Jiangmiao Pang

Jinkun Cao, Jiangmiao Pang, Xinshuo Weng et al. · cmu

Kalman filter (KF) based methods for multi-object tracking (MOT) make an assumption that objects move linearly. While this assumption is acceptable for very short periods of occlusion, linear estimates of motion for prolonged time can be highly inaccurate. Moreover, when there is no measurement available to update Kalman filter parameters, the standard convention is to trust the priori state estimations for posteriori update. This leads to the accumulation of errors during a period of occlusion. The error causes significant motion direction variance in practice. In this work, we show that a basic Kalman filter can still obtain state-of-the-art tracking performance if proper care is taken to fix the noise accumulated during occlusion. Instead of relying only on the linear state estimate (i.e., estimation-centric approach), we use object observations (i.e., the measurements by object detector) to compute a virtual trajectory over the occlusion period to fix the error accumulation of filter parameters during the occlusion period. This allows more time steps to correct errors accumulated during occlusion. We name our method Observation-Centric SORT (OC-SORT). It remains Simple, Online, and Real-Time but improves robustness during occlusion and non-linear motion. Given off-the-shelf detections as input, OC-SORT runs at 700+ FPS on a single CPU. It achieves state-of-the-art on multiple datasets, including MOT17, MOT20, KITTI, head tracking, and especially DanceTrack where the object motion is highly non-linear. The code and models are available at \url{https://github.com/noahcao/OC_SORT}.

Runsen Xu, Tai Wang, Wenwei Zhang et al. · cmu

This paper introduces the Masked Voxel Jigsaw and Reconstruction (MV-JAR) method for LiDAR-based self-supervised pre-training and a carefully designed data-efficient 3D object detection benchmark on the Waymo dataset. Inspired by the scene-voxel-point hierarchy in downstream 3D object detectors, we design masking and reconstruction strategies accounting for voxel distributions in the scene and local point distributions within the voxel. We employ a Reversed-Furthest-Voxel-Sampling strategy to address the uneven distribution of LiDAR points and propose MV-JAR, which combines two techniques for modeling the aforementioned distributions, resulting in superior performance. Our experiments reveal limitations in previous data-efficient experiments, which uniformly sample fine-tuning splits with varying data proportions from each LiDAR sequence, leading to similar data diversity across splits. To address this, we propose a new benchmark that samples scene sequences for diverse fine-tuning splits, ensuring adequate model convergence and providing a more accurate evaluation of pre-training methods. Experiments on our Waymo benchmark and the KITTI dataset demonstrate that MV-JAR consistently and significantly improves 3D detection performance across various data scales, achieving up to a 6.3% increase in mAPH compared to training from scratch. Codes and the benchmark will be available at https://github.com/SmartBot-PJLab/MV-JAR .

Zeqi Xiao, Tai Wang, Jingbo Wang et al. · cmu

Human-Scene Interaction (HSI) is a vital component of fields like embodied AI and virtual reality. Despite advancements in motion quality and physical plausibility, two pivotal factors, versatile interaction control and the development of a user-friendly interface, require further exploration before the practical application of HSI. This paper presents a unified HSI framework, UniHSI, which supports unified control of diverse interactions through language commands. This framework is built upon the definition of interaction as Chain of Contacts (CoC): steps of human joint-object part pairs, which is inspired by the strong correlation between interaction types and human-object contact regions. Based on the definition, UniHSI constitutes a Large Language Model (LLM) Planner to translate language prompts into task plans in the form of CoC, and a Unified Controller that turns CoC into uniform task execution. To facilitate training and evaluation, we collect a new dataset named ScenePlan that encompasses thousands of task plans generated by LLMs based on diverse scenarios. Comprehensive experiments demonstrate the effectiveness of our framework in versatile task execution and generalizability to real scanned scenes. The project page is at https://github.com/OpenRobotLab/UniHSI .

Xiangtai Li, Henghui Ding, Haobo Yuan et al.

Visual segmentation seeks to partition images, video frames, or point clouds into multiple segments or groups. This technique has numerous real-world applications, such as autonomous driving, image editing, robot sensing, and medical analysis. Over the past decade, deep learning-based methods have made remarkable strides in this area. Recently, transformers, a type of neural network based on self-attention originally designed for natural language processing, have considerably surpassed previous convolutional or recurrent approaches in various vision processing tasks. Specifically, vision transformers offer robust, unified, and even simpler solutions for various segmentation tasks. This survey provides a thorough overview of transformer-based visual segmentation, summarizing recent advancements. We first review the background, encompassing problem definitions, datasets, and prior convolutional methods. Next, we summarize a meta-architecture that unifies all recent transformer-based approaches. Based on this meta-architecture, we examine various method designs, including modifications to the meta-architecture and associated applications. We also present several closely related settings, including 3D point cloud segmentation, foundation model tuning, domain-aware segmentation, efficient segmentation, and medical segmentation. Additionally, we compile and re-evaluate the reviewed methods on several well-established datasets. Finally, we identify open challenges in this field and propose directions for future research. The project page can be found at https://github.com/lxtGH/Awesome-Segmentation-With-Transformer. We will also continually monitor developments in this rapidly evolving field.

Tobias Fischer, Thomas E. Huang, Jiangmiao Pang et al. · eth-zurich, mit

Similarity learning has been recognized as a crucial step for object tracking. However, existing multiple object tracking methods only use sparse ground truth matching as the training objective, while ignoring the majority of the informative regions in images. In this paper, we present Quasi-Dense Similarity Learning, which densely samples hundreds of object regions on a pair of images for contrastive learning. We combine this similarity learning with multiple existing object detectors to build Quasi-Dense Tracking (QDTrack), which does not require displacement regression or motion priors. We find that the resulting distinctive feature space admits a simple nearest neighbor search at inference time for object association. In addition, we show that our similarity learning scheme is not limited to video data, but can learn effective instance similarity even from static input, enabling a competitive tracking performance without training on videos or using tracking supervision. We conduct extensive experiments on a wide variety of popular MOT benchmarks. We find that, despite its simplicity, QDTrack rivals the performance of state-of-the-art tracking methods on all benchmarks and sets a new state-of-the-art on the large-scale BDD100K MOT benchmark, while introducing negligible computational overhead to the detector.

Xiangtai Li, Wenwei Zhang, Jiangmiao Pang et al.

This paper presents Video K-Net, a simple, strong, and unified framework for fully end-to-end video panoptic segmentation. The method is built upon K-Net, a method that unifies image segmentation via a group of learnable kernels. We observe that these learnable kernels from K-Net, which encode object appearances and contexts, can naturally associate identical instances across video frames. Motivated by this observation, Video K-Net learns to simultaneously segment and track "things" and "stuff" in a video with simple kernel-based appearance modeling and cross-temporal kernel interaction. Despite the simplicity, it achieves state-of-the-art video panoptic segmentation results on Citscapes-VPS, KITTI-STEP, and VIPSeg without bells and whistles. In particular, on KITTI-STEP, the simple method can boost almost 12\% relative improvements over previous methods. On VIPSeg, Video K-Net boosts almost 15\% relative improvements and results in 39.8 % VPQ. We also validate its generalization on video semantic segmentation, where we boost various baselines by 2\% on the VSPW dataset. Moreover, we extend K-Net into clip-level video framework for video instance segmentation, where we obtain 40.5% mAP for ResNet50 backbone and 54.1% mAP for Swin-base on YouTube-2019 validation set. We hope this simple, yet effective method can serve as a new, flexible baseline in unified video segmentation design. Both code and models are released at https://github.com/lxtGH/Video-K-Net.

