arXiv subjects

Yangming Shi

Publications and source records attributed to Yangming Shi.

13 recordsLinked to original sources

arXiv2026

VLAQuantBench: Closed-Loop Evaluation of Post-Training Quantization for Vision-Language-Action Models

Post-training quantization reduces the memory requirements of vision-language-action (VLA) models, but precision selection must account for the interaction between layer scope, numerical format, and calibration. We introduce \textbf{VLAQuantBench}, a controlled evaluation with 409 runs and 94,574 simulation episodes: four models on LIBERO, with X-VLA additionally evaluated on three simulation benchmark families. Under uncalibrated W4A4 round-to-nearest quantization, expanding a $π_{0.5}$ action-head subset from 126 to 167 layers raises success from 7.0\% to 70.5\%. Fixed-observation replay confirms a corresponding numerical recovery. Two-episode calibration removes the severe joint failures in the tested subsets, whereas the same smoothing-and-clipping recipe lowers $π_0$ success and does not recover OpenVLA-OFT end-to-end. For OpenVLA-OFT, protecting one 28,672-parameter output projection instead restores near-baseline success: the remaining 441 eligible linear layers retain W3 on LIBERO-Long or eight-bit activations across all four suites. Task-clustered intervals support the large failure and recovery contrasts. These results establish recipe-dependent interactions and identify concrete precision assignments, rather than universal layer-sensitivity rules. Real-kernel and physical-robot measurements complement the accuracy analysis. Code, configurations, and episode records are publicly available at https://github.com/jiuyixu25/VLAQuantBench.

cs.RO↗

arXiv2026

Predict Before You Deploy: Offline Prediction of Quantization-Induced Task Degradation for World Action Models

World action models (WAMs) rely on video-generation backbones, requiring substantial memory and compute for deployment. Post-training quantization reduces memory and can accelerate inference, but bit width, grouping, and quantizer choice define a large configuration space. Identifying configurations that preserve task performance through exhaustive closed-loop evaluation is costly. We propose PreDE (Predict Before You Deploy), a policy-calibrated framework for predicting quantization-induced task degradation from offline action deviations. Using closed-loop outcomes from a small development set, PreDE calibrates two thresholds and accepts, rejects, or defers new configurations using a fixed observation log. Under a within-setting label-ordering hypothesis, the rule issues decisions where all thresholds consistent with the development labels agree. Across five WAMs and four benchmark settings, quantization produces configuration-dependent task losses that cannot be explained by bit width alone or a shared deviation threshold. Across 28 held-out configurations from two policies, PreDE issued 21 decisions before observing closed-loop outcomes (75% coverage), all matching the observed acceptable or degraded labels. Deferred candidates included both acceptable outcomes and a 33-percentage-point loss. In 450 Franka Research 3 trials across two independently fine-tuned policies, all configurations assigned to high-deviation groups before testing showed significant degradation, while low-deviation comparisons showed no statistically significant degradation. On the real robot, W4A4 achieved a 1.37x action-query speedup and approximately 44% lower peak memory. These results support policy-specific behavioral calibration for quantization configuration selection while identifying candidates that require closed-loop evaluation. The code is available at https://github.com/jiuyixu25/PreDE.

cs.RO↗

arXiv2026

Mamoda2.5: Enhancing Unified Multimodal Model with DiT-MoE

We present Mamoda2.5, a unified AR-Diffusion framework that seamlessly integrates multimodal understanding and generation within a single architecture. To efficiently enhance the model's generation capability, we equip the Diffusion Transformer backbone with a fine-grained Mixture-of-Experts (MoE) design (128 experts, Top-8 routing), yielding a 25B-parameter model that activates only 3B parameters, significantly reducing training costs while scaling up the model capacity. Mamoda2.5 achieves top-tier generation performance on VBench 2.0 and sets a new record in video editing quality, surpassing evaluated open-source models and matching the performance of current top-tier proprietary models, including the Kling O1 on OpenVE-Bench. Furthermore, we introduce a joint few-step distillation and reinforcement learning framework that compresses the 30-step editing model into a 4-step model and greatly accelerates model inference. Compared to open-source baselines, Mamoda2.5 achieves up to $95.9\times$ faster video editing inference. In real-world applications, Mamoda2.5 has been successfully deployed for content moderation and creative restoration tasks in advertising scenarios, achieving a 98% success rate in internal advertising video editing scenario.

