arXiv subjects

Wendi Chen

Publications and source records attributed to Wendi Chen.

17 recordsLinked to original sources

arXiv2026

FA-RDP: A Frequency-Adaptive Reactive Diffusion Policy for Contact-Rich Manipulation

In contact-rich manipulation, action multimodality and reactivity dominate different stages of a single episode. Before contact, multiple trajectories might be equally valid, making it important to preserve diverse action modes. After contact, geometric constraints and force limits narrow the solution space, while successful execution demands rapid responses to force feedback. However, standard diffusion policies use a fixed inference frequency and sampling steps throughout the episode, forcing a fundamental compromise: low-frequency, multi-step sampling better preserves pre-contact multimodality but responds slowly to force feedback, whereas high-frequency sampling improves reactivity but tends to collapse distinct pre-contact modes. To resolve this tradeoff, we present FA-RDP, a frequency-adaptive reactive diffusion policy. A shared multi-frequency visual-force Transformer predicts action chunks at both low and high frequencies, while a learned multimodality indicator dynamically selects multi-step low-frequency sampling before contact and one-step high-frequency sampling as action ambiguity decreases. We further introduce Manifold Consistency Distillation (MCD), which reparameterizes the diffusion network to predict actions on the robot action manifold while retaining DDPM-based residual supervision. Experiments on three contact-rich manipulation tasks show that FA-RDP achieves the highest success rate while preserving diverse pre-contact trajectory modes. Code and videos are available at https://fa-rdp.github.io.

cs.RO↗

arXiv2026

Never Too Late for Force: Accelerating VLA Post-Training with Reactive Force Injection

Pretrained vision-language-action (VLA) policies provide strong language-conditioned manipulation knowledge, but they remain largely vision-driven and can struggle once manipulation enters contact states where the scene is occluded, depth is ambiguous, or small force errors push execution off the offline demonstration distribution. We present LIFT (Late Reactive Injection of Force for VLA Post-Training), a force-aware post-training framework that adds contact reactivity to a pretrained VLA policy while preserving its general manipulation knowledge. LIFT grafts a reactive action expert beside the original action expert, initializes it from pretrained action weights, and injects recent 6D end-effector force through causal force memory and zero-initialized cross attention, enabling actions to be refreshed during execution. To address the policy-dependent distribution shift of contact feedback, LIFT further couples reactive force injection with an online DAgger loop that trains on a mixture of offline task-alignment data and human-corrected online rollouts. Across towel folding, book insertion, and Hanoi ring placement, LIFT learns faster and reaches higher performance than vision-only post-training, while ablations show that reactive force memory and online corrective data are both important for robust contact-rich manipulation. Our code and data will be publicly available.

cs.RO↗

arXiv2026

FTP-1: A Generalist Foundation Tactile Policy Across Tactile Sensors for Contact-Rich Manipulation

Despite the success of vision-based generalist robotic policies, existing tactile-based policies remain tied to fixed embodiments and sensor setups. This is because tactile signals are highly heterogeneous across hardware, making cross-sensor generalization difficult. We present FTP-1,the first generalist foundation tactile policy pretrained to acquire transferable tactile manipulation abilities across diverse sensors and embodiments. FTP-1 supports varied tactile inputs, including image-, array-, and state-based signals, by using heterogeneous encoders to project them into unified morphology-aware latent tokens that are jointly modeled by a shared tactile Transformer expert. Pretrained on around 3,000 hours of tactile manipulation data aggregated from 26 data sources, spanning human and robot demonstrations across 21 sensors, FTP-1 learns tactile skills that transfer beyond the sensors seen during pretraining. Across downstream finetuning experiments spanning 5 hardware configurations, FTP-1 improves contact-rich manipulation on seen sensor setups by +17.2% and, surprisingly, transfers to two previously unseen tactile-sensor setups, achieving a +31% gain in success rate. FTP-1 establishes the first unified foundation baseline for tactile manipulation, providing future tactile policies with a shared model-level starting point. Pretrained models, datasets, training code and more visualization at https://ftp1-policy.github.io.

cs.RO↗

arXiv2026

Generation of fission yield covariance matrices and its application in uncertainty analysis of decay heat