Tai Wang, Jiangmiao Pang, Dahua Lin

Perceiving 3D objects from monocular inputs is crucial for robotic systems, given its economy compared to multi-sensor settings. It is notably difficult as a single image can not provide any clues for predicting absolute depth values. Motivated by binocular methods for 3D object detection, we take advantage of the strong geometry structure provided by camera ego-motion for accurate object depth estimation and detection. We first make a theoretical analysis on this general two-view case and notice two challenges: 1) Cumulative errors from multiple estimations that make the direct prediction intractable; 2) Inherent dilemmas caused by static cameras and matching ambiguity. Accordingly, we establish the stereo correspondence with a geometry-aware cost volume as the alternative for depth estimation and further compensate it with monocular understanding to address the second problem. Our framework, named Depth from Motion (DfM), then uses the established geometry to lift 2D image features to the 3D space and detects 3D objects thereon. We also present a pose-free DfM to make it usable when the camera pose is unavailable. Our framework outperforms state-of-the-art methods by a large margin on the KITTI benchmark. Detailed quantitative and qualitative analyses also validate our theoretical conclusions. The code will be released at https://github.com/Tai-Wang/Depth-from-Motion.

Zeqi Xiao, Wenwei Zhang, Tai Wang et al.

DEtection TRansformer (DETR) started a trend that uses a group of learnable queries for unified visual perception. This work begins by applying this appealing paradigm to LiDAR-based point cloud segmentation and obtains a simple yet effective baseline. Although the naive adaptation obtains fair results, the instance segmentation performance is noticeably inferior to previous works. By diving into the details, we observe that instances in the sparse point clouds are relatively small to the whole scene and often have similar geometry but lack distinctive appearance for segmentation, which are rare in the image domain. Considering instances in 3D are more featured by their positional information, we emphasize their roles during the modeling and design a robust Mixed-parameterized Positional Embedding (MPE) to guide the segmentation process. It is embedded into backbone features and later guides the mask prediction and query update processes iteratively, leading to Position-Aware Segmentation (PA-Seg) and Masked Focal Attention (MFA). All these designs impel the queries to attend to specific regions and identify various instances. The method, named Position-guided Point cloud Panoptic segmentation transFormer (P3Former), outperforms previous state-of-the-art methods by 3.4% and 1.2% PQ on SemanticKITTI and nuScenes benchmark, respectively. The source code and models are available at https://github.com/SmartBot-PJLab/P3Former .

Shilong Zhang, Xinjiang Wang, Jiaqi Wang et al.

One-to-one label assignment in object detection has successfully obviated the need for non-maximum suppression (NMS) as postprocessing and makes the pipeline end-to-end. However, it triggers a new dilemma as the widely used sparse queries cannot guarantee a high recall, while dense queries inevitably bring more similar queries and encounter optimization difficulties. As both sparse and dense queries are problematic, then what are the expected queries in end-to-end object detection? This paper shows that the solution should be Dense Distinct Queries (DDQ). Concretely, we first lay dense queries like traditional detectors and then select distinct ones for one-to-one assignments. DDQ blends the advantages of traditional and recent end-to-end detectors and significantly improves the performance of various detectors including FCN, R-CNN, and DETRs. Most impressively, DDQ-DETR achieves 52.1 AP on MS-COCO dataset within 12 epochs using a ResNet-50 backbone, outperforming all existing detectors in the same setting. DDQ also shares the benefit of end-to-end detectors in crowded scenes and achieves 93.8 AP on CrowdHuman. We hope DDQ can inspire researchers to consider the complementarity between traditional methods and end-to-end detectors. The source code can be found at \url{https://github.com/jshilong/DDQ}.

Wenwei Zhang, Jiangmiao Pang, Kai Chen et al.

This paper presents Dense Siamese Network (DenseSiam), a simple unsupervised learning framework for dense prediction tasks. It learns visual representations by maximizing the similarity between two views of one image with two types of consistency, i.e., pixel consistency and region consistency. Concretely, DenseSiam first maximizes the pixel level spatial consistency according to the exact location correspondence in the overlapped area. It also extracts a batch of region embeddings that correspond to some sub-regions in the overlapped area to be contrasted for region consistency. In contrast to previous methods that require negative pixel pairs, momentum encoders or heuristic masks, DenseSiam benefits from the simple Siamese network and optimizes the consistency of different granularities. It also proves that the simple location correspondence and interacted region embeddings are effective enough to learn the similarity. We apply DenseSiam on ImageNet and obtain competitive improvements on various downstream tasks. We also show that only with some extra task-specific losses, the simple framework can directly conduct dense prediction tasks. On an existing unsupervised semantic segmentation benchmark, it surpasses state-of-the-art segmentation methods by 2.1 mIoU with 28% training costs. Code and models are released at https://github.com/ZwwWayne/DenseSiam.

Runsen Xu, Xiaolong Wang, Tai Wang et al.

The unprecedented advancements in Large Language Models (LLMs) have shown a profound impact on natural language processing but are yet to fully embrace the realm of 3D understanding. This paper introduces PointLLM, a preliminary effort to fill this gap, enabling LLMs to understand point clouds and offering a new avenue beyond 2D visual data. PointLLM understands colored object point clouds with human instructions and generates contextually appropriate responses, illustrating its grasp of point clouds and common sense. Specifically, it leverages a point cloud encoder with a powerful LLM to effectively fuse geometric, appearance, and linguistic information. We collect a novel dataset comprising 660K simple and 70K complex point-text instruction pairs to enable a two-stage training strategy: aligning latent spaces and subsequently instruction-tuning the unified model. To rigorously evaluate the perceptual and generalization capabilities of PointLLM, we establish two benchmarks: Generative 3D Object Classification and 3D Object Captioning, assessed through three different methods, including human evaluation, GPT-4/ChatGPT evaluation, and traditional metrics. Experimental results reveal PointLLM's superior performance over existing 2D and 3D baselines, with a notable achievement in human-evaluated object captioning tasks where it surpasses human annotators in over 50% of the samples. Codes, datasets, and benchmarks are available at https://github.com/OpenRobotLab/PointLLM .

Meng Wei, Xiaoyu Yue, Wenwei Zhang et al.

Segmenting and recognizing diverse object parts is a crucial ability in applications spanning various computer vision and robotic tasks. While significant progress has been made in object-level Open-Vocabulary Semantic Segmentation (OVSS), i.e., segmenting objects with arbitrary text, the corresponding part-level research poses additional challenges. Firstly, part segmentation inherently involves intricate boundaries, while limited annotated data compounds the challenge. Secondly, part segmentation introduces an open granularity challenge due to the diverse and often ambiguous definitions of parts in the open world. Furthermore, the large-scale vision and language models, which play a key role in the open vocabulary setting, struggle to recognize parts as effectively as objects. To comprehensively investigate and tackle these challenges, we propose an Open-Vocabulary Part Segmentation (OV-PARTS) benchmark. OV-PARTS includes refined versions of two publicly available datasets: Pascal-Part-116 and ADE20K-Part-234. And it covers three specific tasks: Generalized Zero-Shot Part Segmentation, Cross-Dataset Part Segmentation, and Few-Shot Part Segmentation, providing insights into analogical reasoning, open granularity and few-shot adapting abilities of models. Moreover, we analyze and adapt two prevailing paradigms of existing object-level OVSS methods for OV-PARTS. Extensive experimental analysis is conducted to inspire future research in leveraging foundational models for OV-PARTS. The code and dataset are available at https://github.com/OpenRobotLab/OV_PARTS.