cs.CV↗

arXiv2026

UniPrototype: Humn-Robot Skill Learning with Uniform Prototypes

Data scarcity remains a fundamental challenge in robot learning. While human demonstrations benefit from abundant motion capture data and vast internet resources, robotic manipulation suffers from limited training examples. To bridge this gap between human and robot manipulation capabilities, we propose UniPrototype, a novel framework that enables effective knowledge transfer from human to robot domains via shared motion primitives. ur approach makes three key contributions: (1) We introduce a compositional prototype discovery mechanism with soft assignments, enabling multiple primitives to co-activate and thus capture blended and hierarchical skills; (2) We propose an adaptive prototype selection strategy that automatically adjusts the number of prototypes to match task complexity, ensuring scalable and efficient representation; (3) We demonstrate the effectiveness of our method through extensive experiments in both simulation environments and real-world robotic systems. Our results show that UniPrototype successfully transfers human manipulation knowledge to robots, significantly improving learning efficiency and task performance compared to existing approaches.The code and dataset will be released upon acceptance at an anonymous repository.

cs.RO↗

arXiv2026

LowDiff: Efficient Diffusion Sampling with Low-Resolution Condition

Diffusion models have achieved remarkable success in image generation but their practical application is often hindered by the slow sampling speed. Prior efforts of improving efficiency primarily focus on compressing models or reducing the total number of denoising steps, largely neglecting the possibility to leverage multiple input resolutions in the generation process. In this work, we propose LowDiff, a novel and efficient diffusion framework based on a cascaded approach by generating increasingly higher resolution outputs. Besides, LowDiff employs a unified model to progressively refine images from low resolution to the desired resolution. With the proposed architecture design and generation techniques, we achieve comparable or even superior performance with much fewer high-resolution sampling steps. LowDiff is applicable to diffusion models in both pixel space and latent space. Extensive experiments on both conditional and unconditional generation tasks across CIFAR-10, FFHQ and ImageNet demonstrate the effectiveness and generality of our method. Results show over 50% throughput improvement across all datasets and settings while maintaining comparable or better quality. On unconditional CIFAR-10, LowDiff achieves an FID of 2.11 and IS of 9.87, while on conditional CIFAR-10, an FID of 1.94 and IS of 10.03. On FFHQ 64x64, LowDiff achieves an FID of 2.43, and on ImageNet 256x256, LowDiff built on LightningDiT-B/1 produces high-quality samples with a FID of 4.00 and an IS of 195.06, together with substantial efficiency gains.

cs.CV↗

arXiv2026

Many-for-Many: Unify the Training of Multiple Video and Image Generation and Manipulation Tasks

Diffusion models have shown impressive performance in many visual generation and manipulation tasks. Many existing methods focus on training a model for a specific task, especially, text-to-video (T2V) generation, while many other works focus on finetuning the pretrained T2V model for image-to-video (I2V), video-to-video (V2V), image and video manipulation tasks, etc. However, training a strong T2V foundation model requires a large amount of high-quality annotations, which is very costly. In addition, many existing models can perform only one or several tasks. In this work, we introduce a unified framework, namely many-for-many, which leverages the available training data from many different visual generation and manipulation tasks to train a single model for those different tasks. Specifically, we design a lightweight adapter to unify the different conditions in different tasks, then employ a joint image-video learning strategy to progressively train the model from scratch. Our joint learning leads to a unified visual generation and manipulation model with improved video generation performance. In addition, we introduce depth maps as a condition to help our model better perceive the 3D space in visual generation. Two versions of our model are trained with different model sizes (8B and 2B), each of which can perform more than 10 different tasks. In particular, our 8B model demonstrates highly competitive performance in video generation tasks compared to open-source and even commercial engines. Our models and source codes are available at https://github.com/leeruibin/MfM.git.

cs.CV↗

arXiv2025

IDU: Incremental Dynamic Update of Existing 3D Virtual Environments with New Imagery Data

