CoRL
Collection
Accepted papers for CoRL (Conference on Robot Learning), one dataset per year. • 10 items • Updated
paper_id stringlengths 10 10 | title stringlengths 21 134 | paper_url stringlengths 42 42 | authors listlengths 2 35 | type stringclasses 2
values | primary_area stringclasses 0
values | abstract large_stringlengths 709 2.19k | keywords listlengths 1 10 | TL;DR large_stringlengths 45 250 ⌀ | submission_number int64 5 1.15k | arxiv_id stringlengths 10 10 ⌀ | arxiv_id_source stringclasses 2
values |
|---|---|---|---|---|---|---|---|---|---|---|---|
9uKL9FJBiz | Pseudo-Simulation for Autonomous Driving | https://openreview.net/forum?id=9uKL9FJBiz | [
"Wei Cao",
"Marcel Hallgarten",
"Tianyu Li",
"Daniel Dauner",
"Xunjiang Gu",
"Caojun Wang",
"Yakov Miron",
"Marco Aiello",
"Hongyang Li",
"Igor Gilitschenski",
"Boris Ivanovic",
"Marco Pavone",
"Andreas Geiger",
"Kashyap Chitta"
] | Poster | null | Existing evaluation paradigms for Autonomous Vehicles (AVs) face critical limitations. Real-world evaluation is often challenging due to safety concerns and a lack of reproducibility, whereas closed-loop simulation can face insufficient realism or high computational costs. Open-loop evaluation, while being efficient an... | [
"Simulation",
"Benchmarking",
"Autonomous Driving"
] | We introduce Pseudo-Simulation, a novel AV evaluation methodology that augments real data with synthetic observations near the planned trajectory. It achieves strong correlation with closed-loop simulation while being much faster and easier to scale. | 5 | 2506.04218 | title_snapshot |
1otaE496Vm | CaRL: Learning Scalable Planning Policies with Simple Rewards | https://openreview.net/forum?id=1otaE496Vm | [
"Bernhard Jaeger",
"Daniel Dauner",
"Jens Beißwenger",
"Simon Gerstenecker",
"Kashyap Chitta",
"Andreas Geiger"
] | Poster | null | We investigate reinforcement learning (RL) for privileged planning in autonomous driving.
State-of-the-art approaches for this task are rule-based, but these methods do not scale to the long tail.
RL, on the other hand, is scalable and does not suffer from compounding errors like imitation learning.
Contemporary RL ap... | [
"Autonomous Driving",
"Reinforcement Learning",
"Planning"
] | null | 6 | 2504.17838 | title_snapshot |
7iaYcss56y | ImMimic: Cross-Domain Imitation from Human Videos via Mapping and Interpolation | https://openreview.net/forum?id=7iaYcss56y | [
"Yangcen Liu",
"Woo Chul Shin",
"Yunhai Han",
"Zhenyang Chen",
"Harish Ravichandar",
"Danfei Xu"
] | Oral | null | Learning robot manipulation from abundant human videos offers a scalable alternative to costly robot-specific data collection. However, domain gaps across visual, morphological, and physical aspects hinder direct imitation. To effectively bridge the domain gap, we propose ImMimic, an embodiment-agnostic co-training fra... | [
"Learning from Human",
"Imitation learning",
"Dexterous Manipulation"
] | null | 26 | 2509.10952 | title_snapshot |
bi8o9p6h2R | Improving Efficiency of Sampling-based Motion Planning via Message-Passing Monte Carlo | https://openreview.net/forum?id=bi8o9p6h2R | [
"Makram Chahine",
"T. Konstantin Rusch",
"Zach J Patterson",
"Daniela Rus"
] | Poster | null | Sampling-based motion planning methods, while effective in high-dimensional spaces, often suffer from inefficiencies due to irregular sampling distributions, leading to suboptimal exploration of the configuration space. In this paper, we propose an approach that enhances the efficiency of these methods by utilizing low... | [
"Graph Neural Networks",
"Discrepancy Theory",
"Motion Planning"
] | This paper introduces a new sampling method for robot motion planning, using graph neural networks to generate more uniform sample distributions, which leads to improved planning efficiency in complex environments. | 31 | 2410.03909 | title_snapshot |
4Po2mqLjrQ | Motion Blender Gaussian Splatting for Dynamic Reconstruction | https://openreview.net/forum?id=4Po2mqLjrQ | [
"Xinyu Zhang",
"Haonan Chang",
"Yuhan Liu",
"Abdeslam Boularias"
] | Poster | null | Gaussian splatting has emerged as a powerful tool for high-fidelity reconstruction of dynamic scenes. However, existing methods primarily rely on implicit motion representations, such as encoding motions into neural networks or per-Gaussian parameters, which makes it difficult to further manipulate the reconstructed m... | [
"Dynamic Reconstruction",
"Gaussian Splatting"
] | A framework that uses graphs as an explicit and sparse motion representation for Gaussian splatting to reconstruct dynamic world, with applications in animation, robot data synthesis, and planning. | 32 | 2503.09040 | title_judge |
SebHZk78aS | GraspMolmo: Generalizable Task-Oriented Grasping via Large-Scale Synthetic Data Generation | https://openreview.net/forum?id=SebHZk78aS | [
"Abhay Deshpande",
"Yuquan Deng",
"Jordi Salvador",
"Arijit Ray",
"Winson Han",
"Jiafei Duan",
"Rose Hendrix",
"Yuke Zhu",
"Ranjay Krishna"
] | Poster | null | We present GraspMolmo, a generalizable open-vocabulary task-oriented grasping (TOG) model. GraspMolmo predicts semantically appropriate, stable grasps conditioned on a natural language instruction and a single RGB-D frame. For instance, given "pour me some tea", GraspMolmo selects a grasp on a teapot handle rather than... | [
"task-oriented grasping",
"functional grasping",
"robotic manipulation"
] | We present GraspMolmo, an open-vocabulary task-oriented grasping model that predicts semantically appropriate grasps from natural language and RGB-D input, using large-scale synthetic data to achieve SOTA performance in real-world manipulation tasks. | 34 | 2505.13441 | title_snapshot |
B6knAJsB9P | COLLAGE: Adaptive Fusion-based Retrieval for Augmented Policy Learning | https://openreview.net/forum?id=B6knAJsB9P | [
"Sateesh Kumar",
"Shivin Dass",
"Georgios Pavlakos",
"Roberto Martín-Martín"
] | Poster | null | In this work, we study the problem of data retrieval for few-shot imitation learning: select data from a large dataset to train a performant policy for a specific task, given only a few target demonstrations. Prior methods retrieve data using a single-feature distance heuristic, assuming that the best demonstrations ar... | [
"Data Retrieval",
"Few-shot Learning",
"Imitation Learning"
] | COLLAGE improves augmented policy learning by adaptively reweighting data retrieved using multiple feature cues based on task relevance. | 36 | 2508.01131 | title_snapshot |
LbeMepi89R | RoboChemist: Long-Horizon and Safety-Compliant Robotic Chemical Experimentation | https://openreview.net/forum?id=LbeMepi89R | [
"Zongzheng Zhang",
"Chenghao Yue",
"Haobo Xu",
"Minwen Liao",
"Xianglin Qi",
"Huan-ang Gao",
"Ziwei Wang",
"Hao Zhao"
] | Poster | null | Robotic chemists promise to both liberate human experts from repetitive tasks and accelerate scientific discovery, yet remain in their infancy. Chemical experiments involve long-horizon procedures over hazardous and deformable substances, where success requires not only task completion but also strict compliance with e... | [
"Robotic Chemistry",
"VLA Models",
"Visual Prompting"
] | RoboChemist combines VLMs and VLA models in a close-loop framework to safely and reliably automate complex chemical experiments. | 38 | 2509.08820 | title_snapshot |
HAmi1X11BO | Elucidating the Design Space of Torque-aware Vision-Language-Action Models | https://openreview.net/forum?id=HAmi1X11BO | [
"Zongzheng Zhang",
"Haobo Xu",
"Zhuo Yang",
"Chenghao Yue",
"Zehao Lin",
"Huan-ang Gao",
"Ziwei Wang",
"Hao Zhao"
] | Poster | null | Many robotic manipulation tasks require sensing and responding to force signals such as torque to assess whether the task has been successfully completed and to enable closed-loop control. However, current Vision-Language-Action (VLA) models lack the ability to integrate such subtle physical feedback. In this work, we ... | [
"Torque Integration",
"VLA Models"
] | Embedding torque history as a single decoder token and jointly predicting future torque alongside actions in pretrained vision-language-action models significantly boosts performance on contact-rich manipulation tasks. | 39 | 2509.07962 | title_judge |
EBcb4LemZZ | Co-Design of Soft Gripper with Neural Physics | https://openreview.net/forum?id=EBcb4LemZZ | [
"Sha Yi",
"Xueqian Bai",
"Adabhav Singh",
"Jianglong Ye",
"Michael T. Tolley",
"Xiaolong Wang"
] | Poster | null | For robot manipulation, both the controller and end-effector design are crucial. Compared with rigid grippers, soft grippers are more generalizable by deforming to different geometries, but designing such a gripper and finding its grasp pose remains challenging. In this paper, we propose a co-design framework that gene... | [
"soft robot",
"manipulation",
"design optimization"
] | null | 45 | 2505.20404 | title_snapshot |
VqmAvBkFhw | LocoFormer: Generalist Locomotion via Long-context Adaptation | https://openreview.net/forum?id=VqmAvBkFhw | [
"Min Liu",
"Deepak Pathak",
"Ananye Agarwal"
] | Oral | null | Humans and animals exhibit flexible locomotion strategies, such as learning to walk within minutes, and efficient adaptation to changes in morphology. In contrast, modern locomotion controllers are manually tuned for specific embodiments. In this paper, we present LocoFormer, a generalist policy that can control previo... | [
"Cross-Embodied Learning",
"Legged Locomotion",
"Online Adaptation"
] | LocoFormer is a generalist locomotion policy trained via large-scale RL and extended context memory, enabling it to robustly adapt to diverse, unseen legged and wheeled robots without prior knowledge. | 50 | 2509.23745 | title_snapshot |