Chenming Zhu, Jingli Lin, Yilin Long et al.

While Vision-Language Models (VLMs) have shown strong visual reasoning capabilities, their spatial reasoning abilities remain largely constrained to the observed images and text-oriented chain-of-thought. They often struggle to infer unobserved layouts, maintain cross-view consistency, and reason from alternative viewpoints when only limited egocentric observations are available. In this work, we study this problem as thinking with imagination, where a VLM actively acquires imagined visual evidence by interacting with a world simulator during reasoning. We propose Astra, an agentic spatial reasoning framework that empowers VLMs with action-conditioned visual imagination. Specifically, Astra couples Astra-VL, an RL-trained VLM policy, with Astra-WM, a Bagel-based world simulator that generates novel-view observations from context images and natural-language camera motions. To provide reliable imagined evidence, Astra-WM is trained with view consistency tuning to improve pose and content consistency across views. In the RL stage, we propose a world-simulator-in-the-loop two-phase RL curriculum to stabilize tool-use exploration and advance the model's ability to invoke the simulator only when imagined observations improve over direct answering. Experiments demonstrate that both the world simulator and the agentic policy are necessary: Astra-WM improves simulator-augmented Gemini-3-Flash on MMSI-Bench from 45.1 to 49.5, while Astra-VL improves the Qwen3-VL backbone from 29.8 to 38.8 on MMSI-Bench and from 36.8 to 42.7 on MindCube. These results show that imagined observations can provide useful spatial evidence, but effective world-model-augmented reasoning requires learning when, where, and how to imagine.

Hanqing Wang, Jiahe Chen, Wensi Huang et al.

Recent works have been exploring the scaling laws in the field of Embodied AI. Given the prohibitive costs of collecting real-world data, we believe the Simulation-to-Real (Sim2Real) paradigm is a crucial step for scaling the learning of embodied models. This paper introduces project GRUtopia, the first simulated interactive 3D society designed for various robots. It features several advancements: (a) The scene dataset, GRScenes, includes 100k interactive, finely annotated scenes, which can be freely combined into city-scale environments. In contrast to previous works mainly focusing on home, GRScenes covers 89 diverse scene categories, bridging the gap of service-oriented environments where general robots would be initially deployed. (b) GRResidents, a Large Language Model (LLM) driven Non-Player Character (NPC) system that is responsible for social interaction, task generation, and task assignment, thus simulating social scenarios for embodied AI applications. (c) The benchmark, GRBench, supports various robots but focuses on legged robots as primary agents and poses moderately challenging tasks involving Object Loco-Navigation, Social Loco-Navigation, and Loco-Manipulation. We hope that this work can alleviate the scarcity of high-quality data in this field and provide a more comprehensive assessment of Embodied AI research. The project is available at https://github.com/OpenRobotLab/GRUtopia.

Shilong Zhang, Xinjiang Wang, Jiaqi Wang et al.

End-to-end object detection is rapidly progressed after the emergence of DETR. DETRs use a set of sparse queries that replace the dense candidate boxes in most traditional detectors. In comparison, the sparse queries cannot guarantee a high recall as dense priors. However, making queries dense is not trivial in current frameworks. It not only suffers from heavy computational cost but also difficult optimization. As both sparse and dense queries are imperfect, then \emph{what are expected queries in end-to-end object detection}? This paper shows that the expected queries should be Dense Distinct Queries (DDQ). Concretely, we introduce dense priors back to the framework to generate dense queries. A duplicate query removal pre-process is applied to these queries so that they are distinguishable from each other. The dense distinct queries are then iteratively processed to obtain final sparse outputs. We show that DDQ is stronger, more robust, and converges faster. It obtains 44.5 AP on the MS COCO detection dataset with only 12 epochs. DDQ is also robust as it outperforms previous methods on both object detection and instance segmentation tasks on various datasets. DDQ blends advantages from traditional dense priors and recent end-to-end detectors. We hope it can serve as a new baseline and inspires researchers to revisit the complementarity between traditional methods and end-to-end detectors. The source code is publicly available at \url{https://github.com/jshilong/DDQ}.

Qing Lian, Tai Wang, Dahua Lin et al.

Recent multi-camera 3D object detectors usually leverage temporal information to construct multi-view stereo that alleviates the ill-posed depth estimation. However, they typically assume all the objects are static and directly aggregate features across frames. This work begins with a theoretical and empirical analysis to reveal that ignoring the motion of moving objects can result in serious localization bias. Therefore, we propose to model Dynamic Objects in RecurrenT (DORT) to tackle this problem. In contrast to previous global Bird-Eye-View (BEV) methods, DORT extracts object-wise local volumes for motion estimation that also alleviates the heavy computational burden. By iteratively refining the estimated object motion and location, the preceding features can be precisely aggregated to the current frame to mitigate the aforementioned adverse effects. The simple framework has two significant appealing properties. It is flexible and practical that can be plugged into most camera-based 3D object detectors. As there are predictions of object motion in the loop, it can easily track objects across frames according to their nearest center distances. Without bells and whistles, DORT outperforms all the previous methods on the nuScenes detection and tracking benchmarks with 62.5\% NDS and 57.6\% AMOTA, respectively. The source code will be released.

Chenming Zhu, Tai Wang, Wenwei Zhang et al.

Recent advancements in Large Multimodal Models (LMMs) have greatly enhanced their proficiency in 2D visual understanding tasks, enabling them to effectively process and understand images and videos. However, the development of LMMs with 3D scene understanding capabilities has been hindered by the lack of large-scale 3D vision-language datasets and powerful 3D encoders. In this paper, we introduce a simple yet effective framework called LLaVA-3D. Leveraging the strong 2D visual understanding priors from LLaVA, our LLaVA-3D efficiently adapts LLaVA for 3D scene understanding without compromising 2D understanding capabilities. To achieve this, we utilize the 3D position embeddings to enhance the 2D CLIP Patches with 3D spatial context information and construct 3D patches. By integrating the 3D position embeddings into 2D LMMs and employing joint 2D and 3D vision-language instruction tuning, we establish a unified architecture for both 2D visual understanding and 3D scene understanding. In contrast to previous 3D LMMs, LLaVA-3D supports decoding accurate 3D spatial perception outputs, e.g., 3D bounding boxes, directly from these 3D patches, without relying on the time-consuming off-the-shelf 3D segmentors. Experimental results show that LLaVA-3D converges 3.5x faster than existing 3D LMMs when trained on 3D vision-language datasets. Moreover, LLaVA-3D not only achieves state-of-the-art performance across various 3D tasks but also maintains comparable 2D visual understanding and vision-language conversation capabilities with LLaVA.

Jingli Lin, Runsen Xu, Shaohao Zhu et al.