The uncertainties and covariance matrices of fission yield are important in the uncertainty analysis of decay heat. At present, there are no covariance matrixes of fission yield given in the evaluated nuclear data library, although they have provided the uncertainties with good estimates. In this work, the generalized least squares (GLS) updating approach was adopted to evaluate the fission yield covariances with the constraints from basic physical conservation equation and chain yield data, using the nuclear data files from ENDF/B-VIII.0, JENDL-5 and JEFF-3.3. Based on these original and updated data, summation calculation was performed for fission pulse decay heat of thermal neutron-induced fission of $^{235}$U. The uncertainties of decay heat were obtained through generalized perturbation theory, including the uncertainties propagated from fission yield, decay energy, decay constant and branching ratio. The original uncorrelated yield data contributes a $\sim 4 \%$ uncertainty at all times and dominates the decay heat uncertainty at cooling times longer than \SI{100}{s}. With the generated covariance matrixes, the uncertainty of calculated decay heat is strongly reduced and decay energy data makes a major contribution in general. The relative uncertainties at cooling time \SI{0.1}{\second} are $\sim$10$\%$ for ENDF/V-VIII.0 and JEFF-3.3 and $\sim$5$\%$ for JENDL-5 and those at cooling time 10$^{5}$ s are about 1$\%$ for three libraries. The influence of the GLS updating procedure on the contributions of important fission products to decay heat and their sensitive coefficients was also discussed.

nucl-th↗

arXiv2026

VTAM: Video-Tactile-Action Models for Complex Physical Interaction Beyond VLAs

Video-Action Models (VAMs) have emerged as a promising framework for embodied intelligence, learning implicit world dynamics from raw video streams to produce temporally consistent action predictions. Although such models demonstrate strong performance on long-horizon tasks through visual reasoning, they remain limited in contact-rich scenarios where critical interaction states are only partially observable from vision alone. In particular, fine-grained force modulation and contact transitions are not reliably encoded in visual tokens, leading to unstable or imprecise behaviors. To bridge this gap, we introduce the Video-Tactile Action Model (VTAM), a multimodal world modeling framework that incorporates tactile perception as a complementary grounding signal. VTAM augments a pretrained video transformer with tactile streams via a lightweight modality transfer finetuning, enabling efficient cross-modal representation learning without tactile-language paired data or independent tactile pretraining. To stabilize multimodal fusion, we introduce a tactile regularization loss that enforces balanced cross-modal attention, preventing visual latent dominance in the action model. VTAM demonstrates superior performance in contact-rich manipulation, maintaining a robust success rate of 90 percent on average. In challenging scenarios such as potato chip pick-and-place requiring high-fidelity force awareness, VTAM outperforms the pi 0.5 baseline by 80 percent. Our findings demonstrate that integrating tactile feedback is essential for correcting visual estimation errors in world action models, providing a scalable approach to physically grounded embodied foundation models.

cs.RO↗

arXiv2026

RoboPocket: Improve Robot Policies Instantly with Your Phone

Scaling imitation learning is fundamentally constrained by the efficiency of data collection. While handheld interfaces have emerged as a scalable solution for in-the-wild data acquisition, they predominantly operate in an open-loop manner: operators blindly collect demonstrations without knowing the underlying policy's weaknesses, leading to inefficient coverage of critical state distributions. Conversely, interactive methods like DAgger effectively address covariate shift but rely on physical robot execution, which is costly and difficult to scale. To reconcile this trade-off, we introduce RoboPocket, a portable system that enables Robot-Free Instant Policy Iteration using single consumer smartphones. Its core innovation is a Remote Inference framework that visualizes the policy's predicted trajectory via Augmented Reality (AR) Visual Foresight. This immersive feedback allows collectors to proactively identify potential failures and focus data collection on the policy's weak regions without requiring a physical robot. Furthermore, we implement an asynchronous Online Finetuning pipeline that continuously updates the policy with incoming data, effectively closing the learning loop in minutes. Extensive experiments demonstrate that RoboPocket adheres to data scaling laws and doubles the data efficiency compared to offline scaling strategies, overcoming their long-standing efficiency bottleneck. Moreover, our instant iteration loop also boosts sample efficiency by up to 2$\times$ in distributed environments a small number of interactive corrections per person. Project page and videos: https://robo-pocket.github.io.

cs.RO↗

arXiv2026

Rethinking Camera Choice: An Empirical Study on Fisheye Camera Properties in Robotic Manipulation