jedBaI1fgU | BranchOut: Capturing Realistic Multimodality in Autonomous Driving Decisions | https://openreview.net/forum?id=jedBaI1fgU | [
"Hee Jae Kim",
"Zekai Yin",
"Lei Lai",
"Jason Lee",
"Eshed Ohn-Bar"
] | Poster | null | Modeling the nuanced, multimodal nature of human driving remains a core challenge for autonomous systems, as existing methods often fail to capture the diversity of plausible behaviors in complex real-world scenarios. In this work, we introduce a novel benchmark and end-to-end planner for modeling realistic multimodali... | [
"Autonomous Driving",
"Human-in-the-Loop Simulation",
"Multi-modal Planning and Evaluation"
] | null | 55 | null | null |
ZqBXnR6ppz | Generalist Robot Manipulation beyond Action Labeled Data | https://openreview.net/forum?id=ZqBXnR6ppz | [
"Alexander Spiridonov",
"Jan-Nico Zaech",
"Nikolay Nikolov",
"Luc Van Gool",
"Danda Pani Paudel"
] | Poster | null | Recent advances in generalist robot manipulation leverage pre-trained Vision–Language Models (VLMs) and large-scale robot demonstrations to tackle diverse tasks in a zero-shot manner. A key challenge remains: scaling high-quality, action-labeled robot demonstration data, which existing methods rely on for robustness an... | [
"Vision-Language-Action Models",
"Learning from Videos"
] | null | 60 | 2509.19958 | title_snapshot |
hh9afiQMb2 | Real-Time Out-of-Distribution Failure Prevention via Multi-Modal Reasoning | https://openreview.net/forum?id=hh9afiQMb2 | [
"Milan Ganai",
"Rohan Sinha",
"Christopher Agia",
"Daniel Morton",
"Luigi Di Lillo",
"Marco Pavone"
] | Oral | null | Foundation models can provide robust high-level reasoning on appropriate safety interventions in hazardous scenarios beyond a robot's training data, i.e. out-of-distribution (OOD) failures. However, due to the high inference latency of Large Vision and Language Models, current methods rely on manually defined intervent... | [
"Multi-modal Reasoning in Robotics",
"OOD Safety",
"Fallback Synthesis"
] | An algorithm that prevents OOD Failures in open-world environments by rapidly generating semantically safe fallback plans with multi-modal reasoning. | 61 | 2505.10547 | title_snapshot |
Vo1tL9dhpk | Uncertainty-aware Accurate Elevation Modeling for Off-road Navigation via Neural Processes | https://openreview.net/forum?id=Vo1tL9dhpk | [
"Sanghun Jung",
"Daehoon Gwak",
"Byron Boots",
"James Hays"
] | Poster | null | Terrain elevation modeling for off-road navigation aims to accurately estimate changes in terrain geometry in real-time and quantify the corresponding uncertainties. Having precise estimations and uncertainties plays a crucial role in planning and control algorithms to explore safe and reliable maneuver strategies. How... | [
"Robot perception",
"Terrain modeling",
"Neural processes"
] | We propose semantic-conditioned Neural Processes with ball-query attention for sharp elevation estimates along the associated uncertainties. | 65 | 2508.03890 | title_snapshot |
KUSYJIlKor | Omni-Perception: Omnidirectional Collision Avoidance of Legged Robots in Dynamic Environments | https://openreview.net/forum?id=KUSYJIlKor | [
"Zifan Wang",
"Teli Ma",
"Yufei Jia",
"Xun Yang",
"Jiaming Zhou",
"Wenlong OUYANG",
"Qiang Zhang",
"Junwei Liang"
] | Oral | null | Agile locomotion in complex 3D environments requires robust spatial awareness to safely avoid diverse obstacles such as aerial clutter, uneven terrain, and dynamic agents. Depth-based perception approaches often struggle with sensor noise, lighting variability, computational overhead from intermediate representations (... | [
"Locomotion",
"Reinforcement Learning",
"Legged robot"
] | null | 66 | 2505.19214 | title_judge |
dT45OMevL5 | 3DS-VLA: A 3D Spatial-Aware Vision Language Action Model for Robust Multi-Task Manipulation | https://openreview.net/forum?id=dT45OMevL5 | [
"Xiaoqi Li",
"Liang Heng",
"Jiaming Liu",
"Yan Shen",
"Chenyang Gu",
"Zhuoyang Liu",
"Hao Chen",
"Nuowei Han",
"Renrui Zhang",
"Hao Tang",
"Shanghang Zhang",
"Hao Dong"
] | Poster | null | Recently, 2D vision-language-action (VLA) models have made significant strides in multi-task manipulation. However, these models struggle to reason about 3D spatial relationships from 2D image inputs. Although an increasing number of 3D approaches explicitly integrate 3D information, they encounter challenges such as l... | [
"Vision-Language-Action; Robotic Manipulation; Imitation Learning"
] | null | 71 | null | null |
1D6XYy6ofW | LaVA-Man: Learning Visual Action Representations for Robot Manipulation | https://openreview.net/forum?id=1D6XYy6ofW | [
"Chaoran Zhu",
"Hengyi Wang",
"Yik Lung Pang",
"Changjae Oh"
] | Poster | null | Visual-textual understanding is essential for language-guided robot manipulation. Recent works leverage pre-trained vision-language models to measure the similarity between encoded visual observations and textual instructions, and then train a model to map this similarity to robot actions. However, this two-step approa... | [
"Robot manipulation",
"self-supervised representation learning"
] | Self-supervised pre-training method for langauge-conditioned robot manipulation | 72 | 2508.19391 | title_snapshot |
AE299O0tph | See, Point, Fly: A Learning-Free VLM Framework for Universal Unmanned Aerial Navigation | https://openreview.net/forum?id=AE299O0tph | [
"Chih Yao Hu",
"Yang-Sen Lin",
"Yuna Lee",
"Chih-Hai Su",
"Jie-Ying Lee",
"Shr-Ruei Tsai",
"Chin-Yang Lin",
"Kuan-Wen Chen",
"Tsung-Wei Ke",
"Yu-Lun Liu"
] | Poster | null | We present See, Point, Fly (SPF), a training-free aerial vision-and-language navigation (AVLN) framework built atop vision-language models (VLMs). SPF is capable of navigating to any goal based on any type of free-form instructions in any kind of environment. In contrast to existing VLM-based approaches that treat acti... | [
"Vision-Language Models",
"Zero-shot UAV Navigation",
"2D-to-3D",
"Waypoint Prompting"
] | null | 74 | 2509.22653 | title_snapshot |
oRwcxFuN25 | Robot Trains Robot: Automatic Real-World Policy Adaptation and Learning for Humanoids | https://openreview.net/forum?id=oRwcxFuN25 | [
"Kaizhe Hu",
"Haochen Shi",
"Yao He",
"Weizhuo Wang",
"Karen Liu",
"Shuran Song"
] | Poster | null | Simulation-based reinforcement learning (RL) has significantly advanced humanoid locomotion tasks, yet direct real-world RL from scratch or starting from pretrained policies remains rare, limiting the full potential of humanoid robots. Real-world training, despite being crucial for overcoming the sim-to-real gap, faces... | [
"Humanoid Robots",
"Sim-to-Real Adaptation",
"Real-World RL"
] | Robot-Trains-Robot uses a robot arm teacher to actively train a humanoid student, aiming for practical and highly efficient real-world humanoid adaptation and learning. | 76 | 2508.12252 | title_snapshot |
FwLMCbs47K | CDP: Towards Robust Autoregressive Visuomotor Policy Learning via Causal Diffusion | https://openreview.net/forum?id=FwLMCbs47K | [
"Jiahua Ma",
"Yiran Qin",
"Yixiong Li",
"Xuanqi Liao",
"Yulan Guo",
"Ruimao Zhang"
] | Poster | null | Diffusion Policy (DP) enables robots to learn complex behaviors by imitating expert demonstrations through action diffusion. However, in practical applications, hardware limitations often degrade data quality, while real-time constraints restrict model inference to instantaneous state and scene observations. These limi... | [
"Autoregressive",
"Causal Diffusion Policy"
] | null | 77 | 2506.14769 | title_snapshot |
DKXx17oaUf | OPAL: Visibility-aware LiDAR-to-OpenStreetMap Place Recognition via Adaptive Radial Fusion | https://openreview.net/forum?id=DKXx17oaUf | [
"Shuhao Kang",
"Youqi Liao",
"Yan Xia",
"Olaf Wysocki",
"Boris Jutzi",
"Daniel Cremers"
] | Poster | null | LiDAR place recognition is a critical capability for autonomous navigation and cross-modal localization in large-scale outdoor environments. Existing approaches predominantly depend on pre-built 3D dense maps or aerial imagery, which impose significant storage overhead and lack real-time adaptability. In this paper, we... | [
"Place Recognition",
"OpenStreetMap",
"Point Cloud"
] | A novel learning-based method for point cloud-to-OpenStreetMap place recognition | 79 | 2504.19258 | title_snapshot |
b86nyIOJWq | exUMI: Extensible Robot Teaching System with Action-aware Task-agnostic Tactile Representation | https://openreview.net/forum?id=b86nyIOJWq | [
"Yue Xu",
"Litao Wei",
"Pengyu An",
"Qingyu Zhang",
"Yong-Lu Li"
] | Poster | null | Tactile-aware robot learning faces critical challenges in data collection and representation due to data scarcity and sparsity, and the absence of force feedback in existing systems. To address these limitations, we introduce a tactile robot learning system with both hardware and algorithm innovations. We present exUMI... | [
"Tactile Sensing",
"Robot Data Collection System",
"Imitation Learning"
] | null | 85 | 2509.14688 | title_snapshot |
htgNQHa6Ta | TWIST: Teleoperated Whole-Body Imitation System | https://openreview.net/forum?id=htgNQHa6Ta | [
"Yanjie Ze",
"Zixuan Chen",
"Joao Pedro Araujo",
"Zi-ang Cao",
"Xue Bin Peng",
"Jiajun Wu",
"Karen Liu"
] | Poster | null | Teleoperating humanoid robots in a whole-body manner marks a fundamental step toward developing general-purpose robotic intelligence, with human motion providing an ideal interface for controlling all degrees of freedom. Yet, most current humanoid teleoperation systems fall short of enabling coordinated whole-body beha... | [
"humanoid robots",
"whole-body teleoperation",
"learning-based control"
] | A capable system for whole-body humanoid teleoperation | 88 | 2505.02833 | title_snapshot |
GH9kURIRlx | ToddlerBot: Open-Source ML-Compatible Humanoid Platform for Loco-Manipulation | https://openreview.net/forum?id=GH9kURIRlx | [