Spatial understanding over continuous visual input is crucial for MLLMs to evolve into general-purpose assistants in physical environments. Yet there is still no comprehensive benchmark that holistically assesses the progress toward this goal. In this work, we introduce MMSI-Video-Bench, a fully human-annotated benchmark for video-based spatial intelligence in MLLMs. It operationalizes a four-level framework, Perception, Planning, Prediction, and Cross-Video Reasoning, through 1,106 questions grounded in 1,278 clips from 25 datasets and in-house videos. Each item is carefully designed and reviewed by 3DV experts with explanatory rationales to ensure precise, unambiguous grounding. Leveraging its diverse data sources and holistic task coverage, MMSI-Video-Bench also supports three domain-oriented sub-benchmarks (Indoor Scene Perception Bench, Robot Bench and Grounding Bench) for targeted capability assessment. We evaluate 25 strong open-source and proprietary MLLMs, revealing a striking human--AI gap: many models perform near chance, and the best reasoning model lags humans by nearly 60%. We further find that spatially fine-tuned models still fail to generalize effectively on our benchmark. Fine-grained error analysis exposes systematic failures in geometric reasoning, motion grounding, long-horizon prediction, and cross-video correspondence. We also show that typical frame-sampling strategies transfer poorly to our reasoning-intensive benchmark, and that neither 3D spatial cues nor chain-of-thought prompting yields meaningful gains. We expect our benchmark to establish a solid testbed for advancing video-based spatial intelligence.

Xiangtai Li, Haobo Yuan, Wenwei Zhang et al.

Video segmentation aims to segment and track every pixel in diverse scenarios accurately. In this paper, we present Tube-Link, a versatile framework that addresses multiple core tasks of video segmentation with a unified architecture. Our framework is a near-online approach that takes a short subclip as input and outputs the corresponding spatial-temporal tube masks. To enhance the modeling of cross-tube relationships, we propose an effective way to perform tube-level linking via attention along the queries. In addition, we introduce temporal contrastive learning to instance-wise discriminative features for tube-level association. Our approach offers flexibility and efficiency for both short and long video inputs, as the length of each subclip can be varied according to the needs of datasets or scenarios. Tube-Link outperforms existing specialized architectures by a significant margin on five video segmentation datasets. Specifically, it achieves almost 13% relative improvements on VIPSeg and 4% improvements on KITTI-STEP over the strong baseline Video K-Net. When using a ResNet50 backbone on Youtube-VIS-2019 and 2021, Tube-Link boosts IDOL by 3% and 4%, respectively.

Sizhe Yang, Linning Xu, Hao Li et al.

3D spatial perception is fundamental to generalizable robotic manipulation, yet obtaining reliable, high-quality 3D geometry remains challenging. Depth sensors suffer from noise and material sensitivity, while existing reconstruction models lack the precision and metric consistency required for physical interaction. We introduce Robo3R, a feed-forward, manipulation-ready 3D reconstruction model that predicts accurate, metric-scale scene geometry directly from RGB images and robot states in real time. Robo3R jointly infers scale-invariant local geometry and relative camera poses, which are unified into the scene representation in the canonical robot frame via a learned global similarity transformation. To meet the precision demands of manipulation, Robo3R employs a masked point head for sharp, fine-grained point clouds, and a keypoint-based Perspective-n-Point (PnP) formulation to refine camera extrinsics and global alignment. Trained on Robo3R-4M, a curated large-scale synthetic dataset with four million high-fidelity annotated frames, Robo3R consistently outperforms state-of-the-art reconstruction methods and depth sensors. Across downstream tasks including imitation learning, sim-to-real transfer, grasp synthesis, and collision-free motion planning, we observe consistent gains in performance, suggesting the promise of this alternative 3D sensing module for robotic manipulation.

Yuchun Feng, Jinliang Zheng, Zhihao Wang et al. · tsinghua

The specification of the action space plays a pivotal role in imitation-based robotic manipulation policy learning, fundamentally shaping the optimization landscape of policy learning. While recent advances have focused heavily on scaling training data and model capacity, the choice of action space remains guided by ad-hoc heuristics or legacy designs, leading to an ambiguous understanding of robotic policy design philosophies. To address this ambiguity, we conducted a large-scale and systematic empirical study, confirming that the action space does have significant and complex impacts on robotic policy learning. We dissect the action design space along temporal and spatial axes, facilitating a structured analysis of how these choices govern both policy learnability and control stability. Based on 13,000+ real-world rollouts on a bimanual robot and evaluation on 500+ trained models over four scenarios, we examine the trade-offs between absolute vs. delta representations, and joint-space vs. task-space parameterizations. Our large-scale results suggest that properly designing the policy to predict delta actions consistently improves performance, while joint-space and task-space representations offer complementary strengths, favoring control stability and generalization, respectively.

Haotian Liang, Xinyi Chen, Bin Wang et al.

A generalist robotic policy needs both semantic understanding for task planning and the ability to interact with the environment through predictive capabilities. To tackle this, we present MM-ACT, a unified Vision-Language-Action (VLA) model that integrates text, image, and action in shared token space and performs generation across all three modalities. MM-ACT adopts a re-mask parallel decoding strategy for text and image generation, and employs a one-step parallel decoding strategy for action generation to improve efficiency. We introduce Context-Shared Multimodal Learning, a unified training paradigm that supervises generation in all three modalities from a shared context, enhancing action generation through cross-modal learning. Experiments were conducted on the LIBERO simulation and Franka real-robot setups as well as RoboTwin2.0 to assess in-domain and out-of-domain performances respectively. Our approach achieves a success rate of 96.3% on LIBERO, 72.0% across three tasks of real Franka, and 52.38% across eight bimanual tasks of RoboTwin2.0 with an additional gain of 9.25% from cross-modal learning. We release our codes, models and data at https://github.com/HHYHRHY/MM-ACT.

Sihao Lin, Zerui Li, Xunyi Zhao et al.

Despite remarkable progress in Vision-Language Navigation (VLN), existing benchmarks remain confined to fixed, small-scale datasets with naive physical simulation. These shortcomings limit the insight that the benchmarks provide into sim-to-real generalization, and create a significant research gap. Furthermore, task fragmentation prevents unified/shared progress in the area, while limited data scales fail to meet the demands of modern LLM-based pretraining. To overcome these limitations, we introduce VLNVerse: a new large-scale, extensible benchmark designed for Versatile, Embodied, Realistic Simulation, and Evaluation. VLNVerse redefines VLN as a scalable, full-stack embodied AI problem. Its Versatile nature unifies previously fragmented tasks into a single framework and provides an extensible toolkit for researchers. Its Embodied design moves beyond intangible and teleporting "ghost" agents that support full-kinematics in a Realistic Simulation powered by a robust physics engine. We leverage the scale and diversity of VLNVerse to conduct a comprehensive Evaluation of existing methods, from classic models to MLLM-based agents. We also propose a novel unified multi-task model capable of addressing all tasks within the benchmark. VLNVerse aims to narrow the gap between simulated navigation and real-world generalization, providing the community with a vital tool to boost research towards scalable, general-purpose embodied locomotion agents.

Xiaoyu Yue, Lei Bai, Meng Wei et al.

Masked AutoEncoder (MAE) has revolutionized the field of self-supervised learning with its simple yet effective masking and reconstruction strategies. However, despite achieving state-of-the-art performance across various downstream vision tasks, the underlying mechanisms that drive MAE's efficacy are less well-explored compared to the canonical contrastive learning paradigm. In this paper, we first propose a local perspective to explicitly extract a local contrastive form from MAE's reconstructive objective at the patch level. And then we introduce a new empirical framework, called Local Contrastive MAE (LC-MAE), to analyze both reconstructive and contrastive aspects of MAE. LC-MAE reveals that MAE learns invariance to random masking and ensures distribution consistency between the learned token embeddings and the original images. Furthermore, we dissect the contribution of the decoder and random masking to MAE's success, revealing both the decoder's learning mechanism and the dual role of random masking as data augmentation and effective receptive field restriction. Our experimental analysis sheds light on the intricacies of MAE and summarizes some useful design methodologies, which can inspire more powerful visual self-supervised methods.

