Collecting high-quality data for training large-scale robotic models typically relies on real robot platforms, which is labor-intensive and costly, whether via teleoperation or scripted demonstrations. To scale data collection, many researchers have turned to leveraging human manipulation videos available online. However, current methods predominantly focus on hand detection or object pose estimation, failing to fully exploit the rich interaction cues embedded in these videos. In this work, we propose a novel approach that combines large foundation models for video understanding with point tracking techniques to extract dense trajectories of all task-relevant keypoints during manipulation. This enables more comprehensive utilization of Internet-scale human demonstration videos. Experimental results demonstrate that our method can accurately track keypoints throughout the entire manipulation process, paving the way for more scalable and data-efficient robot learning.

Coordinating a team of robots to reposition multiple objects in cluttered environments requires reasoning jointly about where robots should establish contact, how to manipulate objects once contact is made, and how to navigate safely and efficiently at scale. Prior approaches typically fall into two extremes -- either learning the entire task or relying on privileged information and hand-designed planners -- both of which struggle to handle diverse objects in long-horizon tasks. To address these challenges, we present a unified framework for collaborative multi-robot, multi-object non-prehensile manipulation that integrates flow-matching co-generation with anonymous multi-robot motion planning. Within this framework, a generative model co-generates contact formations and manipulation trajectories from visual observations, while a novel motion planner conveys robots at scale. Crucially, the same planner also supports coordination at the object level, assigning manipulated objects to larger target structures and thereby unifying robot- and object-level reasoning within a single algorithmic framework. Experiments in challenging simulated environments demonstrate that our approach outperforms baselines in both motion planning and manipulation tasks, highlighting the benefits of generative co-design and integrated planning for scaling collaborative manipulation to complex multi-agent, multi-object settings. Visit gco-paper.github.io for code and demonstrations.

Egocentric videos capture how humans manipulate objects and tools, providing diverse motion cues for learning object manipulation. Unlike the costly, expert-driven manual teleoperation commonly used in training Vision-Language-Action models (VLAs), egocentric videos offer a scalable alternative. However, prior studies that leverage such videos for training robot policies typically rely on auxiliary annotations, such as detailed hand-pose recordings. Consequently, it remains unclear whether VLAs can be trained directly from raw egocentric videos. In this work, we address this challenge by leveraging EgoScaler, a framework that extracts 6DoF object manipulation trajectories from egocentric videos without requiring auxiliary recordings. We apply EgoScaler to four large-scale egocentric video datasets and automatically refine noisy or incomplete trajectories, thereby constructing a new large-scale dataset for VLA pre-training. Our experiments with a state-of-the-art $π_0$ architecture in both simulated and real-robot environments yield three key findings: (i) pre-training on our dataset improves task success rates by over 20\% compared to training from scratch, (ii) the performance is competitive with that achieved using real-robot datasets, and (iii) combining our dataset with real-robot data yields further improvements. These results demonstrate that egocentric videos constitute a promising and scalable resource for advancing VLA research.

-- Object manipulation skills are necessary for robots operating in various daily-life scenarios, ranging from warehouses to hospitals. They allow the robots to manipulate the given object to their desired arrangement in the cluttered environment. The existing approaches to solving object manipulations are either inefficient sampling based techniques, require expert demonstrations, or learn by trial and error, making them less ideal for practical scenarios. In this paper, we propose a novel, multimodal physics-informed neural network (PINN) for solving object manipulation tasks. Our approach efficiently learns to solve the Eikonal equation without expert data and finds object manipulation trajectories fast in complex, cluttered environments. Our method is multimodal as it also reactively replans the robot's grasps during manipulation to achieve the desired object poses. We demonstrate our approach in both simulation and real-world scenarios and compare it against state-of-the-art baseline methods. The results indicate that our approach is effective across various objects, has efficient training compared to previous learning-based methods, and demonstrates high performance in planning time, trajectory length, and success rates. Our demonstration videos can be found at https://youtu.be/FaQLkTV9knI.