"Haochen Shi",
"Weizhuo Wang",
"Shuran Song",
"Karen Liu"
] | Poster | null | Learning-based robotics research driven by data demands a new approach to robot hardware design—one that serves as both a platform for policy execution and a tool for embodied data collection. We introduce ToddlerBot, a low-cost, open-source humanoid robot platform designed for robotics and AI research. ToddlerBot enab... | [
"Humanoid Robots",
"Mechanisms & Design",
"Robot Modeling & Simulation"
] | ToddlerBot is a low-cost, open-source humanoid robot platform designed for scalable policy learning and research in robotics and AI. | 92 | 2502.00893 | title_snapshot |
Tx54fkQ3Cq | Humanoid Policy ~ Human Policy | https://openreview.net/forum?id=Tx54fkQ3Cq | [
"Ri-Zhao Qiu",
"Shiqi Yang",
"Xuxin Cheng",
"Chaitanya Chawla",
"Jialong Li",
"Tairan He",
"Ge Yan",
"David J. Yoon",
"Ryan Hoque",
"Lars Paulsen",
"Ge Yang",
"Jian Zhang",
"Sha Yi",
"Guanya Shi",
"Xiaolong Wang"
] | Poster | null | Training manipulation policies for humanoid robots with diverse data enhances their robustness and generalization across tasks and platforms. However, learning solely from robot demonstrations is labor-intensive, requiring expensive tele-operated data collection,n which is difficult to scale. This paper investigates a ... | [
"Robot Manipulation",
"Humanoid",
"Learning from Human"
] | Task-oriented egocentric human data can directly be used as a cross-embodiment learning source. | 94 | 2503.13441 | title_snapshot |
8DHSyMFLbB | Sim-to-Real Reinforcement Learning for Vision-Based Dexterous Manipulation on Humanoids | https://openreview.net/forum?id=8DHSyMFLbB | [
"Toru Lin",
"Kartik Sachdev",
"Linxi Fan",
"Jitendra Malik",
"Yuke Zhu"
] | Poster | null | Learning generalizable robot manipulation policies, especially for complex multi-fingered humanoids, remains a significant challenge. Existing approaches primarily rely on extensive data collection and imitation learning, which are expensive, labor-intensive, and difficult to scale. Sim-to-real reinforcement learning (... | [
"Humanoids",
"Vision-Based Dexterous Manipulation",
"Reinforcement Learning",
"Sim-to-Real"
] | null | 95 | 2502.20396 | title_snapshot |
SNvUSjVm6C | Robust Dexterous Grasping of General Objects | https://openreview.net/forum?id=SNvUSjVm6C | [
"Hui Zhang",
"Zijian wu",
"Linyi Huang",
"Sammy Christen",
"Jie Song"
] | Poster | null | The ability to robustly grasp a variety of objects is essential for dexterous robots. In this paper, we present a framework for zero-shot dynamic dexterous grasping using single-view visual inputs, designed to be resilient to various disturbances. Our approach utilizes a hand-centric object shape representation based o... | [
"Dexterous Grasping; Reinforcement Learning; Robot Manipulation"
] | An RL-based pipeline that achieves robust (94.6% success rate) dexterous grasping of various (500+) unseen objects using single-view observation, while performing adaptive motions to disturbances. | 96 | null | null |
b8RqTaDyb3 | From Tabula Rasa to Emergent Abilities: Discovering Robot Skills via Real-World Unsupervised Quality-Diversity | https://openreview.net/forum?id=b8RqTaDyb3 | [
"Luca Grillotti",
"Lisa Coiffard",
"Oscar Pang",
"Maxence Faldor",
"Antoine Cully"
] | Poster | null | Autonomous skill discovery aims to enable robots to acquire diverse be-haviors without explicit supervision. Learning such behaviors directly on physical hardware remains challenging due to safety and data efficiency constraints. Existing methods, including Quality-Diversity Actor-Critic (QDAC), require manually define... | [
"Quality-Diversity",
"Reinforcement Learning",
"Real-World Learning",
"Robotics"
] | We present URSA, a method enabling robots to autonomously discover and learn their abilities in the real world without extensive human guidance. | 100 | 2508.19172 | title_snapshot |
lRKGIidrYZ | Point Policy: Unifying Observations and Actions with Key Points for Robot Manipulation | https://openreview.net/forum?id=lRKGIidrYZ | [
"Siddhant Haldar",
"Lerrel Pinto"
] | Poster | null | Building robotic agents capable of operating across diverse environments and object types remains a significant challenge, often requiring extensive data collection. This is particularly restrictive in robotics, where each data point must be physically executed in the real world. Consequently, there is a critical need ... | [
"Robot Learning",
"Imitation Learning",
"Robot Perception",
"Sensing & Vision"
] | We present a method that unifies robot observations and actions with key points and enables learning generalizable robot policies exclusively from human videos. | 109 | 2502.20391 | title_snapshot |
1NBdplgILy | Vision in Action: Learning Active Perception from Human Demonstrations | https://openreview.net/forum?id=1NBdplgILy | [
"Haoyu Xiong",
"Xiaomeng Xu",
"Jimmy Wu",
"Yifan Hou",
"Jeannette Bohg",
"Shuran Song"
] | Poster | null | We present Vision in Action (ViA), an active perception system for bimanual robot manipulation. ViA learns task-relevant active perceptual strategies (e.g., searching, tracking, and focusing) directly from human demonstrations. On the hardware side, ViA employs a simple yet effective 6-DoF robotic neck to enable flexib... | [
"Active Perception",
"Bimanual Manipulation",
"Imitation Learning",
"Teleoperation Systems"
] | A robotic system for learning bimanual manipulation using active perception to address tasks with visual occlusions. | 120 | 2506.15666 | title_snapshot |
MpJTyAqA0t | Learning a Unified Policy for Position and Force Control in Legged Loco-Manipulation | https://openreview.net/forum?id=MpJTyAqA0t | [
"Peiyuan Zhi",
"Peiyang Li",
"Jianqin Yin",
"Baoxiong Jia",
"Siyuan Huang"
] | Oral | null | Robotic loco-manipulation tasks often involve contact-rich interactions with the environment, requiring the joint modeling of contact force and robot position. However, recent visuomotor policies often focus solely on position or force control, overlooking their integration. In this work, we propose a unified policy fo... | [
"Unified Force and Position Control",
"Force-aware Imitation Learning"
] | null | 123 | 2505.20829 | title_snapshot |
R5Y7Sr1DIe | LLM-Guided Probabilistic Program Induction for POMDP Model Estimation | https://openreview.net/forum?id=R5Y7Sr1DIe | [
"Aidan Curtis",
"Hao Tang",
"Thiago Veloso",
"Kevin Ellis",
"Joshua B. Tenenbaum",
"Tomás Lozano-Pérez",
"Leslie Pack Kaelbling"
] | Poster | null | Partially Observable Markov Decision Processes (POMDPs) model decision making under uncertainty. While there are many approaches to approximately solving POMDPs, we aim to address the problem of learning such models. In particular, we are interested in a subclass of POMDPs wherein the components of the model, including... | [
"POMDP",
"LLM",
"Inference",
"Probabilistic",
"Induction",
"Robotics",
"Uncertainty"
] | We use an LLM to infer POMDP models from data | 125 | 2505.02216 | title_snapshot |
JM2vDI6DlP | VLM-AD: End-to-End Autonomous Driving through Vision-Language Model Supervision | https://openreview.net/forum?id=JM2vDI6DlP | [
"Yi Xu",
"Yuxin Hu",
"Zaiwei Zhang",
"Gregory P. Meyer",
"Siva Karthik Mustikovela",
"Siddhartha Srinivasa",
"Eric M. Wolff",
"Xin Huang"
] | Poster | null | Human drivers rely on commonsense reasoning to navigate diverse and dynamic real-world scenarios. Existing end-to-end (E2E) autonomous driving (AD) models are typically optimized to mimic driving patterns observed in data, without capturing the underlying reasoning processes. This limitation constrains their ability t... | [
"End-to-End Autonomous Driving",
"Vision-Language Model"
] | null | 129 | 2412.14446 | title_snapshot |
Hu3NoPMAg4 | One Demo is Worth a Thousand Trajectories: Action-View Augmentation for Visuomotor Policies | https://openreview.net/forum?id=Hu3NoPMAg4 | [
"Chuer Pan",
"Litian Liang",
"Dominik Bauer",
"Eric Cousineau",
"Benjamin Burchfiel",
"Siyuan Feng",
"Shuran Song"
] | Poster | null | Visuomotor policies for manipulation have demonstrated remarkable potential in modeling complex robotic behaviors, yet minor alterations in the robot’s initial configuration and unseen obstacles easily lead to out-of-distribution observations. Without extensive data collection effort, these result in catastrophic execu... | [
"visuomotor policies",
"LfD",
"manipulation",
"3D vision",
"gaussian splat",
"motion planning"
] | null | 130 | 2606.19586 | title_snapshot |
XjjXLxfPou | ReWiND: Language-Guided Rewards Teach Robot Policies without New Demonstrations | https://openreview.net/forum?id=XjjXLxfPou | [
"Jiahui Zhang",
"Yusen Luo",
"Abrar Anwar",
"Sumedh Anand Sontakke",
"Joseph J Lim",
"Jesse Thomason",
"Erdem Biyik",
"Jesse Zhang"
] | Oral | null | We introduce ReWiND, a framework for learning robot manipulation tasks solely from language instructions without per-task demonstrations. Standard reinforcement learning (RL) and imitation learning methods require expert supervision through human-designed reward functions or demonstrations for every new task. In contra... | [
"Reinforcement Learning",
"Offline Reinforcement Learning",
"Reward Learning",
"Reward Modeling",
"Language"
] | ReWiND enables real-world robots to efficiently learn new manipulation tasks from language instructions alone by training a generalizable reward model trained on a small initial demonstration dataset. | 132 | 2505.10911 | title_snapshot |
CgGSFtjplI | Crossing the Human-Robot Embodiment Gap with Sim-to-Real RL using One Human Demonstration | https://openreview.net/forum?id=CgGSFtjplI | [
"Tyler Ga Wei Lum",
"Olivia Y. Lee",
"Karen Liu",
"Jeannette Bohg"