For simulation and training purposes, military organizations have made substantial investments in developing high-resolution 3D virtual environments through extensive imaging and 3D scanning. However, the dynamic nature of battlefield conditions-where objects may appear or vanish over time-makes frequent full-scale updates both time-consuming and costly. In response, we introduce the Incremental Dynamic Update (IDU) pipeline, which efficiently updates existing 3D reconstructions, such as 3D Gaussian Splatting (3DGS), with only a small set of newly acquired images. Our approach starts with camera pose estimation to align new images with the existing 3D model, followed by change detection to pinpoint modifications in the scene. A 3D generative AI model is then used to create high-quality 3D assets of the new elements, which are seamlessly integrated into the existing 3D model. The IDU pipeline incorporates human guidance to ensure high accuracy in object identification and placement, with each update focusing on a single new object at a time. Experimental results confirm that our proposed IDU pipeline significantly reduces update time and labor, offering a cost-effective and targeted solution for maintaining up-to-date 3D models in rapidly evolving military scenarios.

cs.CV↗

arXiv2024

Open-Vocabulary High-Resolution 3D (OVHR3D) Data Segmentation and Annotation Framework

In the domain of the U.S. Army modeling and simulation, the availability of high quality annotated 3D data is pivotal to creating virtual environments for training and simulations. Traditional methodologies for 3D semantic and instance segmentation, such as KpConv, RandLA, Mask3D, etc., are designed to train on extensive labeled datasets to obtain satisfactory performance in practical tasks. This requirement presents a significant challenge, given the inherent scarcity of manually annotated 3D datasets, particularly for the military use cases. Recognizing this gap, our previous research leverages the One World Terrain data repository manually annotated databases, as showcased at IITSEC 2019 and 2021, to enrich the training dataset for deep learning models. However, collecting and annotating large scale 3D data for specific tasks remains costly and inefficient. To this end, the objective of this research is to design and develop a comprehensive and efficient framework for 3D segmentation tasks to assist in 3D data annotation. This framework integrates Grounding DINO and Segment anything Model, augmented by an enhancement in 2D image rendering via 3D mesh. Furthermore, the authors have also developed a user friendly interface that facilitates the 3D annotation process, offering intuitive visualization of rendered images and the 3D point cloud.

cs.CV↗

arXiv2024

To Shelter or Not To Shelter: Exploring the Influence of Different Modalities in Virtual Reality on Individuals' Tornado Mitigation Behaviors

Timely and adequate risk communication before natural hazards can reduce losses from extreme weather events and provide more resilient disaster preparedness. However, existing natural hazard risk communications have been abstract, ineffective, not immersive, and sometimes counterproductive. The implementation of virtual reality (VR) for natural hazard risk communication presents a promising alternative to the existing risk communication system by offering immersive and engaging experiences. However, it is still unknown how different modalities in VR could affect individuals' mitigation behaviors related to incoming natural hazards. In addition, it is also not clear how the repetitive risk communication of different modalities in the VR system leads to the effect of risk habituation. To fill the knowledge gap, we developed a VR system with a tornado risk communication scenario and conducted a mixed-design human subject experiment (N = 24). We comprehensively investigated our research using both quantitative and qualitative results.

cs.HC↗

arXiv2024

Factorized-Dreamer: Training A High-Quality Video Generator with Limited and Low-Quality Data

Text-to-video (T2V) generation has gained significant attention due to its wide applications to video generation, editing, enhancement and translation, \etc. However, high-quality (HQ) video synthesis is extremely challenging because of the diverse and complex motions existed in real world. Most existing works struggle to address this problem by collecting large-scale HQ videos, which are inaccessible to the community. In this work, we show that publicly available limited and low-quality (LQ) data are sufficient to train a HQ video generator without recaptioning or finetuning. We factorize the whole T2V generation process into two steps: generating an image conditioned on a highly descriptive caption, and synthesizing the video conditioned on the generated image and a concise caption of motion details. Specifically, we present \emph{Factorized-Dreamer}, a factorized spatiotemporal framework with several critical designs for T2V generation, including an adapter to combine text and image embeddings, a pixel-aware cross attention module to capture pixel-level image information, a T5 text encoder to better understand motion description, and a PredictNet to supervise optical flows. We further present a noise schedule, which plays a key role in ensuring the quality and stability of video generation. Our model lowers the requirements in detailed captions and HQ videos, and can be directly trained on limited LQ datasets with noisy and brief captions such as WebVid-10M, largely alleviating the cost to collect large-scale HQ video-text pairs. Extensive experiments in a variety of T2V and image-to-video generation tasks demonstrate the effectiveness of our proposed Factorized-Dreamer. Our source codes are available at \url{https://github.com/yangxy/Factorized-Dreamer/}.