Tai Wang, Xiaohan Mao, Chenming Zhu et al.

In the realm of computer vision and robotics, embodied agents are expected to explore their environment and carry out human instructions. This necessitates the ability to fully understand 3D scenes given their first-person observations and contextualize them into language for interaction. However, traditional research focuses more on scene-level input and output setups from a global view. To address the gap, we introduce EmbodiedScan, a multi-modal, ego-centric 3D perception dataset and benchmark for holistic 3D scene understanding. It encompasses over 5k scans encapsulating 1M ego-centric RGB-D views, 1M language prompts, 160k 3D-oriented boxes spanning over 760 categories, some of which partially align with LVIS, and dense semantic occupancy with 80 common categories. Building upon this database, we introduce a baseline framework named Embodied Perceptron. It is capable of processing an arbitrary number of multi-modal inputs and demonstrates remarkable 3D perception capabilities, both within the two series of benchmarks we set up, i.e., fundamental 3D perception tasks and language-grounded tasks, and in the wild. Codes, datasets, and benchmarks will be available at https://github.com/OpenRobotLab/EmbodiedScan.

Junli Ren, Junfeng Long, Tao Huang et al.

We present a reinforcement learning framework for autonomous goalkeeping with humanoid robots in real-world scenarios. While prior work has demonstrated similar capabilities on quadrupedal platforms, humanoid goalkeeping introduces two critical challenges: (1) generating natural, human-like whole-body motions, and (2) covering a wider guarding range with an equivalent response time. Unlike existing approaches that rely on separate teleoperation or fixed motion tracking for whole-body control, our method learns a single end-to-end RL policy, enabling fully autonomous, highly dynamic, and human-like robot-object interactions. To achieve this, we integrate multiple human motion priors conditioned on perceptual inputs into the RL training via an adversarial scheme. We demonstrate the effectiveness of our method through real-world experiments, where the humanoid robot successfully performs agile, autonomous, and naturalistic interceptions of fast-moving balls. In addition to goalkeeping, we demonstrate the generalization of our approach through tasks such as ball escaping and grabbing. Our work presents a practical and scalable solution for enabling highly dynamic interactions between robots and moving objects, advancing the field toward more adaptive and lifelike robotic behaviors.

Xiaoyang Lyu, Yang-Tian Sun, Yi-Hua Huang et al. · stanford

In this paper, we present an implicit surface reconstruction method with 3D Gaussian Splatting (3DGS), namely 3DGSR, that allows for accurate 3D reconstruction with intricate details while inheriting the high efficiency and rendering quality of 3DGS. The key insight is incorporating an implicit signed distance field (SDF) within 3D Gaussians to enable them to be aligned and jointly optimized. First, we introduce a differentiable SDF-to-opacity transformation function that converts SDF values into corresponding Gaussians' opacities. This function connects the SDF and 3D Gaussians, allowing for unified optimization and enforcing surface constraints on the 3D Gaussians. During learning, optimizing the 3D Gaussians provides supervisory signals for SDF learning, enabling the reconstruction of intricate details. However, this only provides sparse supervisory signals to the SDF at locations occupied by Gaussians, which is insufficient for learning a continuous SDF. Then, to address this limitation, we incorporate volumetric rendering and align the rendered geometric attributes (depth, normal) with those derived from 3D Gaussians. This consistency regularization introduces supervisory signals to locations not covered by discrete 3D Gaussians, effectively eliminating redundant surfaces outside the Gaussian sampling range. Our extensive experimental results demonstrate that our 3DGSR method enables high-quality 3D surface reconstruction while preserving the efficiency and rendering quality of 3DGS. Besides, our method competes favorably with leading surface reconstruction techniques while offering a more efficient learning process and much better rendering qualities. The code will be available at https://github.com/CVMI-Lab/3DGSR.

Qingwei Ben, Feiyu Jia, Jia Zeng et al.

Generalizable humanoid loco-manipulation poses significant challenges, requiring coordinated whole-body control and precise, contact-rich object manipulation. To address this, this paper introduces HOMIE, a semi-autonomous teleoperation system that combines a reinforcement learning policy for body control mapped to a pedal, an isomorphic exoskeleton arm for arm control, and motion-sensing gloves for hand control, forming a unified cockpit to freely operate humanoids and establish a data flywheel. The policy incorporates novel designs, including an upper-body pose curriculum, a height-tracking reward, and symmetry utilization. These features enable the system to perform walking and squatting to specific heights while seamlessly adapting to arbitrary upper-body poses. The exoskeleton, by eliminating the reliance on inverse dynamics, delivers faster and more precise arm control. The gloves utilize Hall sensors instead of servos, allowing even compact devices to achieve 15 or more degrees of freedom and freely adapt to any model of dexterous hands. Compared to previous teleoperation systems, HOMIE stands out for its exceptional efficiency, completing tasks in half the time; its expanded working range, allowing users to freely reach high and low areas as well as interact with any objects; and its affordability, with a price of just $500. The system is fully open-source, demos and code can be found in our https://homietele.github.io/.

Yufei Xue, YunFeng Lin, Wentao Dong et al.

Learning-based whole-body controllers have become a key driver for humanoid robots, yet most existing approaches require robot-specific training. In this paper, we study the problem of cross-embodiment humanoid control and show that a single policy can robustly generalize across a wide range of humanoid robot designs with one-time training. We introduce XHugWBC, a novel cross-embodiment training framework that enables generalist humanoid control through: (1) physics-consistent morphological randomization, (2) semantically aligned observation and action spaces across diverse humanoid robots, and (3) effective policy architectures modeling morphological and dynamical properties. XHugWBC is not tied to any specific robot. Instead, it internalizes a broad distribution of morphological and dynamical characteristics during training. By learning motion priors from diverse randomized embodiments, the policy acquires a strong structural bias that supports zero-shot transfer to previously unseen robots. Experiments on twelve simulated humanoids and seven real-world robots demonstrate the strong generalization and robustness of the resulting universal controller.

Jiahe Chen, ZiRui Wang, Feiyu Jia et al.

Whole-body Humanoid-Object Interaction (HOI) is bottlenecked by the scarcity of high-fidelity 3D data. While video generative priors offer a promising alternative, existing methods suffer from \textit{Representation Misalignment} due to their reliance on geometric priors (e.g., explicit CAD models), and \textit{Retargeting Complexity} arising from intensive morphing and morphological mismatch. We propose Imagine2Real, a zero-shot HOI framework for flexible, geometry-free interaction. To resolve misalignment, we formulate robot and object motions as unified 4D point trajectories. To overcome retargeting complexity, our Keypoints Tracker tracks only sparse critical points (base, hands, and object), entirely bypassing the error-amplifying retargeting process. To maintain natural gaits despite these sparse signals, we utilize the latent space of a Behavior Foundation Model (BFM) as the tracker's search domain. Using a progressive training strategy, Imagine2Real learns robust behaviors with simple tracking rewards, enabling zero-shot physical deployment within a motion capture(mocap) system.

Runsen Xu, Zhiwei Huang, Tai Wang et al.