Humans subconsciously choose robust ways of selecting and using tools, based on years of embodied experience -- for example, choosing a ladle instead of a flat spatula to serve meatballs. However, robustness under uncertainty remains underexplored in robotic tool-use planning. This paper presents a robustness-aware framework that jointly selects tools and plans contact-rich manipulation trajectories, explicitly optimizing for robustness against environmental disturbances. At the core of our approach is a learned, energy-based robustness metric, which guides the planner towards robust manipulation behaviors. We formulate a hierarchical optimization pipeline that first identifies a tool and configuration that optimizes robustness, and then plans a corresponding manipulation trajectory that maintains robustness throughout execution. We evaluate our approach across three representative tool-use tasks. Simulation and real-world results demonstrate that our approach consistently selects robust tools and generates disturbance-resilient manipulation plans.

Cross-embodiment robotic manipulation synthesis for complicated tasks is challenging, partially due to the scarcity of paired cross-embodiment datasets and the impediment of designing intricate controllers. Inspired by robotic learning via guided human expert demonstration, we here propose a novel cross-embodiment robotic manipulation algorithm via CycleVAE and human behavior transformer. First, we utilize unsupervised CycleVAE together with a bidirectional subspace alignment algorithm to align latent motion sequences between cross-embodiments. Second, we propose a casual human behavior transformer design to learn the intrinsic motion dynamics of human expert demonstrations. During the test case, we leverage the proposed transformer for the human expert demonstration generation, which will be aligned using CycleVAE for the final human-robotic manipulation synthesis. We validated our proposed algorithm through extensive experiments using a dexterous robotic manipulator with the robotic hand. Our results successfully generate smooth trajectories across intricate tasks, outperforming prior learning-based robotic motion planning algorithms. These results have implications for performing unsupervised cross-embodiment alignment and future autonomous robotics design. Complete video demonstrations of our experiments can be found in https://sites.google.com/view/humanrobots/home.

DexTrack learns a generalizable neural tracking controller for dexterous manipulation from human references. It generates hand action commands from kinematic references, ensuring close tracking of input trajectories (Fig. (a)), generalizes to novel and challenging tasks involving thin objects, complex movements and intricate in-hand manipulations (Fig. (b)), and demonstrates robustness to large kinematics noise and utility in real-world scenarios (Fig. (c)). Kinematic references are illustrated in orange rectangles and background. We address the challenge of developing a generalizable neural tracking controller for dexterous manipulation from human references. This controller aims to manage a dexterous robot hand to manipulate diverse objects for various purposes defined by kinematic human-object interactions. Developing such a controller is complicated by the intricate contact dynamics of dexterous manipulation and the need for adaptivity, generalizability, and robustness. Current reinforcement learning and trajectory optimization methods often fall short due to their dependence on task-specific rewards or precise system models. We introduce an approach that curates large-scale successful robot tracking demonstrations, comprising pairs of human references and robot actions, to train a neural controller. Utilizing a data flywheel, we iteratively enhance the controller's performance, as well as the number and quality of successful tracking demonstrations. We exploit available tracking demonstrations and carefully integrate reinforcement learning and imitation learning to boost the controller's performance in dynamic environments. At the same time, to obtain high-quality tracking demonstrations, we individually optimize per-trajectory tracking by leveraging the learned tracking controller in a homotopy optimization method. The homotopy optimization, mimicking chainof-thought, aids in solving challenging trajectory tracking problems to increase demonstration diversity. Our method achieves over a 10% improvement in success rates compared to leading baselines. The project website with animated results is available at DexTrack. Robotic dexterous manipulation refers to the ability of a robot hand skillfully handling and manipulating objects for various target states with precision and adaptability. Achieving human-level robotic dexterous manipulation is challenging due to two main difficulties: the intricate dynamics of contact-rich manipulation, which complicates optimization (Pang & Tedrake, 2021; Pang et al., 2023; Liu et al., 2024a; Jin, 2024), and the need for robots to master a wide range of versatile skills beyond specific tasks.