The adoption of fisheye cameras in robotic manipulation, driven by their exceptionally wide Field of View (FoV), is rapidly outpacing a systematic understanding of their downstream effects on policy learning. This paper presents the first comprehensive empirical study to bridge this gap, rigorously analyzing the properties of wrist-mounted fisheye cameras for imitation learning. Through extensive experiments in both simulation and the real world, we investigate three critical research questions: spatial localization, scene generalization, and hardware generalization. Our investigation reveals that: (1) The wide FoV significantly enhances spatial localization, but this benefit is critically contingent on the visual complexity of the environment. (2) Fisheye-trained policies, while prone to overfitting in simple scenes, unlock superior scene generalization when trained with sufficient environmental diversity. (3) While naive cross-camera transfer leads to failures, we identify the root cause as scale overfitting and demonstrate that hardware generalization performance can be improved with a simple Random Scale Augmentation (RSA) strategy. Collectively, our findings provide concrete, actionable guidance for the large-scale collection and effective use of fisheye datasets in robotic learning. More results and videos are available on https://robo-fisheye.github.io/

cs.RO↗

arXiv2026

Perturbative calculations of nucleon-deuteron elastic scattering in chiral effective field theory

We develop a framework for calculating nucleon-deuteron scattering using strict perturbation theory for treating subleading interactions in chiral effective field theory (ChEFT). Rather than using direct evaluations in the distorted-wave expansion, our approach solves a hierarchy of integral equations to obtain subleading scattering amplitudes. A benchmark with the wave packet continuum-discretization is performed. This framework benefits from the fact that the renormalization-group invariance chiral forces involves only a limited number of two-body partial waves at leading order. We use it to calculate nucleon-deuteron elastic scattering differential cross sections and analyzing powers up to next-to-leading order.

nucl-th↗

arXiv2025

ImplicitRDP: An End-to-End Visual-Force Diffusion Policy with Structural Slow-Fast Learning

Human-level contact-rich manipulation relies on the distinct roles of two key modalities: vision provides spatially rich but temporally slow global context, while force sensing captures rapid local contact dynamics. Integrating these signals is challenging due to their fundamental frequency and informational disparities. In this work, we propose ImplicitRDP, a unified end-to-end visual-force diffusion policy that integrates visual planning and reactive force control within a single network. We introduce Structural Slow-Fast Learning, a mechanism utilizing causal attention to simultaneously process asynchronous visual and force tokens, allowing the policy to perform rapid force control at the action rate while maintaining the temporal coherence of action chunks. Furthermore, to mitigate modality collapse where end-to-end models fail to adjust the weights across different modalities, we propose Virtual-target-based Representation Regularization. This auxiliary objective maps force feedback into the same space as the action, providing a stronger, physics-grounded learning signal than raw force prediction. Extensive experiments on contact-rich tasks demonstrate that ImplicitRDP significantly outperforms both vision-only and hierarchical baselines, achieving superior reactivity and success rates with a streamlined training pipeline. Code and videos are available at https://implicit-rdp.github.io.

cs.RO↗

arXiv2025

SOE: Sample-Efficient Robot Policy Self-Improvement via On-Manifold Exploration

Intelligent agents progress by continually refining their capabilities through actively exploring environments. Yet robot policies often lack sufficient exploration capability due to action mode collapse. Existing methods that encourage exploration typically rely on random perturbations, which are unsafe and induce unstable, erratic behaviors, thereby limiting their effectiveness. We propose Self-Improvement via On-Manifold Exploration (SOE), a framework that enhances policy exploration and improvement in robotic manipulation. SOE learns a compact latent representation of task-relevant factors and constrains exploration to the manifold of valid actions, ensuring safety, diversity, and effectiveness. It can be seamlessly integrated with arbitrary policy models as a plug-in module, augmenting exploration without degrading the base policy performance. Moreover, the structured latent space enables human-guided exploration, further improving efficiency and controllability. Extensive experiments in both simulation and real-world tasks demonstrate that SOE consistently outperforms prior methods, achieving higher task success rates, smoother and safer exploration, and superior sample efficiency. These results establish on-manifold exploration as a principled approach to sample-efficient policy self-improvement. Project website: https://ericjin2002.github.io/SOE