] | Poster | null | Teaching robots dexterous manipulation skills often requires collecting hundreds of demonstrations using wearables or teleoperation, a process that is challenging to scale. Videos of human-object interactions are easier to collect and scale, but leveraging them directly for robot learning is difficult due to the lack o... | [
"Dexterous Manipulation",
"Reinforcement Learning",
"Sim-to-Real"
] | Human2Sim2Robot trains dexterous manipulation policies from one human RGB-D video. We use object trajectories and pre-manipulation poses to guide RL in sim and thereby bridge the human-robot embodiment gap, achieving zero-shot sim-to-real transfer. | 134 | 2504.12609 | title_snapshot |
7wGYX11BJB | ParticleFormer: A 3D Point Cloud World Model for Multi-Object, Multi-Material Robotic Manipulation | https://openreview.net/forum?id=7wGYX11BJB | [
"Suning Huang",
"Qianzhong Chen",
"Xiaohan Zhang",
"Jiankai Sun",
"Mac Schwager"
] | Poster | null | 3D world models (i.e., learning-based 3D dynamics models) offer a promising approach to generalizable robotic manipulation by capturing the underlying physics of environment evolution conditioned on robot actions. However, existing 3D world models are primarily limited to single-material dynamics using a particle-based... | [
"Learning-based Dynamics Modeling",
"Model-based Planning"
] | ParticleFormer is a powerful dynamics modeling framework that excels at capturing complex multi-object, multi-material interactions, and enables effective downstream visuomotor control. | 135 | 2506.23126 | title_snapshot |
EgSDP6AOF1 | UniSkill: Imitating Human Videos via Cross-Embodiment Skill Representations | https://openreview.net/forum?id=EgSDP6AOF1 | [
"Hanjung Kim",
"Jaehyun Kang",
"Hyolim Kang",
"Meedeum Cho",
"Seon Joo Kim",
"Youngwoon Lee"
] | Poster | null | Mimicry is a fundamental learning mechanism in humans, enabling individuals to learn new tasks by observing and imitating experts. However, applying this ability to robots presents significant challenges due to the inherent differences between human and robot embodiments in both their visual appearance and physical cap... | [
"Robot Learning",
"Imitation Learning"
] | null | 138 | 2505.08787 | title_snapshot |
bU15EK0oqk | CASPER: Inferring Diverse Intents for Assistive Teleoperation with Vision Language Models | https://openreview.net/forum?id=bU15EK0oqk | [
"Huihan Liu",
"Rutav Shah",
"Shuijing Liu",
"Jack Pittenger",
"Mingyo Seo",
"Yuchen Cui",
"Yonatan Bisk",
"Roberto Martín-Martín",
"Yuke Zhu"
] | Poster | null | Assistive teleoperation, where control is shared between a human and a robot, enables efficient and intuitive human-robot collaboration in diverse and unstructured environments. A central challenge in real-world assistive teleoperation is for the robot to infer a wide range of human intentions from user control inputs ... | [
"Assistive Teleoperation",
"Mobile Manipulation",
"Vision Language Models"
] | We develop an assistive teleoperation system that harnesses commonsense knowledge embedded in pre-trained visual language models (VLMs) for real-time human intent inference and skill execution. | 143 | 2506.14727 | title_snapshot |
4eSv0QeYlz | Deep Reactive Policy: Learning Reactive Manipulator Motion Planning for Dynamic Environments | https://openreview.net/forum?id=4eSv0QeYlz | [
"Jiahui Yang",
"Jason Jingzhou Liu",
"Yulong Li",
"Youssef Khaky",
"Kenneth Shaw",
"Deepak Pathak"
] | Poster | null | Generating collision-free motion in dynamic, partially observable environments is a fundamental challenge for robotic manipulators. Classical motion planners can compute globally optimal trajectories but require full environment knowledge and are typically too slow for dynamic scenes. Neural motion policies offer a pro... | [
"Motion Planning",
"Visuo-Motor Policy",
"Reactive Control"
] | We introduce Deep Reactive Policy (DRP), a novel hierarchical framework capable of global planning while reacting to dynamic obstacles, directly operating on point cloud observations. | 145 | 2509.06953 | title_snapshot |
FCHl2ybdAm | Robot Learning from Any Images | https://openreview.net/forum?id=FCHl2ybdAm | [
"Siheng Zhao",
"Jiageng Mao",
"Wei Chow",
"Zeyu Shangguan",
"Tianheng Shi",
"Rong Xue",
"Yuxi Zheng",
"Yijia Weng",
"Yang You",
"Daniel Seita",
"Leonidas Guibas",
"Sergey Zakharov",
"Vitor Campagnolo Guizilini",
"Yue Wang"
] | Poster | null | We introduce RoLA, a framework that transforms any in‑the‑wild image into an interactive, physics‑enabled robotic environment. Unlike previous methods, RoLA operates directly on a single image without requiring additional hardware or digital assets. Our framework democratizes robotic data generation by producing massi... | [
"Robotic Manipulation",
"Real-to-Sim-to-Real",
"Robotic Data Collection"
] | We present a framework that transforms any single image into a robotic environment and enables scalable robotic data collection. | 163 | 2509.22970 | title_snapshot |
etSYDtRO0Z | ManiFlow: A General Robot Manipulation Policy via Consistency Flow Training | https://openreview.net/forum?id=etSYDtRO0Z | [
"Ge Yan",
"Jiyue Zhu",
"Yuquan Deng",
"Shiqi Yang",
"Ri-Zhao Qiu",
"Xuxin Cheng",
"Marius Memmel",
"Ranjay Krishna",
"Ankit Goyal",
"Xiaolong Wang",
"Dieter Fox"
] | Poster | null | Generative models based on flow matching offer significant potential for learning robot policies, particularly in generating high-dimensional, dexterous behaviors that are conditioned on diverse observations. In this work, we introduce ManiFlow, an advanced flow matching model specifically designed to support dexterous... | [
"Imitation Learning",
"Dexterous Manipulation"
] | null | 166 | 2509.01819 | title_snapshot |
o2w2iiMyEU | LaDi-WM: A Latent Diffusion-Based World Model for Predictive Manipulation | https://openreview.net/forum?id=o2w2iiMyEU | [
"Yuhang Huang",
"Jiazhao Zhang",
"Shilong Zou",
"Xinwang Liu",
"Ruizhen Hu",
"Kai Xu"
] | Poster | null | Predictive manipulation has recently gained considerable attention in the Embodied AI community due to its potential to improve robot policy performance by leveraging predicted states. However, generating accurate future visual states of robot-object interactions from world models remains a well-known challenge, partic... | [
"Robot Learning: Model Learning",
"Control & Dynamics",
"Grasping & Manipulation"
] | null | 167 | 2505.11528 | title_snapshot |
HT34hQcU91 | GLOVER++: Unleashing the Potential of Affordance Learning from Human Behaviors for Robotic Manipulation | https://openreview.net/forum?id=HT34hQcU91 | [
"Teli Ma",
"Jia Zheng",
"Zifan Wang",
"Ziyao Gao",
"Jiaming Zhou",
"Junwei Liang"
] | Poster | null | Learning manipulation skills from human demonstration videos offers a promising path toward generalizable and interpretable robotic intelligence—particularly through the lens of *actionable affordances*. However, transferring such knowledge remains challenging due to: 1) a lack of large-scale datasets with precise af... | [
"Actionable Affordance",
"Affordance Transfer",
"Vision-Language Model",
"Human Demonstrations",
"Robotic Manipulation"
] | null | 174 | 2505.11865 | title_snapshot |
GIHDBntBu9 | TrackVLA: Embodied Visual Tracking in the Wild | https://openreview.net/forum?id=GIHDBntBu9 | [
"Shaoan Wang",
"Jiazhao Zhang",
"Minghan Li",
"Jiahang Liu",
"Anqi Li",
"Kui Wu",
"Fangwei Zhong",
"Junzhi Yu",
"Zhizheng Zhang",
"He Wang"
] | Poster | null | Embodied visual tracking is a fundamental skill in Embodied AI, enabling an agent to follow a specific target in dynamic environments using only egocentric vision. This task is inherently challenging as it requires both accurate target recognition and effective trajectory planning under conditions of severe occlusion a... | [
"Embodied Visual Tracking",
"Vision-Language-Action Model"
] | TrackVLA is a unified Vision-Language-Action model jointly learns object recognition and trajectory planning, achieving state-of-the-art embodied visual tracking performance in dynamic environments. | 177 | 2505.23189 | title_snapshot |
kbrb6W7bUL | SDS – See it, Do it, Sorted: Quadruped Skill Synthesis from Single Video Demonstration | https://openreview.net/forum?id=kbrb6W7bUL | [
"Maria Stamatopoulou",
"Jeffrey Li",
"Dimitrios Kanoulas"
] | Poster | null | Imagine a robot learning locomotion skills from any single video, without labels or reward engineering. We introduce SDS ("See it. Do it. Sorted."), an automated pipeline for skill acquisition from unstructured video demonstrations. Using GPT-4o, SDS applies novel prompting techniques, in the form of spatio-temporal gr... | [
"Skill Imitation",
"Robot Skill Learning",
"Quadrupedal Robot"
] | Fully automated pipeline for quadruped locomotion skill learning from a single video via GPT-4o–generated reward functions, requiring no manual tuning. | 178 | 2410.11571 | title_snapshot |
PTJopl2uaU | Shortcut Learning in Generalist Robot Policies: The Role of Dataset Diversity and Fragmentation | https://openreview.net/forum?id=PTJopl2uaU | [
"Youguang Xing",
"Xu Luo",
"Junlin Xie",
"Lianli Gao",
"Heng Tao Shen",
"Jingkuan Song"
] | Poster | null | Generalist robot policies trained on large-scale datasets such as Open X-Embodiment (OXE) demonstrate strong performance across a wide range of tasks. However, they often struggle to generalize beyond the distribution of their training data. In this paper, we investigate the underlying cause of this limited generalizat... | [
"Generalist Robot Policies",
"Shortcut Learning",
"Large-scale Robotic Datasets"
] | Identifying shortcut learning as the key impediment to the generalization of generalist robot policies and providing a comprehensive analysis. | 179 | 2508.06426 | title_snapshot |
