Search arXiv⌕ Search

SEARCH · Search arXiv

Results for “robot learning”

Search indexed arXiv papers on artificial intelligence, large language models, computer vision and robotics. Read source abstracts and follow links to arXiv.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,549 records · Page 86Linked to original sources
arXiv2024

Scaling Proprioceptive-Visual Learning with Heterogeneous Pre-trained Transformers

One of the roadblocks for training generalist robotic models today is heterogeneity. Previous robot learning methods often collect data to train with one specific embodiment for one task, which is expensive and prone to overfitting. This work studies the problem of learning policy representations through heterogeneous pre-training on robot data across different embodiments and tasks at scale. We propose Heterogeneous Pre-trained Transformers (HPT), which pre-train a large, shareable trunk of a policy neural network to learn a task and embodiment agnostic shared representation. This general architecture aligns the specific proprioception and vision inputs from distinct embodiments to a short sequence of tokens and then processes such tokens to map to control robots for different tasks. Leveraging the recent large-scale multi-embodiment real-world robotic datasets as well as simulation, deployed robots, and human video datasets, we investigate pre-training policies across heterogeneity. We conduct experiments to investigate the scaling behaviors of training objectives, to the extent of 52 datasets. HPTs outperform several baselines and enhance the fine-tuned policy performance by over 20% on unseen tasks in multiple simulator benchmarks and real-world settings. See the project website (https://liruiw.github.io/hpt/) for code and videos.

cs.RO↗
arXiv2026

Collective Learning for Unified Robotic Manipulation via Predictive Dynamics Adaptation

Learning a collective policy for state-based Cartesian manipulation across diverse robotic embodiments presents a fundamental challenge due to embodiment-dependent transition dynamics. Collective learning aggregates multi-robot experience to discover shared task structure, yet naive aggregation can average incompatible action modes when different embodiments realize the same Cartesian command through different transition behavior. This leads to strategy ambiguity and performance collapse on dynamically outlying embodiments. To address this, we propose PACE, a two-stage collective learning framework that separates shared task-space intent from embodiment-dependent transition dynamics. The first stage employs a representation layer pre-trained on offline trajectories to extract task-level context from collective data. The second stage introduces a decision layer equipped with a predictive dynamics adapter. By forcing the latent state to predict future transitions, this adapter acts as a latent regularizer that grounds the policy in the transition behavior of the executing robot while preserving real-time inference without online planning. Evaluations across simulated embodiments demonstrate that our method prevents performance collapse on distinct robots. It achieves over 93\% overall success and 89\% on dynamically out-of-distribution zero-shot simulated robots, outperforming the Transformer behavior cloning (82\%) and diffusion policy baselines (66\%). Physical KUKA deployment further validates the feasibility of the full state-based pipeline under zero-shot sim-to-real transfer.

cs.RO↗
arXiv2023

LARG, Language-based Automatic Reward and Goal Generation

Goal-conditioned and Multi-Task Reinforcement Learning (GCRL and MTRL) address numerous problems related to robot learning, including locomotion, navigation, and manipulation scenarios. Recent works focusing on language-defined robotic manipulation tasks have led to the tedious production of massive human annotations to create dataset of textual descriptions associated with trajectories. To leverage reinforcement learning with text-based task descriptions, we need to produce reward functions associated with individual tasks in a scalable manner. In this paper, we leverage recent capabilities of Large Language Models (LLMs) and introduce \larg, Language-based Automatic Reward and Goal Generation, an approach that converts a text-based task description into its corresponding reward and goal-generation functions We evaluate our approach for robotic manipulation and demonstrate its ability to train and execute policies in a scalable manner, without the need for handcrafted reward functions.

cs.CL↗
arXiv2017

GPLAC: Generalizing Vision-Based Robotic Skills using Weakly Labeled Images

We tackle the problem of learning robotic sensorimotor control policies that can generalize to visually diverse and unseen environments. Achieving broad generalization typically requires large datasets, which are difficult to obtain for task-specific interactive processes such as reinforcement learning or learning from demonstration. However, much of the visual diversity in the world can be captured through passively collected datasets of images or videos. In our method, which we refer to as GPLAC (Generalized Policy Learning with Attentional Classifier), we use both interaction data and weakly labeled image data to augment the generalization capacity of sensorimotor policies. Our method combines multitask learning on action selection and an auxiliary binary classification objective, together with a convolutional neural network architecture that uses an attentional mechanism to avoid distractors. We show that pairing interaction data from just a single environment with a diverse dataset of weakly labeled data results in greatly improved generalization to unseen environments, and show that this generalization depends on both the auxiliary objective and the attentional architecture that we propose. We demonstrate our results in both simulation and on a real robotic manipulator, and demonstrate substantial improvement over standard convolutional architectures and domain adaptation methods.

