparing with some relevant ones proposed before. The model can deal with positive or negative autocorrelation data. The analysis of the number of differenced daily new cases in Barbados is done using the suggested model.

stat.AP↗

arXiv2022

A standing Leidenfrost drop with Sufi-whirling

The mobility of Leidenfrost drop has been exploited for the manipulation of drop motions. In the classical model, the Leidenfrost drop was levitated by a vapor cushion, in the absence of touch to the surface. Here we report a standing Leidenfrost state on a heated hydrophobic surface where drop stands on the surface with partial adhesion and further self-rotates like Sufi-whirling. To elucidate this new phenomenon, we imaged the evolution of the partial adhesion, the inner circulation, and the ellipsoidal rotation of the drop. The stable partial adhesion is accompanied by thermal and mechanical equilibrium, and further drives the development of the drop rotation.

physics.flu-dyn↗