xzR8rBRgPp | Estimating Value of Assistance for Online POMDP Robotic Agents | https://openreview.net/forum?id=xzR8rBRgPp | [
"Yuval Goshen",
"Sarah Keren"
] | Poster | null | Robotic agents operating in dynamic, partially observable environments often benefit from teammate assistance. We address the challenge of determining when and how to assist in multi-robot systems where agents can modify the physical environment, such as moving obstacles that block perception or manipulation. For robot... | [
"POMDP",
"Online Planning"
] | null | 181 | null | null |
rJRFFDVTnf | O$^3$Afford: One-Shot 3D Object-to-Object Affordance Grounding for Generalizable Robotic Manipulation | https://openreview.net/forum?id=rJRFFDVTnf | [
"Tongxuan Tian",
"Xuhui Kang",
"Yen-Ling Kuo"
] | Poster | null | Grounding object affordance is fundamental to robotic manipulation as it establishes the critical link between perception and action among interacting objects. However, prior works predominantly focus on predicting single-object affordance, overlooking the fact that most real-world interactions involve relationships be... | [
"Affordance Grounding",
"Few-shot Learning",
"Vision Foundation Models"
] | We propose a novel framework for one-shot affordance learning that leverages vision foundation models to enable effective robotic manipulation. | 187 | 2509.06233 | title_snapshot |
Lv5lmtLGIQ | Towards Embodiment Scaling Laws in Robot Locomotion | https://openreview.net/forum?id=Lv5lmtLGIQ | [
"Bo Ai",
"Liu Dai",
"Nico Bohlinger",
"Dichen Li",
"Tongzhou Mu",
"Zhanxin Wu",
"K. Fay",
"Henrik I Christensen",
"Jan Peters",
"Hao Su"
] | Poster | null | Developing generalist agents that operate across diverse tasks, environments, and robot embodiments is a grand challenge in robotics and artificial intelligence. While substantial progress has been made in cross-task and cross-environment generalization, achieving broad generalization to novel embodiments remains elusi... | [
"Cross-Embodiment Learning",
"Robot Locomotion",
"Robotic Foundation Models",
"Reinforcement Learning",
"Imitation Learning"
] | We conduct a large-scale study on embodiment scaling laws using robot locomotion as a testbed. | 194 | 2505.05753 | title_snapshot |
yOWUy97hmd | Learning Long-Horizon Robot Manipulation Skills via Privileged Action | https://openreview.net/forum?id=yOWUy97hmd | [
"Xiaofeng Mao",
"Yucheng XU",
"Zhaole Sun",
"Elle Miller",
"Daniel Layeghi",
"Michael Mistry"
] | Poster | null | Long-horizon contact-rich tasks are challenging to learn with reinforcement learning, due to ineffective exploration of high-dimensional state spaces with sparse rewards. The learning process often gets stuck in local optimum and demands task-specific reward fine-tuning for complex scenarios. In this work, we propose ... | [
"Reinforcement learning",
"Grasping & Manipulation",
"Currirulum Learning"
] | We propose a framework leveraging privileged actions and curriculum learning to efficiently solve long-horizon, contact-rich manipulation tasks, achieving robust behaviors and outperforming state-of-the-art methods. | 196 | 2502.15442 | title_snapshot |
qAGHsCsA1O | Predictive Red Teaming: Breaking Policies Without Breaking Robots | https://openreview.net/forum?id=qAGHsCsA1O | [
"Anirudha Majumdar",
"Mohit Sharma",
"Dmitry Kalashnikov",
"Sumeet Singh",
"Pierre Sermanet",
"Vikas Sindhwani"
] | Poster | null | Visuomotor policies trained via imitation learning are capable of performing challenging manipulation tasks, but are often extremely brittle to lighting, visual distractors, and object locations. These vulnerabilities can depend unpredictably on the specifics of training, and are challenging to expose without time-cons... | [
"red teaming",
"policy evaluation",
"targeted data collection"
] | Discovering vulnerabilities of a visuomotor policy with respect to changes in environmental factors without hardware evaluations in off-nominal conditions. | 197 | 2502.06575 | title_snapshot |
eLeCrM5PEO | Self-supervised perception for tactile skin covered dexterous hands | https://openreview.net/forum?id=eLeCrM5PEO | [
"Akash Sharma",
"Carolina Higuera",
"Chaithanya Krishna Bodduluri",
"Zixi Liu",
"Taosha Fan",
"Tess Hellebrekers",
"Mike Lambeta",
"Byron Boots",
"Michael Kaess",
"Tingfan Wu",
"Francois Robert Hogan",
"Mustafa Mukadam"
] | Poster | null | We present PercepSkin, a pre-trained encoder for magnetic skin sensors distributed across the fingertips, phalanges, and palm of a dexterous robot hand.
Magnetic tactile skins offer a flexible form factor for hand-wide coverage with fast response times, in contrast to vision-based tactile sensors that are restricted t... | [
"Robot Perception",
"Sensing & Vision",
"Representation learning",
"Foundation models",
"Tactile sensing"
] | null | 198 | 2505.11420 | title_snapshot |
HMcBBIg1Th | Extracting Visual Plans from Unlabeled Videos via Symbolic Guidance | https://openreview.net/forum?id=HMcBBIg1Th | [
"Wenyan Yang",
"Ahmet Tikna",
"Yi Zhao",
"Yuying Zhang",
"Luigi Palopoli",
"Marco Roveri",
"Joni Pajarinen"
] | Poster | null | Visual planning, by offering a sequence of intermediate visual subgoals to a goal-conditioned low-level policy, achieves promising performance on long-horizon manipulation tasks. To obtain the subgoals, existing methods typically resort to video generation models but suffer from model hallucination and computational co... | [
"offline imitation learning",
"planning"
] | null | 202 | 2505.08444 | title_snapshot |
19LSN4QnV4 | FOMO-3D: Using Vision Foundation Models for Long-Tailed 3D Object Detection | https://openreview.net/forum?id=19LSN4QnV4 | [
"Anqi Joyce Yang",
"James Tu",
"Nikita Dvornik",
"Enxu Li",
"Raquel Urtasun"
] | Poster | null | In order to navigate complex traffic environments, self-driving vehicles
must recognize many semantic classes pertaining to vulnerable road users or
traffic control devices. However, many safety-critical objects (e.g.,
construction worker) appear infrequently in nominal traffic conditions, leading to a
severe shortage... | [
"Long-Tailed 3D Object Detection",
"Vision Foundation Model",
"Multimodal Fusion",
"Autonomous Vehicles"
] | We propose FOMO-3D, the first to leverage vision foundation models (specifically OWLv2 for 2D object detection and Metric3Dv2 for dense depths) with novel multi-modal fusion designs to tackle long-tailed 3D object detection. | 203 | 2603.08611 | title_snapshot |
sUWOSP6SUJ | Adapt3R: Adaptive 3D Scene Representation for Domain Transfer in Imitation Learning | https://openreview.net/forum?id=sUWOSP6SUJ | [
"Albert Wilcox",
"Mohamed Ghanem",
"Masoud Moghani",
"Pierre Barroso",
"Benjamin Joffe",
"Animesh Garg"
] | Poster | null | Imitation Learning can train robots to perform complex and diverse manipulation tasks, but learned policies are brittle with observations outside of the training distribution. 3D scene representations that incorporate observations from calibrated RGBD cameras have been proposed as a way to mitigate this, but in our eva... | [
"Imitation Learning",
"3D Perception",
"Cross-Embodiment Learning"
] | null | 204 | 2503.04877 | title_snapshot |
bOVF8Rj33i | VT-Refine: Learning Bimanual Assembly with Visuo-Tactile Feedback via Simulation Fine-Tuning | https://openreview.net/forum?id=bOVF8Rj33i | [
"Binghao Huang",
"Jie Xu",
"Iretiayo Akinola",
"Wei Yang",
"Balakumar Sundaralingam",
"Rowland O'Flaherty",
"Dieter Fox",
"Xiaolong Wang",
"Arsalan Mousavian",
"Yu-Wei Chao",
"Yunzhu Li"
] | Poster | null | Humans excel at bimanual assembly tasks by adapting to rich tactile feedback—a capability that remains difficult to replicate in robots through behavioral cloning alone, due to the suboptimality and limited diversity of human demonstrations. In this work, we present VT-Refine, a visuo-tactile policy learning framework ... | [
"Tactile Simulation",
"Bimanual Manipulation",
"RL Fine-Tuning"
] | null | 208 | 2510.14930 | title_snapshot |
7OOMC7pzaw | TypeTele: Releasing Dexterity in Teleoperation by Dexterous Manipulation Types | https://openreview.net/forum?id=7OOMC7pzaw | [
"Yuhao Lin",
"Yi-Lin Wei",
"Haoran Liao",
"Mu Lin",
"Chengyi Xing",
"Hao Li",
"Dandan Zhang",
"Mark Cutkosky",
"Wei-Shi Zheng"
] | Poster | null | Dexterous teleoperation plays a crucial role in robotic manipulation for real-world data collection and remote robot control. Previous dexterous teleoperation mostly relies on hand retargeting to closely mimic human hand postures. However, these approaches may fail to fully leverage the inherent dexterity of dexterous ... | [
"Teleoperation",
"Dexterous",
"Manipulation"
] | null | 211 | 2507.01857 | title_snapshot |
TFbT7kHD89 | Constrained Style Learning from Imperfect Demonstrations under Task Optimality | https://openreview.net/forum?id=TFbT7kHD89 | [
"Kehan Wen",
"Chenhao Li",
"Junzhe He",
"Marco Hutter"
] | Poster | null | Learning from demonstration has proven effective in robotics for acquiring natural behaviors, such as stylistic motions and lifelike agility, particularly when explicitly defining style-oriented reward functions is challenging. Synthesizing stylistic motions for real-world tasks usually requires balancing task performa... | [
"Constrained Markov Decision Process",
"Imitation Learning",
"Legged Robots"
] | A Constraint Markov Decision Process (CMDP)–based policy optimization approach for learning from imperfect demonstrations to achieve lifelike agility and energy-efficient motions while enforcing near-optimal task performance. | 212 | 2507.09371 | title_snapshot |
mjYKNIRqpy | GC-VLN: Instruction as Graph Constraints for Training-free Vision-and-Language Navigation | https://openreview.net/forum?id=mjYKNIRqpy | [