3D visual grounding is crucial for robots, requiring integration of natural language and 3D scene understanding. Traditional methods depending on supervised learning with 3D point clouds are limited by scarce datasets. Recently zero-shot methods leveraging LLMs have been proposed to address the data issue. While effective, these methods only use object-centric information, limiting their ability to handle complex queries. In this work, we present VLM-Grounder, a novel framework using vision-language models (VLMs) for zero-shot 3D visual grounding based solely on 2D images. VLM-Grounder dynamically stitches image sequences, employs a grounding and feedback scheme to find the target object, and uses a multi-view ensemble projection to accurately estimate 3D bounding boxes. Experiments on ScanRefer and Nr3D datasets show VLM-Grounder outperforms previous zero-shot methods, achieving 51.6% Acc@0.25 on ScanRefer and 48.0% Acc on Nr3D, without relying on 3D geometry or object priors. Codes are available at https://github.com/OpenRobotLab/VLM-Grounder .

Haifeng Huang, Yilun Chen, Zehan Wang et al.

Recent advancements in multi-modal large language models (MLLMs) have shown strong potential for 3D scene understanding. However, existing methods struggle with fine-grained object grounding and contextual reasoning, limiting their ability to interpret and interact with complex 3D environments. In this paper, we present Chat-Scene++, an MLLM framework that represents 3D scenes as context-rich object sequences. By structuring scenes as sequences of objects with contextual semantics, Chat-Scene++ enables object-centric representation and interaction. It decomposes a 3D scene into object representations paired with identifier tokens, allowing LLMs to follow instructions across diverse 3D vision-language tasks. To capture inter-object relationships and global semantics, Chat-Scene++ extracts context-rich object features using large-scale pre-trained 3D scene-level and 2D image-level encoders, unlike the isolated per-object features in Chat-Scene. Its flexible object-centric design also supports grounded chain-of-thought (G-CoT) reasoning, enabling the model to distinguish objects at both category and spatial levels during multi-step inference. Without the need for additional task-specific heads or fine-tuning, Chat-Scene++ achieves state-of-the-art performance on five major 3D vision-language benchmarks: ScanRefer, Multi3DRefer, Scan2Cap, ScanQA, and SQA3D. These results highlight its effectiveness in scene comprehension, object grounding, and spatial reasoning. Additionally, without reconstructing 3D worlds through computationally expensive processes, we demonstrate its applicability to real-world scenarios using only 2D inputs.

Mu Huang, Hui Wang, Kerui Ren et al.

Simulating deformable objects under rich interactions remains a fundamental challenge for real-to-sim robot manipulation, with dynamics jointly driven by environmental effects and robot actions. Existing simulators rely on predefined physics or data-driven dynamics without robot-conditioned control, limiting accuracy, stability, and generalization. This paper presents SoMA, a 3D Gaussian Splat simulator for soft-body manipulation. SoMA couples deformable dynamics, environmental forces, and robot joint actions in a unified latent neural space for end-to-end real-to-sim simulation. Modeling interactions over learned Gaussian splats enables controllable, stable long-horizon manipulation and generalization beyond observed trajectories without predefined physical models. SoMA improves resimulation accuracy and generalization on real-world robot manipulation by 20%, enabling stable simulation of complex tasks such as long-horizon cloth folding.

Jiaqi Peng, Wenzhe Cai, Yuqiang Yang et al.

Trajectory planning in unstructured environments is a fundamental and challenging capability for mobile robots. Traditional modular pipelines suffer from latency and cascading errors across perception, localization, mapping, and planning modules. Recent end-to-end learning methods map raw visual observations directly to control signals or trajectories, promising greater performance and efficiency in open-world settings. However, most prior end-to-end approaches still rely on separate localization modules that depend on accurate sensor extrinsic calibration for self-state estimation, thereby limiting generalization across embodiments and environments. We introduce LoGoPlanner, a localization-grounded, end-to-end navigation framework that addresses these limitations by: (1) finetuning a long-horizon visual-geometry backbone to ground predictions with absolute metric scale, thereby providing implicit state estimation for accurate localization; (2) reconstructing surrounding scene geometry from historical observations to supply dense, fine-grained environmental awareness for reliable obstacle avoidance; and (3) conditioning the policy on implicit geometry bootstrapped by the aforementioned auxiliary tasks, thereby reducing error propagation. We evaluate LoGoPlanner in both simulation and real-world settings, where its fully end-to-end design reduces cumulative error while metric-aware geometry memory enhances planning consistency and obstacle avoidance, leading to more than a 27.3\% improvement over oracle-localization baselines and strong generalization across embodiments and environments. The code and models have been made publicly available on the https://steinate.github.io/logoplanner.github.io.

Yinhuai Wang, Qihan Zhao, Yuen Fui Lau et al.

Enabling humanoid robots to perform agile and adaptive interactive tasks has long been a core challenge in robotics. Current approaches are bottlenecked by either the scarcity of realistic interaction data or the need for meticulous, task-specific reward engineering, which limits their scalability. To narrow this gap, we present HumanX, a full-stack framework that compiles human video into generalizable, real-world interaction skills for humanoids, without task-specific rewards. HumanX integrates two co-designed components: XGen, a data generation pipeline that synthesizes diverse and physically plausible robot interaction data from video while supporting scalable data augmentation; and XMimic, a unified imitation learning framework that learns generalizable interaction skills. Evaluated across five distinct domains--basketball, football, badminton, cargo pickup, and reactive fighting--HumanX successfully acquires 10 different skills and transfers them zero-shot to a physical Unitree G1 humanoid. The learned capabilities include complex maneuvers such as pump-fake turnaround fadeaway jumpshots without any external perception, as well as interactive tasks like sustained human-robot passing sequences over 10 consecutive cycles--learned from a single video demonstration. Our experiments show that HumanX achieves over 8 times higher generalization success than prior methods, demonstrating a scalable and task-agnostic pathway for learning versatile, real-world robot interactive skills.

Sihan Yang, Runsen Xu, Yiman Xie et al. · pku

Spatial intelligence is essential for multimodal large language models (MLLMs) operating in the complex physical world. Existing benchmarks, however, probe only single-image relations and thus fail to assess the multi-image spatial reasoning that real-world deployments demand. We introduce MMSI-Bench, a VQA benchmark dedicated to multi-image spatial intelligence. Six 3D-vision researchers spent more than 300 hours meticulously crafting 1,000 challenging, unambiguous multiple-choice questions from over 120,000 images, each paired with carefully designed distractors and a step-by-step reasoning process. We conduct extensive experiments and thoroughly evaluate 34 open-source and proprietary MLLMs, observing a wide gap: the strongest open-source model attains roughly 30% accuracy and OpenAI's o3 reasoning model reaches 40%, while humans score 97%. These results underscore the challenging nature of MMSI-Bench and the substantial headroom for future research. Leveraging the annotated reasoning processes, we also provide an automated error analysis pipeline that diagnoses four dominant failure modes, including (1) grounding errors, (2) overlap-matching and scene-reconstruction errors, (3) situation-transformation reasoning errors, and (4) spatial-logic errors, offering valuable insights for advancing multi-image spatial intelligence. Project page: https://runsenxu.com/projects/MMSI_Bench .

Kerui Ren, Guanghao Li, Changjian Jiang et al.