Robotic manipulation, owing to its multi-modal nature, often faces significant training ambiguity, necessitating explicit instructions to clearly delineate the manipulation details in tasks. In this work, we highlight that vision instruction is naturally more comprehensible to recent robotic policies than the commonly adopted text instruction, as these policies are born with some vision understanding ability like human infants. Building on this premise and drawing inspiration from cognitive science, we introduce the robotic imagery paradigm, which realizes large-scale robotic data pre-training without text annotations. Additionally, we propose the robotic gaze strategy that emulates the human eye gaze mechanism, thereby guiding subsequent actions and focusing the attention of the policy on the manipulated object. Leveraging these innovations, we develop VIRT, a fully Transformer-based policy. We design comprehensive tasks using both a physical robot and simulated environments to assess the efficacy of VIRT. The results indicate that VIRT can complete very competitive tasks like "opening the lid of a tightly sealed bottle", and the proposed techniques boost the success rates of the baseline policy on diverse challenging tasks from nearly 0% to more than 65%. The key insight that supports this work is existing robotic policies are akin to human infants, who are born with visual perception and reasoning abilities but do not comprehend natural language according to previous cognitive science literature (Colombo & Mitchell, 2009). Specifically, visual signal serves as the primary information source of recent robotic policies, and the backbones of these policies are pre-trained with large-scale image datasets before the robotic data based training (He et al., 2016; Oquab et al., 2024). Therefore, the policies begin with a basic visual understanding capability like human infants. By contrast, natural language inputs are rarely incorporated into the process of pre-training these backbones, suggesting the lack of natural language knowledge in these policies.

Visual affordance learning is a key component for robots to understand how to interact with objects. Conventional approaches in this field rely on pre-defined objects and actions, falling short of capturing diverse interactions in realworld scenarios. The key idea of our approach is employing textual instruction, targeting various affordances for a wide range of objects. This approach covers both hand-object and tool-object interactions. We introduce text-driven affordance learning, aiming to learn contact points and manipulation trajectories from an egocentric view following textual instruction. In our task, contact points are represented as heatmaps, and the manipulation trajectory as sequences of coordinates that incorporate both linear and rotational movements for various manipulations. However, when we gather data for this task, manual annotations of these diverse interactions are costly. To this end, we propose a pseudo dataset creation pipeline and build a large pseudo-training dataset: TextAFF80K, consisting of over 80K instances of the contact points, trajectories, images, and text tuples. We extend existing referring expression comprehension models for our task, and experimental results show that our approach robustly handles multiple affordances, serving as a new standard for affordance learning in real-world scenarios.

Dexterous manipulation, particularly adept coordinating and grasping, constitutes a fundamental and indispensable capability for robots, facilitating the emulation of human-like behaviors. Integrating this capability into robots empowers them to supplement and even supplant humans in undertaking increasingly intricate tasks in both daily life and industrial settings. Unfortunately, contemporary methodologies encounter serious challenges in devising manipulation trajectories owing to the intricacies of tasks, the expansive robotic manipulation space, and dynamic obstacles. We propose a novel approach, APEX, to address all these difficulties by introducing a collision-free latent diffusion model for both robotic motion planning and manipulation. Firstly, we simplify the complexity of real-life ambidextrous dual-arm robotic manipulation tasks by abstracting them as aligning two vectors. Secondly, we devise latent diffusion models to produce a variety of robotic manipulation trajectories. Furthermore, we integrate obstacle information utilizing a classifier-guidance technique, thereby guaranteeing both the feasibility and safety of the generated manipulation trajectories. Lastly, we validate our proposed algorithm through extensive experiments conducted on the hardware platform of ambidextrous dual-arm robots. Our algorithm consistently generates successful and seamless trajectories across diverse tasks, surpassing conventional robotic motion planning algorithms. These results carry significant implications for the future design of diffusion robots, enhancing their capability to tackle more intricate robotic manipulation tasks with increased efficiency and safety. Complete video demonstrations of our experiments can be found in https://sites.google.com/view/apex-dual-arm/home.