"Hang Yin",
"Haoyu Wei",
"Xiuwei Xu",
"Wenxuan Guo",
"Jie Zhou",
"Jiwen Lu"
] | Poster | null | In this paper, we propose a training-free framework for vision-and-language navigation (VLN). Existing zero-shot VLN methods are mainly designed for discrete environments or involve unsupervised training in continuous simulator environments, which makes it challenging to generalize and deploy them in real-world scenari... | [
"Vision-Language Navigation",
"Graph Constraint"
] | null | 214 | 2509.10454 | title_snapshot |
a9RXjOt5bU | The Sound of Simulation: Learning Multimodal Sim-to-Real Robot Policies with Generative Audio | https://openreview.net/forum?id=a9RXjOt5bU | [
"Renhao Wang",
"Haoran Geng",
"Tingle Li",
"Philipp Wu",
"Feishi Wang",
"Gopala Anumanchipalli",
"Trevor Darrell",
"Boyi Li",
"Pieter Abbeel",
"Jitendra Malik",
"Alexei A Efros"
] | Oral | null | Robots must integrate multiple sensory modalities to act effectively in the real world. Yet, learning such multimodal policies at scale remains challenging. Simulation offers a viable solution, but while vision has benefited from high-fidelity simulators, other modalities (e.g. sound) can be notoriously difficult to si... | [
"generative modeling",
"real2sim",
"sim2real",
"multimodal learning"
] | We augment physics-based simulators with generative models capable of simulating modalities traditionally absent from simulation, enabling learning multimodal sim2real policies that transfer zero-shot to real-world tasks. | 219 | 2507.02864 | title_snapshot |
XrgRvBklWu | DexUMI: Using Human Hand as the Universal Manipulation Interface for Dexterous Manipulation | https://openreview.net/forum?id=XrgRvBklWu | [
"Mengda Xu",
"Han Zhang",
"Yifan Hou",
"Zhenjia Xu",
"Linxi Fan",
"Manuela Veloso",
"Shuran Song"
] | Oral | null | We present DexUMI - a data collection and policy learning framework that uses the human hand as the natural interface to transfer dexterous manipulation skills to various robot hands. DexUMI incorporates hardware and software adaptations to minimize the embodiment gap between the human hand and various robot hands. The... | [
"Dexterous Manipulation",
"Learning from Human",
"Imitation Learning"
] | We present DexUMI - a data collection and policy learning framework that uses the human hand as the natural interface to transfer dexterous manipulation skills to various robot hands. | 220 | 2505.21864 | title_snapshot |
HqqyJ9A2fy | Neural Robot Dynamics | https://openreview.net/forum?id=HqqyJ9A2fy | [
"Jie Xu",
"Eric Heiden",
"Iretiayo Akinola",
"Dieter Fox",
"Miles Macklin",
"Yashraj Narang"
] | Poster | null | Accurate and efficient simulation of modern robots remains challenging due to their high degrees of freedom and intricate mechanisms. Neural simulators have emerged as a promising alternative to traditional analytical simulators, capable of efficiently predicting complex dynamics and adapting to real-world data; howeve... | [
"Robot Model Learning; Robotics Simulation; Neural Simulation"
] | We learn a robot-centric neural physics module that is genearlizable across different tasks and environment setups for a particular robot. | 222 | 2508.15755 | title_snapshot |
SpFH8T7gjM | SLAC: Simulation-Pretrained Latent Action Space for Whole-Body Real-World RL | https://openreview.net/forum?id=SpFH8T7gjM | [
"Jiaheng Hu",
"Peter Stone",
"Roberto Martín-Martín"
] | Poster | null | Building capable household and industrial robots requires mastering the control of versatile, high-degree-of-freedom (DoF) systems such as mobile manipulators. While reinforcement learning (RL) holds promise for autonomously acquiring robot control policies, scaling it to high-DoF embodiments remains challenging. Direc... | [
"Real-world RL",
"Latent Action",
"Whole-body Manipulation"
] | We introduce SLAC, a method that makes real-world RL feasible for high DoF robots by leveraging a low-fidelity simulator to pretrain a task-agnostic latent action space. | 223 | null | null |
bmpDAqsJov | Latent Adaptive Planner for Dynamic Manipulation | https://openreview.net/forum?id=bmpDAqsJov | [
"Donghun Noh",
"Deqian Kong",
"Minglu Zhao",
"Andrew Lizarraga",
"Jianwen Xie",
"Ying Nian Wu",
"Dennis Hong"
] | Poster | null | This paper presents Latent Adaptive Planner (LAP), a novel approach for dynamic nonprehensile manipulation tasks that formulates planning as latent space inference, effectively learned from human demonstration videos.
Our method addresses key challenges in visuomotor policy learning through a principled variationa... | [
"Imitation Learning",
"Robotics",
"Dynamic Nonprehensile Manipulation",
"Latent Space Planning",
"Classical Variational Bayes",
"Video-based Skill Acquisition"
] | null | 224 | 2505.03077 | title_snapshot |
EcOGafgvuC | FACET: Force-Adaptive Control via Impedance Reference Tracking for Legged Robots | https://openreview.net/forum?id=EcOGafgvuC | [
"Botian Xu",
"Haoyang Weng",
"Qingzhou Lu",
"Yang Gao",
"Huazhe Xu"
] | Oral | null | Reinforcement learning (RL) has made significant strides in legged robot control, enabling locomotion across diverse terrains and complex loco-manipulation capabilities.
However, the commonly used position or velocity tracking-based objectives are agnostic to forces experienced by the robot, leading to stiff and potent... | [
"Reinforcement Learning"
] | null | 225 | 2505.06883 | title_snapshot |
9FpccnRarn | BEVCalib: LiDAR-Camera Calibration via Geometry-Guided Bird’s-Eye View Representation | https://openreview.net/forum?id=9FpccnRarn | [
"Weiduo Yuan",
"Jerry Li",
"Justin Yue",
"Divyank Shah",
"Konstantinos Karydis",
"Hang Qiu"
] | Poster | null | Accurate LiDAR-camera calibration is the foundation of accurate multimodal fusion environmental perception for autonomous driving and robotic systems. Traditional calibration methods require extensive data collection in controlled environments and cannot compensate for the transformation changes during the vehicle/robo... | [
"LiDAR-Camera Calibration",
"Autonomous Driving",
"BEV Features"
] | BEVCalib, the first model using bird's-eye view(BEV) features to perform LiDAR-camera calibration from raw data. | 226 | 2506.02587 | title_judge |
tXY6VQlXfA | CHD: Coupled Hierarchical Diffusion for Long-Horizon Tasks | https://openreview.net/forum?id=tXY6VQlXfA | [
"Ce Hao",
"Anxing Xiao",
"Zhiwei Xue",
"Harold Soh"
] | Poster | null | Diffusion-based planners have shown strong performance in short-horizon tasks but often fail in complex, long-horizon settings. We trace the failure to loose coupling between high-level (HL) sub-goal selection and low-level (LL) trajectory generation, which leads to incoherent plans and degraded performance. We propose... | [
"Diffusion Planner",
"Long-horizon Planning",
"Hierarchical Planning"
] | Coupled Hierarchical Diffusion (CHD) jointly models sub-goals and trajectories with feedback-driven correction for scalable long-horizon planning. | 230 | 2505.07261 | title_snapshot |
EXgckdYESp | PicoPose: Progressive Pixel-to-Pixel Correspondence Learning for Novel Object Pose Estimation | https://openreview.net/forum?id=EXgckdYESp | [
"Lihua Liu",
"Jiehong Lin",
"ZhenXin Liu",
"Kui Jia"
] | Poster | null | RGB-based novel object pose estimation is critical for rapid deployment in robotic applications, yet zero-shot generalization remains a key challenge. In this paper, we introduce PicoPose, a novel framework designed to tackle this task using a three-stage pixel-to-pixel correspondence learning process. Firstly, PicoPos... | [
"Novel Object Pose Estimation",
"Robotic Manipulation"
] | null | 234 | 2504.02617 | title_snapshot |
2y7TSgwqAB | GENNAV: Polygon Mask Generation for Generalized Referring Navigable Regions | https://openreview.net/forum?id=2y7TSgwqAB | [
"Kei Katsumata",
"Yui Iioka",
"Naoki Hosomi",
"Teruhisa Misu",
"Kentaro Yamada",
"Komei Sugiura"
] | Poster | null | We focus on the task of identifying the location of target regions from a natural language instruction and a front camera image captured by a mobility.
This task is challenging because it requires both existence prediction and segmentation mask generation, particularly for stuff-type target regions with ambiguous bound... | [
"Autonomous driving",
"Vision and Language",
"Semantic Understanding"
] | We propose GENNAV, a polygon-based segmentation method which predicts the existence of target regions and generates segmentation masks for stuff-type target regions. | 236 | 2508.21102 | title_snapshot |
BcJlmjF1vV | HyperTASR: Hypernetwork-Driven Task-Aware Scene Representations for Robust Manipulation | https://openreview.net/forum?id=BcJlmjF1vV | [
"Li Sun",
"Jiefeng Wu",
"Feng Chen",
"Ruizhe Liu",
"Yanchao Yang"
] | Poster | null | Effective policy learning for robotic manipulation requires scene representations that selectively capture task-relevant environmental features. Current approaches typically employ task-agnostic representation extraction, failing to emulate the dynamic perceptual adaptation observed in human cognition. We present Hyper... | [
"Representation learning",
"manipulation",
"hypernetworks"
] | We present HyperTASR, a novel framework for task-aware scene representations in robotic manipulation that dynamically adapts perceptual processing based on both task objectives and execution progression through hypernetworks. | 240 | 2508.18802 | title_snapshot |
ZPJo9RJL15 | Unsupervised Skill Discovery as Exploration for Learning Agile Locomotion | https://openreview.net/forum?id=ZPJo9RJL15 | [
"Seungeun Rho",
"Kartik Garg",
"Morgan Byrd",
"Sehoon Ha"
] | Poster | null | Exploration is crucial for legged robots to learn agile locomotion behaviors capable of overcoming diverse obstacles.