Streaming reconstruction from uncalibrated monocular video remains challenging, as it requires both high-precision pose estimation and computationally efficient online refinement in dynamic environments. While coupling 3D foundation models with SLAM frameworks is a promising paradigm, a critical bottleneck persists: most multi-view foundation models estimate poses in a feed-forward manner, yielding pixel-level correspondences that lack the requisite precision for rigorous geometric optimization. To address this, we present M^3, which augments the Multi-view foundation model with a dedicated Matching head to facilitate fine-grained dense correspondences and integrates it into a robust Monocular Gaussian Splatting SLAM. M^3 further enhances tracking stability by incorporating dynamic area suppression and cross-inference intrinsic alignment. Extensive experiments on diverse indoor and outdoor benchmarks demonstrate state-of-the-art accuracy in both pose estimation and scene reconstruction. Notably, M^3 reduces ATE RMSE by 64.3% compared to VGGT-SLAM 2.0 and outperforms ARTDECO by 2.11 dB in PSNR on the ScanNet++ dataset.

Xijie Yang, Mulin Yu, Changjian Jiang et al.

Recent reconstruction methods based on radiance field such as NeRF and 3DGS reproduce indoor scenes with high visual fidelity, but break down under scene editing due to baked illumination and the lack of explicit light transport. In contrast, physically based inverse rendering relies on mesh representations and path tracing, which enforce correct light transport but place strong requirements on geometric fidelity, becoming a practical bottleneck for real indoor scenes. In this work, we propose Emission-Aware Gaussians and Path Tracing (EAG-PT), aiming for physically based light transport with a unified 2D Gaussian representation. Our design is based on three cores: (1) using 2D Gaussians as a unified scene representation and transport-friendly geometry proxy that avoids reconstructed mesh, (2) explicitly separating emissive and non-emissive components during reconstruction for further scene editing, and (3) decoupling reconstruction from final rendering by using efficient single-bounce optimization and high-quality multi-bounce path tracing after scene editing. Experiments on synthetic and real indoor scenes show that EAG-PT produces more natural and physically consistent renders after editing than radiant scene reconstructions, while preserving finer geometric detail and avoiding mesh-induced artifacts compared to mesh-based inverse path tracing. These results suggest promising directions for future use in interior design, XR content creation, and embodied AI.

Weiguang Zhao, Haoran Xu, Xingyu Miao et al.

Point tracking aims to follow visual points through complex motion, occlusion, and viewpoint changes, and has advanced rapidly with modern foundation models. Yet progress toward general point tracking remains constrained by limited high-quality data, as existing datasets often provide insufficient diversity and imperfect trajectory annotations. To this end, we introduce SynthVerse, a large-scale, diverse synthetic dataset specifically designed for point tracking. SynthVerse includes several new domains and object types missing from existing synthetic datasets, such as animated-film-style content, embodied manipulation, scene navigation, and articulated objects. SynthVerse substantially expands dataset diversity by covering a broader range of object categories and providing high-quality dynamic motions and interactions, enabling more robust training and evaluation for general point tracking. In addition, we establish a highly diverse point tracking benchmark to systematically evaluate state-of-the-art methods under broader domain shifts. Extensive experiments and analyses demonstrate that training with SynthVerse yields consistent improvements in generalization and reveal limitations of existing trackers under diverse settings.

Zhixuan Liang, Yizhuo Li, Tianshuo Yang et al.

Vision-Language-Action (VLA) models adapt large vision-language backbones to map images and instructions into robot actions. However, prevailing VLAs either generate actions auto-regressively in a fixed left-to-right order or attach separate MLP or diffusion heads outside the backbone, leading to fragmented information pathways and specialized training requirements that hinder a unified, scalable architecture. We present Discrete Diffusion VLA, a unified-transformer policy that models discretized action chunks with discrete diffusion. The design retains diffusion's progressive refinement paradigm while remaining natively compatible with the discrete token interface of VLMs. Our method achieves an adaptive decoding order that resolves easy action elements before harder ones and uses secondary re-masking to revisit uncertain predictions across refinement rounds, which improves consistency and enables robust error correction. This unified decoder preserves pre-trained vision-language priors, supports parallel decoding, breaks the autoregressive bottleneck, and reduces the number of function evaluations. Discrete Diffusion VLA achieves 96.3% avg. success rates on LIBERO, 71.2% visual matching on SimplerEnv-Fractal and 54.2% overall on SimplerEnv-Bridge, improving over autoregressive, MLP decoder and continuous diffusion baselines. These findings indicate that discrete-diffusion VLA supports precise action modeling and consistent training, laying groundwork for scaling VLA to larger models and datasets. Our project page is https://github.com/Liang-ZX/DiscreteDiffusionVLA

Tianshuo Yang, Guanyu Chen, Yutian Chen et al.

While end-to-end Vision-Language-Action (VLA) models offer a promising paradigm for robotic manipulation, fine-tuning them on narrow control data often compromises the profound reasoning capabilities inherited from their base Vision-Language Models (VLMs). To resolve this fundamental trade-off, we propose HiVLA, a visual-grounded-centric hierarchical framework that explicitly decouples high-level semantic planning from low-level motor control. In high-level part, a VLM planner first performs task decomposition and visual grounding to generate structured plans, comprising a subtask instruction and a precise target bounding box. Then, to translate this plan into physical actions, we introduce a flow-matching Diffusion Transformer (DiT) action expert in low-level part equipped with a novel cascaded cross-attention mechanism. This design sequentially fuses global context, high-resolution object-centric crops and skill semantics, enabling the DiT to focus purely on robust execution. Our decoupled architecture preserves the VLM's zero-shot reasoning while allowing independent improvement of both components. Extensive experiments in simulation and the real world demonstrate that HiVLA significantly outperforms state-of-the-art end-to-end baselines, particularly excelling in long-horizon skill composition and the fine-grained manipulation of small objects in cluttered scenes.

Jiahe Chen, Jinkun Cao, Dahua Lin et al.

To predict future trajectories, the normalizing flow with a standard Gaussian prior suffers from weak diversity. The ineffectiveness comes from the conflict between the fact of asymmetric and multi-modal distribution of likely outcomes and symmetric and single-modal original distribution and supervision losses. Instead, we propose constructing a mixed Gaussian prior for a normalizing flow model for trajectory prediction. The prior is constructed by analyzing the trajectory patterns in the training samples without requiring extra annotations while showing better expressiveness and being multi-modal and asymmetric. Besides diversity, it also provides better controllability for probabilistic trajectory generation. We name our method Mixed Gaussian Flow (MGF). It achieves state-of-the-art performance in the evaluation of both trajectory alignment and diversity on the popular UCY/ETH and SDD datasets. Code is available at https://github.com/mulplue/MGF.

Xiaofeng Mao, Zhen Li, Chuanhao Li et al.

Recent approaches have demonstrated the promise of using diffusion models to generate interactive and explorable worlds. However, most of these methods face critical challenges such as excessively large parameter sizes, reliance on lengthy inference steps, and rapidly growing historical context, which severely limit real-time performance and lack text-controlled generation capabilities. To address these challenges, we propose \method, a novel framework designed to generate realistic, interactive, and continuous worlds from a single image or text prompt. \method achieves this through a carefully designed framework that supports keyboard-based exploration of the generated worlds. The framework comprises three core components: (1) a long-video generation framework integrating unified context compression with linear attention; (2) a real-time streaming acceleration strategy powered by bidirectional attention distillation and an enhanced text embedding scheme; (3) a text-controlled method for generating world events. We have provided the codebase in the supplementary material.

Haozhan Li, Yuxin Zuo, Jiale Yu et al. · pku, tsinghua

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 $π_0$ 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

Feiyu Jia, Xiaojie Niu, Sizhe Yang et al.