cs.LG↗
arXiv2026

Humanoid Everyday: A Comprehensive Robotic Dataset for Open-World Humanoid Manipulation

From loco-motion to dextrous manipulation, humanoid robots have made remarkable strides in demonstrating complex full-body capabilities. However, the majority of current robot learning datasets and benchmarks mainly focus on stationary robot arms, and the few existing humanoid datasets are either confined to fixed environments or limited in task diversity, often lacking human-humanoid interaction and lower-body locomotion. Moreover, there are a few standardized evaluation platforms for benchmarking learning-based policies on humanoid data. In this work, we present Humanoid Everyday, a large-scale and diverse humanoid manipulation dataset characterized by extensive task variety involving dextrous object manipulation, human-humanoid interaction, locomotion-integrated actions, and more. Leveraging a highly efficient human-supervised teleoperation pipeline, Humanoid Everyday aggregates high-quality multimodal sensory data, including RGB, depth, LiDAR, and tactile inputs, together with natural language annotations, comprising 10.3k trajectories and over 3 million frames of data across 260 tasks across 7 broad categories. In addition, we conduct an analysis of representative policy learning methods on our dataset, providing insights into their strengths and limitations across different task categories. For standardized evaluation, we introduce a cloud-based evaluation platform that allows researchers to seamlessly deploy their policies in our controlled setting and receive performance feedback. By releasing Humanoid Everyday along with our policy learning analysis and a standardized cloud-based evaluation platform, we intend to advance research in general-purpose humanoid manipulation and lay the groundwork for more capable and embodied robotic agents in real-world scenarios. Our dataset, data collection code, and cloud evaluation website are made publicly available on our project website.

cs.RO↗
arXiv2026

EgoLive: A Large-Scale Egocentric Dataset from Real-World Human Tasks

The advancement of robot learning is currently hindered by the scarcity of large-scale, high-quality datasets. While established data collection methods such as teleoperation and universal manipulation interfaces dominate current datasets, they suffer from inherent limitations in scalability and real-world deployability. Human egocentric video collection, by contrast, has emerged as a promising approach to enable scalable, natural and in-the-wild data collection. As such, we present EgoLive, a large-scale, high-quality egocentric dataset designed explicitly for robot manipulation learning. EgoLive establishes three distinctive technical advantages over existing egocentric datasets: first, it represents the largest open-source annotated egocentric dataset focused on real-world task-oriented human routines to date; second, it delivers leading data quality via a customized head-mounted capture device and comprehensive high-precision multi-modal annotations; third, all data is collected exclusively in unconstrained real-world scenarios and encompasses vertical field human working data, including home service, retail, and other practical work scenarios, providing superior diversity and ecological validity. With the introduction of EgoLive, we aim to provide the research community with a scalable, high-quality dataset that accelerates breakthroughs in generalizable robotic models and facilitates the real-world deployment of robot systems.

cs.RO↗
arXiv2026

Continual Model-Based Reinforcement Learning with Hypernetworks

Effective planning in model-based reinforcement learning (MBRL) and model-predictive control (MPC) relies on the accuracy of the learned dynamics model. In many instances of MBRL and MPC, this model is assumed to be stationary and is periodically re-trained from scratch on state transition experience collected from the beginning of environment interactions. This implies that the time required to train the dynamics model - and the pause required between plan executions - grows linearly with the size of the collected experience. We argue that this is too slow for lifelong robot learning and propose HyperCRL, a method that continually learns the encountered dynamics in a sequence of tasks using task-conditional hypernetworks. Our method has three main attributes: first, it includes dynamics learning sessions that do not revisit training data from previous tasks, so it only needs to store the most recent fixed-size portion of the state transition experience; second, it uses fixed-capacity hypernetworks to represent non-stationary and task-aware dynamics; third, it outperforms existing continual learning alternatives that rely on fixed-capacity networks, and does competitively with baselines that remember an ever increasing coreset of past experience. We show that HyperCRL is effective in continual model-based reinforcement learning in robot locomotion and manipulation scenarios, such as tasks involving pushing and door opening. Our project website with videos is at this link https://rvl.cs.toronto.edu/blog/hypercrl

cs.LG↗
arXiv2026

EgoEngine: From Egocentric Human Videos to High-Fidelity Dexterous Robot Demonstrations