For example, a robot may need to try different contact patterns and momentum profiles to successfully jump over an obstacle—but encouraging such diverse exploration is inherently challenging. As a resu... | [
"Unsupervised Reinforcement Learning",
"Locomotion",
"Quadruped",
"Skill Discovery"
] | By applying unsupervised reinforcement learning, we trained agile locomotion behaviors without relying on curriculum or reference motions. | 241 | 2508.08982 | title_snapshot |
VVhAhzr2WV | Real2Render2Real: Scaling Robot Data Without Dynamics Simulation or Robot Hardware | https://openreview.net/forum?id=VVhAhzr2WV | [
"Justin Yu",
"Letian Fu",
"Huang Huang",
"Karim El-Refai",
"Rares Andrei Ambrus",
"Richard Cheng",
"Muhammad Zubair Irshad",
"Ken Goldberg"
] | Oral | null | Scaling robot learning requires vast and diverse datasets. Yet the prevailing data collection paradigm—human teleoperation—remains costly and constrained by manual effort and physical robot access. We introduce Real2Render2Real (R2R2R), a novel approach for generating robot training data without relying on object dynam... | [
"Robot Datasets",
"Imitation Learning",
"Data Augmentation"
] | Real2Render2Real (R2R2R) generates robot data from a single video demo and 3D object scan, without teleop or dynamics sim, by tracking object motion, synthesizing diverse trajectories, and rendering demonstrations for policy learning. | 244 | 2505.09601 | title_snapshot |
iq0wUf6dUy | Reflective Planning: Vision-Language Models for Multi-Stage Long-Horizon Robotic Manipulation | https://openreview.net/forum?id=iq0wUf6dUy | [
"Yunhai Feng",
"Jiaming Han",
"Zhuoran Yang",
"Xiangyu Yue",
"Sergey Levine",
"Jianlan Luo"
] | Poster | null | Solving complex long-horizon robotic manipulation problems requires sophisticated high-level planning capabilities, the ability to reason about the physical world, and reactively choose appropriate motor skills. Vision-language models (VLMs) pretrained on Internet data could in principle offer a framework for tackling ... | [
"Vision-Language Models",
"Long-Horizon Manipulation",
"Manipulation Planning"
] | null | 245 | 2502.16707 | title_snapshot |
PwKsCO6TAF | KineSoft: Learning Proprioceptive Manipulation Policies with Soft Robot Hands | https://openreview.net/forum?id=PwKsCO6TAF | [
"Uksang Yoo",
"Jonathan Francis",
"Jean Oh",
"Jeffrey Ichnowski"
] | Oral | null | Underactuated soft robot hands offer inherent safety and adaptability advantages over rigid systems, but developing dexterous manipulation skills remains challenging. While imitation learning shows promise for complex manipulation tasks, traditional approaches struggle with soft systems due to demonstration collection ... | [
"Soft Robots",
"Proprioceptive Estimation",
"Imitation Learning"
] | We propose KineSoft, a framework that enables soft robot hands to perform dexterous manipulation by combining proprioceptive mesh shape sensing, kinesthetic demonstration, and a behavior cloning policy trained to track deformation trajectories. | 246 | 2503.01078 | title_snapshot |
FdgtV9mO6j | Towards Generalizable Safety in Crowd Navigation via Conformal Uncertainty Handling | https://openreview.net/forum?id=FdgtV9mO6j | [
"Jianpeng Yao",
"Xiaopan Zhang",
"Yu Xia",
"Zejin Wang",
"Amit Roy-Chowdhury",
"Jiachen Li"
] | Poster | null | Mobile robots navigating in crowds trained using reinforcement learning are known to suffer performance degradation when faced with out-of-distribution scenarios. We propose that by properly accounting for the uncertainties of pedestrians, a robot can learn safe navigation policies that are robust to distribution shift... | [
"Crowd Navigation",
"Reinforcement Learning",
"Distribution Shift"
] | We present a reinforcement learning algorithm that maintains safe crowd navigation under distribution shifts by explicitly modeling and constraining prediction uncertainty. | 248 | 2508.05634 | title_snapshot |
RUSscFSEfD | FastUMI: A Scalable and Hardware-Independent Universal Manipulation Interface with Dataset | https://openreview.net/forum?id=RUSscFSEfD | [
"Zhaxizhuoma",
"Kehui Liu",
"Chuyue Guan",
"Zhongjie Jia",
"Ziniu Wu",
"Xin Liu",
"Tianyu Wang",
"Shuai Liang",
"Pengan CHEN",
"Pingrui Zhang",
"Haoming Song",
"Delin Qu",
"Dong Wang",
"Zhigang Wang",
"Nieqing Cao",
"Yan Ding",
"Bin Zhao",
"Xuelong Li"
] | Poster | null | Real-world manipulation datasets for robotic arms remain scarce due to the high costs, rigid hardware dependencies, and complex setup procedures associated with existing data collection methods.
We introduce, a redesigned Universal Manipulation Interface (UMI) that addresses these challenges, enabling low-cost, scalabl... | [
"Imitation Learning",
"Manipulation",
"Data Collection"
] | null | 249 | 2409.19499 | title_snapshot |
XoRtWWjXuC | AgentWorld: An Interactive Simulation Platform for Scene Construction and Mobile Robotic Manipulation | https://openreview.net/forum?id=XoRtWWjXuC | [
"Yizheng Zhang",
"Zhenjun Yu",
"JiaXin Lai",
"Cewu Lu",
"Lei Han"
] | Poster | null | We introduce AgentWorld, an interactive simulation platform for developing household mobile manipulation capabilities. Our platform combines automated scene construction that encompasses layout generation, semantic asset placement, visual material configuration, and physics simulation, with a dual-mode teleoperation sy... | [
"Simulation Platform",
"scene construction",
"teleoperation",
"simulation dataset"
] | null | 254 | 2508.07770 | title_snapshot |
AP7kM1xk2a | Do LLM Modules Generalize? A Study on Motion Generation for Autonomous Driving | https://openreview.net/forum?id=AP7kM1xk2a | [
"Mingyi Wang",
"Jingke Wang",
"Tengju Ye",
"Kaicheng Yu"
] | Poster | null | Recent breakthroughs in large language models (LLMs) have not only advanced natural language processing but also inspired their application in domains with structurally similar problems—most notably, autonomous driving motion generation. Both domains involve autoregressive sequence modeling, token-based representations... | [
"Large Language Model",
"Autonomous Driving",
"Motion Generation",
"Sim Agent"
] | null | 259 | 2509.02754 | title_snapshot |
kto4zVmo4w | One View, Many Worlds: Single-Image to 3D object Meets Generative Domain Randomization for One-Shot 6D Pose Estimation | https://openreview.net/forum?id=kto4zVmo4w | [
"Zheng Geng",
"Nan Wang",
"Shaocong Xu",
"Chongjie Ye",
"Bohan Li",
"Zhaoxi Chen",
"Sida Peng",
"Hao Zhao"
] | Oral | null | Estimating the 6D pose of arbitrary objects from a single reference image is a critical yet challenging task in robotics, especially considering the long-tail distribution of real-world instances. While category-level and model-based approaches have achieved notable progress, they remain limited in generalizing to unse... | [
"Unseen Object Pose Estimation",
"Generative Model",
"Robot Manipulation"
] | Single-Image to 3D object Meets Generative Domain Randomization for One-Shot 6D Pose Estimation | 262 | 2509.07978 | title_snapshot |
5ySSVlJBOn | FetchBot: Learning Generalizable Object Fetching in Cluttered Scenes via Zero-Shot Sim2Real | https://openreview.net/forum?id=5ySSVlJBOn | [
"Weiheng Liu",
"Yuxuan Wan",
"Jilong Wang",
"Yuxuan Kuang",
"Xuesong Shi",
"Haoran Li",
"Dongbin Zhao",
"Zhizheng Zhang",
"He Wang"
] | Oral | null | Generalizable object fetching in cluttered scenes remains a fundamental and application-critical challenge in embodied AI. Closely packed objects cause inevitable occlusions, making safe action generation particularly difficult. Under such partial observability, effective policies must not only generalize across divers... | [
"Generalizable Fetching",
"Sim2Real",
"Occlusion Handling"
] | null | 263 | 2502.17894 | title_snapshot |
AVDCwK1dek | Bipedal Balance Control with Whole-body Musculoskeletal Standing and Falling Simulations | https://openreview.net/forum?id=AVDCwK1dek | [
"Chengtian Ma",
"Yunyue Wei",
"Chenhui Zuo",
"Chen Zhang",
"Yanan Sui"
] | Poster | null | Balance control is important for human and bipedal robotic systems. While dynamic balance during locomotion has received considerable attention, quantitative understanding of static balance and falling remains limited. This work presents a hierarchical control pipeline for simulating human balance via a comprehensive w... | [
"balance control",
"bipedal standing and falling",
"musculoskeletal system"
] | null | 264 | 2506.09383 | title_snapshot |
DiyLz91PKF | UniTac2Pose: A Unified Approach Learned in Simulation for Category-level Visuotactile In-hand Pose Estimation | https://openreview.net/forum?id=DiyLz91PKF | [
"Mingdong Wu",
"Long Yang",
"Jin Liu",
"Weiyao Huang",
"Lehong Wu",
"Zelin Chen",
"Daolin Ma",
"Hao Dong"
] | Poster | null | Accurate estimation of the in-hand pose of an object based on its CAD model is crucial in both industrial applications and everyday tasks—ranging from positioning workpieces and assembling components to seamlessly inserting devices like USB connectors. While existing methods often rely on regression, feature matching, ... | [
"Tactile Object Pose Estimation",
"Diffusion Model",
"Precise Manipulation"
] | We propose a novel framework, trained in pure simulation, that integrates visuotactile in-hand object pose estimation, tracking, and uncertainty estimation into a unified system. | 277 | 2509.15934 | title_snapshot |
4IuTfpWGDR | TReF-6: Inferring Task-Relevant Frames from a Single Demonstration for One-Shot Skill Generalization | https://openreview.net/forum?id=4IuTfpWGDR | [
"Yuxuan Ding",
"Shuangge Wang",
"Tesca Fitzgerald"
] | Poster | null | Robots often struggle to generalize from a single demonstration due to the lack of a transferable and interpretable spatial representation. In this work, we introduce TReF-6, a method that infers a simplified, abstracted 6DoF Task-Relevant Frame from a single trajectory. Our approach identifies an influence point purel... | [
"Spatial Reference Frames",
"One-Shot Imitation Learning",
"Dynamic Movement Primitives"
] | null | 279 | 2509.00310 | title_snapshot |
ksOrtEgIC0 | AirExo-2: Scaling up Generalizable Robotic Imitation Learning with Low-Cost Exoskeletons | https://openreview.net/forum?id=ksOrtEgIC0 | [
"Hongjie Fang",
"Chenxi Wang",
"Yiming Wang",