Teleoperation is a key approach for collecting high-quality, physically consistent demonstrations for robotic manipulation. However, teleoperation for dexterous manipulation remains constrained by: (i) inaccurate hand-robot motion mapping, which limits teleoperated dexterity, and (ii) limited tactile feedback that forces vision-dominated interaction and hinders perception of contact geometry and force variation. To address these challenges, we present TAG, a low-cost glove system that integrates precise hand motion capture with high-resolution tactile feedback, enabling effective tactile-in-the-loop dexterous teleoperation. For motion capture, TAG employs a non-contact magnetic sensing design that provides drift-free, electromagnetically robust 21-DoF joint tracking with joint angle estimation errors below 1 degree. Meanwhile, to restore tactile sensation, TAG equips each finger with a 32-actuator tactile array within a compact 2 cm^2 module, allowing operators to directly feel physical interactions at the robot end-effector through spatial activation patterns. Through real-world teleoperation experiments and user studies, we show that TAG enables reliable real-time perception of contact geometry and dynamic force, improves success rates in contact-rich teleoperation tasks, and increases the reliability of demonstration data collection for learning-based manipulation.

Juncheng Mu, Sizhe Yang, Hojin Bae et al.

Cross-embodiment manipulation is crucial for enhancing the scalability of robot manipulation and reducing the high cost of data collection. However, the significant differences between embodiments, such as variations in action spaces and structural disparities, pose challenges for joint training across multiple sources of data. To address this, we propose One-Policy-Fits-All (OPFA), a framework that enables learning a single, versatile policy across multiple embodiments. We first learn a Geometry-Aware Latent Representation (GaLR), which leverages 3D convolution networks and transformers to build a shared latent action space across different embodiments. Then we design a unified latent retargeting decoder that extracts embodiment-specific actions from the latent representations, without any embodiment-specific decoder tuning. OPFA enables end-to-end co-training of data from diverse embodiments, including various grippers and dexterous hands with arbitrary degrees of freedom, significantly improving data efficiency and reducing the cost of skill transfer. We conduct extensive experiments across 11 different end-effectors. The results demonstrate that OPFA significantly improves policy performance in diverse settings by leveraging heterogeneous embodiment data. For instance, cross-embodiment co-training can improve success rates by more than 50% compared to single-source training. Moreover, by adding only a few demonstrations from a new embodiment (e.g., eight), OPFA can achieve performance comparable to that of a well-trained model with 72 demonstrations.

Zhen Luo, Yixuan Yang, Xudong Xu et al.

Generating simulation-ready tabletop scenes from task instructions is an intriguing and promising research direction in the field of Embodied AI. However, existing task-to-scene generation methods rely exclusively on large language models (LLMs) to predict scene layouts, inevitably yielding object collisions or floating due to LLMs' inherent limitations in 3D spatial reasoning. In this paper, we present STABLE, a semantics-physics dual-system tailored for simulation-ready tabletop scene generation. STABLE consists of two complementary modules: (i) a Semantic Reasoner, a fine-tuned LLM trained on a structured tabletop scene dataset to generate coarse layouts from input task instructions, and (ii) a Physics Corrector, a physics-aware flow-based denoising model that outputs pose updates to refine layouts, which ensures the physical plausibility of scenes while preserves semantic alignment with task instructions. STABLE adopts a progressive generation paradigm: by alternating between the Semantic Reasoner and Physics Corrector, it incrementally expands the scene from task-critical objects to background objects. Experiments demonstrate that STABLE successfully generates simulation-ready tabletop scenes that strictly conform to task instructions and significantly enhances the physical validity of scenes over prior art.

Xinyi Chen, Yilun Chen, Yanwei Fu et al.

We introduce InternVLA-M1, a unified framework for spatial grounding and robot control that advances instruction-following robots toward scalable, general-purpose intelligence. Its core idea is spatially guided vision-language-action training, where spatial grounding serves as the critical link between instructions and robot actions. InternVLA-M1 employs a two-stage pipeline: (i) spatial grounding pre-training on over 2.3M spatial reasoning data to determine ``where to act'' by aligning instructions with visual, embodiment-agnostic positions, and (ii) spatially guided action post-training to decide ``how to act'' by generating embodiment-aware actions through plug-and-play spatial prompting. This spatially guided training recipe yields consistent gains: InternVLA-M1 outperforms its variant without spatial guidance by +14.6% on SimplerEnv Google Robot, +17% on WidowX, and +4.3% on LIBERO Franka, while demonstrating stronger spatial reasoning capability in box, point, and trace prediction. To further scale instruction following, we built a simulation engine to collect 244K generalizable pick-and-place episodes, enabling a 6.2% average improvement across 200 tasks and 3K+ objects. In real-world clustered pick-and-place, InternVLA-M1 improved by 7.3%, and with synthetic co-training, achieved +20.6% on unseen objects and novel configurations. Moreover, in long-horizon reasoning-intensive scenarios, it surpassed existing works by over 10%. These results highlight spatially guided training as a unifying principle for scalable and resilient generalist robots. Code and models are available at https://github.com/InternRobotics/InternVLA-M1.

Xiaofeng Mao, Shaoheng Lin, Zhen Li et al.

Yume aims to use images, text, or videos to create an interactive, realistic, and dynamic world, which allows exploration and control using peripheral devices or neural signals. In this report, we present a preview version of \method, which creates a dynamic world from an input image and allows exploration of the world using keyboard actions. To achieve this high-fidelity and interactive video world generation, we introduce a well-designed framework, which consists of four main components, including camera motion quantization, video generation architecture, advanced sampler, and model acceleration. First, we quantize camera motions for stable training and user-friendly interaction using keyboard inputs. Then, we introduce the Masked Video Diffusion Transformer~(MVDT) with a memory module for infinite video generation in an autoregressive manner. After that, training-free Anti-Artifact Mechanism (AAM) and Time Travel Sampling based on Stochastic Differential Equations (TTS-SDE) are introduced to the sampler for better visual quality and more precise control. Moreover, we investigate model acceleration by synergistic optimization of adversarial distillation and caching mechanisms. We use the high-quality world exploration dataset \sekai to train \method, and it achieves remarkable results in diverse scenes and applications. All data, codebase, and model weights are available on https://github.com/stdstu12/YUME. Yume will update monthly to achieve its original goal. Project page: https://stdstu12.github.io/YUME-Project/.

Yang Li, Zhaxizhuoma, Hongru Jiang et al.

Embodied intelligence for contact-rich manipulation has predominantly relied on position control, while explicit awareness and regulation of interaction forces remain under-explored, limiting stability, precision, and robustness in real-world tasks. We propose ForceVLA2, an end-to-end vision-language-action framework that equips robots with hybrid force-position control and explicit force awareness. ForceVLA2 introduces force-based prompts into the VLM expert to construct force-aware task concepts across stages, and employs a Cross-Scale Mixture-of-Experts (MoE) in the action expert to adaptively fuse these concepts with real-time interaction forces for closed-loop hybrid force-position regulation. To support learning and evaluation, we construct ForceVLA2-Dataset, containing 1,000 trajectories over 5 contact-rich tasks, including wiping, pressing, and assembling, with multi-view images, task prompts, proprioceptive state, and force signals. Extensive experiments show that ForceVLA2 substantially improves success rates and reliability in contact-rich manipulation, outperforming pi0 and pi0.5 by 48.0% and 35.0%, respectively, across the 5 tasks, and mitigating common failure modes such as arm overload and unstable contact, thereby actively advancing force-aware interactive physical intelligence in VLAs. The project page is available at https://sites.google.com/view/force-vla2/home.