Dexterous manipulation is limited by the cost of collecting large-scale robot demonstrations. Egocentric human videos offer a scalable source of diverse manipulation behaviors, but directly using them for robot learning requires bridging two gaps: the visual gap between human and robot observations, and the action gap between human motion and robot-executable action. We propose EgoEngine, a scalable framework for transforming egocentric human manipulation videos into high-fidelity robot data. Given an egocentric RGB video, EgoEngine produces: (i) a high-fidelity robot observation video replacing human with robot while preserving scene context and temporal alignment, and (ii) a task-aligned, executable robot action trajectory under feasibility constraints. Experiments in simulation and on real robots show that EgoEngine enables scalable conversion of human videos into robot data and, to our knowledge, demonstrates the first zero-shot visuomotor dexterous policy learning from egocentric human videos without real-robot demonstrations. Project website: https://egoengine.github.io.

cs.RO↗
arXiv2020

IRIS: Implicit Reinforcement without Interaction at Scale for Learning Control from Offline Robot Manipulation Data

Learning from offline task demonstrations is a problem of great interest in robotics. For simple short-horizon manipulation tasks with modest variation in task instances, offline learning from a small set of demonstrations can produce controllers that successfully solve the task. However, leveraging a fixed batch of data can be problematic for larger datasets and longer-horizon tasks with greater variations. The data can exhibit substantial diversity and consist of suboptimal solution approaches. In this paper, we propose Implicit Reinforcement without Interaction at Scale (IRIS), a novel framework for learning from large-scale demonstration datasets. IRIS factorizes the control problem into a goal-conditioned low-level controller that imitates short demonstration sequences and a high-level goal selection mechanism that sets goals for the low-level and selectively combines parts of suboptimal solutions leading to more successful task completions. We evaluate IRIS across three datasets, including the RoboTurk Cans dataset collected by humans via crowdsourcing, and show that performant policies can be learned from purely offline learning. Additional results at https://sites.google.com/stanford.edu/iris/ .

cs.RO↗
arXiv2022

Toward Force Estimation in Robot-Assisted Surgery using Deep Learning with Vision and Robot State

Knowledge of interaction forces during teleoperated robot-assisted surgery could be used to enable force feedback to human operators and evaluate tissue handling skill. However, direct force sensing at the end-effector is challenging because it requires biocompatible, sterilizable, and cost-effective sensors. Vision-based deep learning using convolutional neural networks is a promising approach for providing useful force estimates, though questions remain about generalization to new scenarios and real-time inference. We present a force estimation neural network that uses RGB images and robot state as inputs. Using a self-collected dataset, we compared the network to variants that included only a single input type, and evaluated how they generalized to new viewpoints, workspace positions, materials, and tools. We found that vision-based networks were sensitive to shifts in viewpoints, while state-only networks were robust to changes in workspace. The network with both state and vision inputs had the highest accuracy for an unseen tool, and was moderately robust to changes in viewpoints. Through feature removal studies, we found that using only position features produced better accuracy than using only force features as input. The network with both state and vision inputs outperformed a physics-based baseline model in accuracy. It showed comparable accuracy but faster computation times than a baseline recurrent neural network, making it better suited for real-time applications.

cs.RO↗
arXiv2024

Surgical Gym: A high-performance GPU-based platform for reinforcement learning with surgical robots

Recent advances in robot-assisted surgery have resulted in progressively more precise, efficient, and minimally invasive procedures, sparking a new era of robotic surgical intervention. This enables doctors, in collaborative interaction with robots, to perform traditional or minimally invasive surgeries with improved outcomes through smaller incisions. Recent efforts are working toward making robotic surgery more autonomous which has the potential to reduce variability of surgical outcomes and reduce complication rates. Deep reinforcement learning methodologies offer scalable solutions for surgical automation, but their effectiveness relies on extensive data acquisition due to the absence of prior knowledge in successfully accomplishing tasks. Due to the intensive nature of simulated data collection, previous works have focused on making existing algorithms more efficient. In this work, we focus on making the simulator more efficient, making training data much more accessible than previously possible. We introduce Surgical Gym, an open-source high performance platform for surgical robot learning where both the physics simulation and reinforcement learning occur directly on the GPU. We demonstrate between 100-5000x faster training times compared with previous surgical learning platforms. The code is available at: https://github.com/SamuelSchmidgall/SurgicalGym.