cs.CV↗

arXiv2024

Motion Guided Token Compression for Efficient Masked Video Modeling

Recent developments in Transformers have achieved notable strides in enhancing video comprehension. Nonetheless, the O($N^2$) computation complexity associated with attention mechanisms presents substantial computational hurdles when dealing with the high dimensionality of videos. This challenge becomes particularly pronounced when striving to increase the frames per second (FPS) to enhance the motion capturing capabilities. Such a pursuit is likely to introduce redundancy and exacerbate the existing computational limitations. In this paper, we initiate by showcasing the enhanced performance achieved through an escalation in the FPS rate. Additionally, we present a novel approach, Motion Guided Token Compression (MGTC), to empower Transformer models to utilize a smaller yet more representative set of tokens for comprehensive video representation. Consequently, this yields substantial reductions in computational burden and remains seamlessly adaptable to increased FPS rates. Specifically, we draw inspiration from video compression algorithms and scrutinize the variance between patches in consecutive video frames across the temporal dimension. The tokens exhibiting a disparity below a predetermined threshold are then masked. Notably, this masking strategy effectively addresses video redundancy while conserving essential information. Our experiments, conducted on widely examined video recognition datasets, Kinetics-400, UCF101 and HMDB51, demonstrate that elevating the FPS rate results in a significant top-1 accuracy score improvement of over 1.6, 1.6 and 4.0. By implementing MGTC with the masking ratio of 25\%, we further augment accuracy by 0.1 and simultaneously reduce computational costs by over 31\% on Kinetics-400. Even within a fixed computational budget, higher FPS rates paired with MGTC sustain performance gains when compared to lower FPS settings.

cs.CV↗

arXiv2022

APRNet: Attention-based Pixel-wise Rendering Network for Photo-Realistic Text Image Generation

Style-guided text image generation tries to synthesize text image by imitating reference image's appearance while keeping text content unaltered. The text image appearance includes many aspects. In this paper, we focus on transferring style image's background and foreground color patterns to the content image to generate photo-realistic text image. To achieve this goal, we propose 1) a content-style cross attention based pixel sampling approach to roughly mimicking the style text image's background; 2) a pixel-wise style modulation technique to transfer varying color patterns of the style image to the content image spatial-adaptively; 3) a cross attention based multi-scale style fusion approach to solving text foreground misalignment issue between style and content images; 4) an image patch shuffling strategy to create style, content and ground truth image tuples for training. Experimental results on Chinese handwriting text image synthesis with SCUT-HCCDoc and CASIA-OLHWDB datasets demonstrate that the proposed method can improve the quality of synthetic text images and make them more photo-realistic.

cs.CV↗

arXiv2019

Low-light Image Enhancement Algorithm Based on Retinex and Generative Adversarial Network

Low-light image enhancement is generally regarded as a challenging task in image processing, especially for the complex visual tasks at night or weakly illuminated. In order to reduce the blurs or noises on the low-light images, a large number of papers have contributed to applying different technologies. Regretfully, most of them had served little purposes in coping with the extremely poor illumination parts of images or test in practice. In this work, the authors propose a novel approach for processing low-light images based on the Retinex theory and generative adversarial network (GAN), which is composed of the decomposition part for splitting the image into illumination image and reflected image, and the enhancement part for generating high-quality image. Such a discriminative network is expected to make the generated image clearer. Couples of experiments have been implemented under the circumstance of different lighting strength on the basis of Converted See-In-the-Dark (CSID) datasets, and the satisfactory results have been achieved with exceeding expectation that much encourages the authors. In a word, the proposed GAN-based network and employed Retinex theory in this work have proven to be effective in dealing with the low-light image enhancement problems, which will benefit the image processing with no doubt.

cs.CV↗