cs.RO↗

arXiv2025

Right-Side-Out: Learning Zero-Shot Sim-to-Real Garment Reversal

Turning garments right-side out is a challenging manipulation task: it is highly dynamic, entails rapid contact changes, and is subject to severe visual occlusion. We introduce Right-Side-Out, a zero-shot sim-to-real framework that effectively solves this challenge by exploiting task structures. We decompose the task into Drag/Fling to create and stabilize an access opening, followed by Insert&Pull to invert the garment. Each step uses a depth-inferred, keypoint-parameterized bimanual primitive that sharply reduces the action space while preserving robustness. Efficient data generation is enabled by our custom-built, high-fidelity, GPU-parallel Material Point Method (MPM) simulator that models thin-shell deformation and provides robust and efficient contact handling for batched rollouts. Built on the simulator, our fully automated pipeline scales data generation by randomizing garment geometry, material parameters, and viewpoints, producing depth, masks, and per-primitive keypoint labels without any human annotations. With a single depth camera, policies trained entirely in simulation deploy zero-shot on real hardware, achieving up to 81.3% success rate. By employing task decomposition and high fidelity simulation, our framework enables tackling highly dynamic, severely occluded tasks without laborious human demonstrations.

cs.RO↗

arXiv2025

Reactive Diffusion Policy: Slow-Fast Visual-Tactile Policy Learning for Contact-Rich Manipulation

Humans can accomplish complex contact-rich tasks using vision and touch, with highly reactive capabilities such as fast response to external changes and adaptive control of contact forces; however, this remains challenging for robots. Existing visual imitation learning (IL) approaches rely on action chunking to model complex behaviors, which lacks the ability to respond instantly to real-time tactile feedback during the chunk execution. Furthermore, most teleoperation systems struggle to provide fine-grained tactile / force feedback, which limits the range of tasks that can be performed. To address these challenges, we introduce TactAR, a low-cost teleoperation system that provides real-time tactile feedback through Augmented Reality (AR), along with Reactive Diffusion Policy (RDP), a novel slow-fast visual-tactile imitation learning algorithm for learning contact-rich manipulation skills. RDP employs a two-level hierarchy: (1) a slow latent diffusion policy for predicting high-level action chunks in latent space at low frequency, (2) a fast asymmetric tokenizer for closed-loop tactile feedback control at high frequency. This design enables both complex trajectory modeling and quick reactive behavior within a unified framework. Through extensive evaluation across three challenging contact-rich tasks, RDP significantly improves performance compared to state-of-the-art visual IL baselines. Furthermore, experiments show that RDP is applicable across different tactile / force sensors. Code and videos are available on https://reactive-diffusion-policy.github.io.

cs.RO↗

arXiv2024

DeformPAM: Data-Efficient Learning for Long-horizon Deformable Object Manipulation via Preference-based Action Alignment

In recent years, imitation learning has made progress in the field of robotic manipulation. However, it still faces challenges when addressing complex long-horizon tasks with deformable objects, such as high-dimensional state spaces, complex dynamics, and multimodal action distributions. Traditional imitation learning methods often require a large amount of data and encounter distributional shifts and accumulative errors in these tasks. To address these issues, we propose a data-efficient general learning framework (DeformPAM) based on preference learning and reward-guided action selection. DeformPAM decomposes long-horizon tasks into multiple action primitives, utilizes 3D point cloud inputs and diffusion models to model action distributions, and trains an implicit reward model using human preference data. During the inference phase, the reward model scores multiple candidate actions, selecting the optimal action for execution, thereby reducing the occurrence of anomalous actions and improving task completion quality. Experiments conducted on three challenging real-world long-horizon deformable object manipulation tasks demonstrate the effectiveness of this method. Results show that DeformPAM improves both task completion quality and efficiency compared to baseline methods even with limited data. Code and data will be available at https://deform-pam.robotflow.ai.

cs.RO↗

arXiv2024

Digital Fingerprinting on Multimedia: A Survey

The explosive growth of multimedia content in the digital economy era has brought challenges in content recognition, copyright protection, and data management. As an emerging content management technology, perceptual hash-based digital fingerprints, serving as compact summaries of multimedia content, have been widely adopted for efficient multimedia content identification and retrieval across different modalities (e.g., text, image, video, audio), attracting significant attention from both academia and industry. Despite the increasing applications of digital fingerprints, there is a lack of systematic and comprehensive literature review on multimedia digital fingerprints. This survey aims to fill this gap and provide an important resource for researchers studying the details and related advancements of multimedia digital fingerprints. The survey first introduces the definition, characteristics, and related concepts (including hash functions, granularity, similarity measures, etc.) of digital fingerprints. It then focuses on analyzing and summarizing the algorithms for extracting unimodal fingerprints of different types of digital content, including text fingerprints, image fingerprints, video fingerprints, and audio fingerprints. Particularly, it provides an in-depth review and summary of deep learning-based fingerprints. Additionally, the survey elaborates on the various practical applications of digital fingerprints and outlines the challenges and potential future research directions. The goal is to promote the continued development of multimedia digital fingerprint research.