cs.RO↗
arXiv2025

AAM-SEALS: Developing Aerial-Aquatic Manipulators in SEa, Air, and Land Simulator

Current mobile manipulators and high-fidelity simulators lack the ability to seamlessly operate and simulate across integrated environments spanning sea, air, and land. To address this gap, we introduce Aerial-Aquatic Manipulators (AAMs) in SEa, Air, and Land Simulator (SEALS), a comprehensive and photorealistic simulator designed for AAMs to operate and learn in these diverse environments. The development of AAM-SEALS tackles several significant challenges, including the creation of integrated controllers for flying, swimming, and manipulation, and the high-fidelity simulation of aerial dynamics and hydrodynamics leveraging particle-based hydrodynamics. Our evaluation demonstrates smooth operation and photorealistic transitions across air, water, and their interfaces. We quantitatively validate the fidelity of particle-based hydrodynamics by comparing position-tracking errors across real-world and simulated systems. AAM-SEALS benefits a broad range of robotics communities, including robot learning, aerial robotics, underwater robotics, mobile manipulation, and robotic simulators. We will open-source our code and data to foster the advancement of research in these fields. The overview video is available at https://youtu.be/MbqIIrYvR78. Visit our project website at https://aam-seals.umd.edu for more details.

cs.RO↗
arXiv2019

Learning Actions from Human Demonstration Video for Robotic Manipulation

Learning actions from human demonstration is an emerging trend for designing intelligent robotic systems, which can be referred as video to command. The performance of such approach highly relies on the quality of video captioning. However, the general video captioning methods focus more on the understanding of the full frame, lacking of consideration on the specific object of interests in robotic manipulations. We propose a novel deep model to learn actions from human demonstration video for robotic manipulation. It consists of two deep networks, grasp detection network (GNet) and video captioning network (CNet). GNet performs two functions: providing grasp solutions and extracting the local features for the object of interests in robotic manipulation. CNet outputs the captioning results by fusing the features of both full frames and local objects. Experimental results on UR5 robotic arm show that our method could produce more accurate command from video demonstration than state-of-the-art work, thereby leading to more robust grasping performance.

cs.CV↗
arXiv2022

Inducing Structure in Reward Learning by Learning Features

Reward learning enables robots to learn adaptable behaviors from human input. Traditional methods model the reward as a linear function of hand-crafted features, but that requires specifying all the relevant features a priori, which is impossible for real-world tasks. To get around this issue, recent deep Inverse Reinforcement Learning (IRL) methods learn rewards directly from the raw state but this is challenging because the robot has to implicitly learn the features that are important and how to combine them, simultaneously. Instead, we propose a divide and conquer approach: focus human input specifically on learning the features separately, and only then learn how to combine them into a reward. We introduce a novel type of human input for teaching features and an algorithm that utilizes it to learn complex features from the raw state space. The robot can then learn how to combine them into a reward using demonstrations, corrections, or other reward learning frameworks. We demonstrate our method in settings where all features have to be learned from scratch, as well as where some of the features are known. By first focusing human input specifically on the feature(s), our method decreases sample complexity and improves generalization of the learned reward over a deepIRL baseline. We show this in experiments with a physical 7DOF robot manipulator, as well as in a user study conducted in a simulated environment.

cs.RO↗
arXiv2025

A Flexible Field-Based Policy Learning Framework for Diverse Robotic Systems and Sensors

We present a cross robot visuomotor learning framework that integrates diffusion policy based control with 3D semantic scene representations from D3Fields to enable category level generalization in manipulation. Its modular design supports diverse robot camera configurations including UR5 arms with Microsoft Azure Kinect arrays and bimanual manipulators with Intel RealSense sensors through a low latency control stack and intuitive teleoperation. A unified configuration layer enables seamless switching between setups for flexible data collection training and evaluation. In a grasp and lift block task the framework achieved an 80 percent success rate after only 100 demonstration episodes demonstrating robust skill transfer between platforms and sensing modalities. This design paves the way for scalable real world studies in cross robotic generalization.