cs.MM↗

arXiv2023

Elastic scattering and total reaction cross sections of $^{6}$Li studied with a microscopic continuum discretized coupled channels model

We present a systematic study of $^{6}$Li elastic scattering and total reaction cross sections at incident energies around the Coulomb barrier within the continuum discretized coupled-channels (CDCC) framework, where $^{6}$Li is treated in an $α$+$d$ two-body model. Collisions with $^{27}$Al, $^{64}$Zn, $^{138}$Ba and $^{208}$Pa are analyzed. The microscopic optical potentials (MOP) based on Skyrme nucleon-nucleon interaction for $α$ and $d$ are adopted in CDCC calculations and satisfactory agreement with the experimental data is obtained without any adjustment on MOPs. For comparison, the $α$ and $d$ global phenomenological optical potentials (GOP) are also used in CDCC analysis and a reduction no less than 50$\%$ on the surface imaginary part of deuteron GOP is required for describing the data. In all cases, the $^6$Li breakup effect is significant and provides repulsive correction to the folding model potential. The reduction on the surface imaginary part of GOP of deuteron reveals a strong suppression of the reaction probability of deuteron as a component of $^{6}$Li as compared with that of a free deuteron. A further investigation is made by taking the $d$ breakup process into account equivalently within the dynamic polarization potential approach and it shows that $d$ behaves like a tightly bound nucleus in $^{6}$Li induced reactions.

nucl-th↗

arXiv2023

Continuum-discretized coupled-channel calculations for $^{6}$Li fusion reactions with closed channels

Fusion reactions induced by the weakly bound nucleus $^{6}$Li with targets $^{28}$Si, $^{64}$Ni, $^{144}$Sm and $^{209}$Bi at energies around the Coulomb barrier are investigated within a three-body model where $^{6}$Li is described with an $α+ d$ cluster model. The total fusion (TF) cross sections are calculated with the continuum-discretized coupled-channel (CDCC) method and the complete fusion (CF) cross sections are extracted through the sum-rule model. The calculations demonstrate that (i) for the TF cross section calculations, the continuum states up to 40 MeV are found to be necessary, which corresponds to the inclusion of closed channels for light and medium mass targets, such as $^{28}$Si, $^{59}$Co and $^{144}$Sm, (ii) the converged CDCC results for TF cross section at energies above the Coulomb barrier are almost the same as single channel results in which the continuum coupling effect is neglected, and (iii) the continuum coupling strongly influences partial wave fusion cross sections and the closed channels play a significant role in the improvement of the description of the CF cross sections at energies below the Coulomb barrier for the $^6$Li+$^{28}$Si, $^{59}$Co and $^{144}$Sm systems.

nucl-th↗

arXiv2022

Application of the Lagrange mesh method in continuum-discretized coupled-channels calculations

We apply the Lagrange-mesh method to discretize continuum states of weakly bound nuclei for continuum-discretized coupled-channel (CDCC) calculations of three-body breakup reactions. This discretization method is compared with the bin method, which is regarded as the standard continuum discretization method, for the $d$ and $^6$Li induced reactions. An improved Numerov algorithm is used to solve the coupled channels equations, which permits a fast integration of equations and a convenient treatment of the closed channels. A new CDCC model code named \emph{CDCC-R} is developed. In all cases, the combination of the Lagrange-mesh method and the Numerov algorithm shows high efficiency and accuracy for the CDCC calculations of the elastic scattering and breakup reactions. Especially, various numerical and physical aspects are discussed for $^6$Li induced reactions. The coupling effects of the continuum states with high orbital angular momentum and closed channels are discussed. Moderate effects are found in the calculations for elastic scattering and breakup reaction when $F$- and $G$-wave continuum states are included in CDCC model space for $^6$Li induced reactions at the incident energies well above the Coulomb barrier. The closed channel effect is found to be indispensable for $^6$Li breakup reaction calculation when the incident energy is around the Coulomb barrier.

nucl-th↗