cs.RO↗
arXiv2026

Intrinsic Robot Rewarding: Reusing VLA Representations for Autonomous Evaluation and Policy Improvement

Vision-language-action (VLA) systems already bring together two valuable resources for robot learning: rich visual representations and demonstrations of successful task execution. Intrinsic Robot Rewarding (IRR) proposes to use these resources for a second, complementary purpose: evaluating the robot's own outcomes and providing feedback for policy improvement. Successful demonstration endpoints define task-specific references, and the policy's frozen visual encoder provides the feature space in which new outcomes are assessed. The core reward mechanism adds a reference bank and a scoring operation to the existing pipeline, without requiring a separate learned evaluator or an additional perception backbone. Our position is that this reuse offers a promising route to lower integration effort, efficient reward computation, and reduced recurring human outcome scoring. Building on established research in visual rewards and learning from experience, IRR brings these ideas into the robot's existing perception and demonstration pipeline. An operational COMAU Racer 3 demonstrator is available at technology readiness level 4 (TRL 4). This laboratory foundation supports the next research step: connecting internal outcome evaluation to physical policy improvement. We present the reward formulation, central research questions, and an evaluation methodology linking reward reliability to task success and supervision effort. The intended contribution is a reusable approach to learn and improve from the data and experience already available in industrial robot systems.

cs.RO↗
arXiv2025

Deep Incremental Model Informed Reinforcement Learning for Continuous Robotic Control

Model-based reinforcement learning attempts to use an available or learned model to improve the data efficiency of reinforcement learning. This work proposes a one-step lookback approach that jointly learns the deep incremental model and the policy to realize the sample-efficient continuous robotic control, wherein the control-theoretical knowledge is utilized to decrease the model learning difficulty and facilitate efficient training. Specifically, we use one-step backward data to facilitate the deep incremental model, an alternative structured representation of the robotic evolution model, that accurately predicts the robotic movement but with low sample complexity. This is because the formulated deep incremental model degrades the model learning difficulty into a parametric matrix learning problem, which is especially favourable to high-dimensional robotic applications. The imagined data from the learned deep incremental model is used to supplement training data to enhance the sample efficiency. Comparative numerical simulations on benchmark continuous robotics control problems are conducted to validate the efficiency of our proposed one-step lookback approach.

cs.RO↗
arXiv2017

Autonomous development and learning in artificial intelligence and robotics: Scaling up deep learning to human--like learning

Autonomous lifelong development and learning is a fundamental capability of humans, differentiating them from current deep learning systems. However, other branches of artificial intelligence have designed crucial ingredients towards autonomous learning: curiosity and intrinsic motivation, social learning and natural interaction with peers, and embodiment. These mechanisms guide exploration and autonomous choice of goals, and integrating them with deep learning opens stimulating perspectives. Deep learning (DL) approaches made great advances in artificial intelligence, but are still far away from human learning. As argued convincingly by Lake et al., differences include human capabilities to learn causal models of the world from very little data, leveraging compositional representations and priors like intuitive physics and psychology. However, there are other fundamental differences between current DL systems and human learning, as well as technical ingredients to fill this gap, that are either superficially, or not adequately, discussed by Lake et al. These fundamental mechanisms relate to autonomous development and learning. They are bound to play a central role in artificial intelligence in the future. Current DL systems require engineers to manually specify a task-specific objective function for every new task, and learn through off-line processing of large training databases. On the contrary, humans learn autonomously open-ended repertoires of skills, deciding for themselves which goals to pursue or value, and which skills to explore, driven by intrinsic motivation/curiosity and social learning through natural interaction with peers. Such learning processes are incremental, online, and progressive. Human child development involves a progressive increase of complexity in a curriculum of learning where skills are explored, acquired, and built on each other, through particular ordering and timing. Finally, human learning happens in the physical world, and through bodily and physical experimentation, under severe constraints on energy, time, and computational resources. In the two last decades, the field of Developmental and Cognitive Robotics (Cangelosi and Schlesinger, 2015, Asada et al., 2009), in strong interaction with developmental psychology and neuroscience, has achieved significant advances in computational

cs.AI↗