"Jingjing Chen",
"Shangning Xia",
"Jun Lv",
"Zihao He",
"Xiyan Yi",
"Yunhan Guo",
"Xinyu Zhan",
"Lixin Yang",
"Weiming Wang",
"Cewu Lu",
"Hao-Shu Fang"
] | Oral | null | Scaling up robotic imitation learning for real-world applications requires efficient and scalable demonstration collection methods. While teleoperation is effective, it depends on costly and inflexible robot platforms. In-the-wild demonstrations offer a promising alternative, but existing collection devices have key li... | [
"Robotic Manipulation",
"Scalable Data Collection",
"Generalizable Imitation Policy"
] | null | 281 | 2503.03081 | title_snapshot |
Bl2VfU9NhF | Hold My Beer: Learning Gentle Humanoid Locomotion and End-Effector Stabilization Control | https://openreview.net/forum?id=Bl2VfU9NhF | [
"Yitang Li",
"Yuanhang Zhang",
"Wenli Xiao",
"Chaoyi Pan",
"Haoyang Weng",
"Guanqi He",
"Tairan He",
"Guanya Shi"
] | Poster | null | Can your humanoid walk up and hand you a full cup of beer—without spilling a drop? While humanoids are increasingly featured in flashy demos—dancing, delivering packages, traversing rough terrain—fine-grained control during locomotion remains a significant challenge. In particular, stabilizing a filled end-effector (EE... | [
"Humanoid Robots",
"Reinforcement Learning",
"Stable Locomotion"
] | We decouple humanoid control into high-frequency arm and low-frequency leg policies, enabling stable end-effector stable control during dynamic locomotion. | 283 | 2505.24198 | title_snapshot |
seSw6ssEid | Residual Neural Terminal Constraint for MPC-based Collision Avoidance in Dynamic Environments | https://openreview.net/forum?id=seSw6ssEid | [
"Bojan Derajic",
"Mohamed-Khalil Bouzidi",
"Sebastian Bernhard",
"Wolfgang Hönig"
] | Poster | null | In this paper, we propose a hybrid MPC local planner that uses a learning-based approximation of a time-varying safe set, derived from local observations and applied as the MPC terminal constraint. This set can be represented as a zero-superlevel set of the value function computed via Hamilton-Jacobi (HJ) reachability ... | [
"MPC",
"Obstacle Avoidance",
"Learning for Control"
] | A hybrid MPC local planner that uses a learning-based approximation of a time-varying safe set, derived from local observations and applied as the MPC terminal constraint. | 286 | 2508.03428 | title_snapshot |
BTUioBmCWo | Phantom: Training Robots Without Robots Using Only Human Videos | https://openreview.net/forum?id=BTUioBmCWo | [
"Marion Lepert",
"Jiaying Fang",
"Jeannette Bohg"
] | Poster | null | Training general-purpose robots requires learning from large and diverse data sources. Current approaches rely heavily on teleoperated demonstrations which are difficult to scale. We present a scalable framework for training manipulation policies directly from human video demonstrations, requiring no robot data. Our me... | [
"Learning from Human Videos",
"Imitation Learning"
] | We present a scalable method for training robot manipulation policies directly from human videos—without any robot data—achieving zero-shot deployment and up to 92% success on diverse tasks. | 289 | 2503.00779 | title_snapshot |
BNCh3SS1Yl | Articulate AnyMesh: Open-vocabulary 3D Articulated Objects Modeling | https://openreview.net/forum?id=BNCh3SS1Yl | [
"Xiaowen Qiu",
"Jincheng Yang",
"Yian Wang",
"Zhehuan Chen",
"Yufei Wang",
"Tsun-Hsuan Wang",
"Zhou Xian",
"Chuang Gan"
] | Poster | null | 3D articulated objects modeling has long been a challenging problem, since it requires to capture both accurate surface geometries and semantically meaningful and spatially precise structures, parts, and joints. Existing methods heavily depend on training data from a limited set of handcrafted articulated object catego... | [
"Articulated objects",
"visual prompting",
"URDF prediction"
] | null | 293 | 2502.02590 | title_snapshot |
HprBJupvvM | Gen2Act: Human Video Generation in Novel Scenarios enables Generalizable Robot Manipulation | https://openreview.net/forum?id=HprBJupvvM | [
"Homanga Bharadhwaj",
"Debidatta Dwibedi",
"Abhinav Gupta",
"Shubham Tulsiani",
"Carl Doersch",
"Ted Xiao",
"Dhruv Shah",
"Fei Xia",
"Dorsa Sadigh",
"Sean Kirmani"
] | Poster | null | How can robot manipulation policies generalize to novel tasks involving unseen object types and new motions? In this paper, we provide a solution in terms of predicting motion information from web data through human video generation and conditioning a robot policy on the generated video. Instead of attempting to scale ... | [
"Robot Manipulation",
"Human Video Generation",
"Human Videos"
] | Casting language-conditioned manipulation as human video generation followed by closed-loop policy execution conditioned on the generated video enables solving diverse real-world tasks involving object/motion types unseen in the robot dataset. | 295 | 2409.16283 | title_snapshot |
o0LBjJxUeS | Learning Long-Context Diffusion Policies via Past-Token Prediction | https://openreview.net/forum?id=o0LBjJxUeS | [
"Marcel Torne Villasevil",
"Andy Tang",
"Yuejiang Liu",
"Chelsea Finn"
] | Poster | null | Reasoning over long sequences of observations and actions is essential for many robotic tasks.
Yet, learning effective long-context policies from demonstrations remains challenging.
As context length increases, training becomes increasingly expensive due to rising memory demands, and policy performance often degrades... | [
"history-conditioned policy learning"
] | null | 300 | 2505.09561 | title_snapshot |
1K3kjo91Q1 | Learning from 10 Demos: Generalisable and Sample-Efficient Policy Learning with Oriented Affordance Frames | https://openreview.net/forum?id=1K3kjo91Q1 | [
"Krishan Rana",
"Jad Abou-Chakra",
"Sourav Garg",
"Robert Lee",
"Ian Reid",
"Niko Suenderhauf"
] | Poster | null | Imitation learning has unlocked the potential for robots to exhibit highly dexterous behaviours. However, it still struggles with long-horizon, multi-object tasks due to poor sample efficiency and limited generalisation. Existing methods require a substantial number of demonstrations to cover possible task variations, ... | [
"behaviour cloning",
"spatial generalisation",
"intra-category generalisation",
"long-horizon tasks",
"affordances"
] | Learning spatial and intra-category invariant diffusion policies using oriented affordance task frames for compositional generalisation in long-horizon, multi-object tasks using only 10 demonstrations. | 316 | 2410.12124 | title_snapshot |
zEC8TOXDkH | GraspVLA: a Grasping Foundation Model Pre-trained on Billion-scale Synthetic Action Data | https://openreview.net/forum?id=zEC8TOXDkH | [
"Shengliang Deng",
"Mi Yan",
"Songlin Wei",
"Haixin Ma",
"Yuxin Yang",
"Jiayi Chen",
"Zhiqi Zhang",
"Taoyu Yang",
"Xuheng Zhang",
"Heming Cui",
"Zhizheng Zhang",
"He Wang"
] | Poster | null | Embodied foundation models are gaining increasing attention for their zero-shot generalization, scalability, and adaptability to new tasks through few-shot post-training. However, existing models rely heavily on real-world data, which is costly and labor-intensive to collect. Synthetic data offers a cost-effective alte... | [
"Vision-Language-Action",
"Large-scale Robot Learning",
"Grasping"
] | null | 317 | 2505.03233 | title_snapshot |
Ggu7Hh2xnn | RobotxR1: Enabling Embodied Robotic Intelligence on Large Language Models through Closed-Loop Reinforcement Learning | https://openreview.net/forum?id=Ggu7Hh2xnn | [
"Liam Boyle",
"Nicolas Baumann",
"Paviththiren Sivasothilingam",
"Michele Magno",
"Luca Benini"
] | Poster | null | Future robotic systems operating in real-world environments require
on-board embodied intelligence without continuous cloud connection, balancing
capabilities with constraints on computational power and memory. This work
presents an extension of the R1-zero approach, which enables the usage of small
parameter-count Lar... | [
"Large Language Models",
"Reinforcement Learning",
"Embodied AI",
"Constrained Hardware"
] | This paper extends R1-zero to embodied AI by training small LLMs via closed-loop RL for autonomous driving, enabling edge deployment with reasoning and adaptability once limited to significantly larger models. | 325 | 2505.03238 | title_snapshot |
Th1kFSnjUW | MoTo: A Zero-shot Plug-in Interaction-aware Navigation for General Mobile Manipulation | https://openreview.net/forum?id=Th1kFSnjUW | [
"Zhenyu Wu",
"Angyuan Ma",
"Xiuwei Xu",
"Hang Yin",
"Yinan Liang",
"Ziwei Wang",
"Jiwen Lu",
"Haibin Yan"
] | Poster | null | Mobile manipulation is the fundamental challenge for robotics in assisting humans with diverse tasks and environments in everyday life.
Conventional mobile manipulation approaches often struggle to generalize across different tasks and environments due to the lack of large-scale training. However, recent advances in ma... | [
"Mobile Manipulation",
"VLA",
"VLM"
] | null | 338 | 2509.01658 | title_snapshot |
O2RGlBfR9H | ImLPR: Image-based LiDAR Place Recognition using Vision Foundation Models | https://openreview.net/forum?id=O2RGlBfR9H | [
"Minwoo Jung",
"Lanke Frank Tarimo Fu",
"Maurice Fallon",
"Ayoung Kim"
] | Poster | null | LiDAR Place Recognition (LPR) is a key component in robotic localization, enabling robots to align current scans with prior maps of their environment. While Visual Place Recognition (VPR) has embraced Vision Foundation Models (VFMs) to enhance descriptor robustness, LPR has relied on task-specific models with limited u... | [
"LiDAR Place Recognition",
"Deep Learning",
"Vision Foundation Model"
] | Range Image Views for LiDAR Place Recognition with a vision foundation model | 340 | 2505.18364 | title_snapshot |
irh5o90Mj1 | Long-VLA: Unleashing Long-Horizon Capability of Vision Language Action Model for Robot Manipulation | https://openreview.net/forum?id=irh5o90Mj1 | [
"Yiguo Fan",
"Shuanghao Bai",
"Xinyang Tong",
"Pengxiang Ding",
"Yuyang Zhu",
"Hongchao Lu",
"Fengqi Dai",
"Wei Zhao",
"Yang Liu",
"Siteng Huang",
"Zhaoxin Fan",
"Badong Chen",
"Donglin Wang"
] | Poster | null | Vision-Language-Action (VLA) models have become a cornerstone in robotic policy learning, leveraging large-scale multimodal data for robust and scalable control. However, existing VLA frameworks primarily address short-horizon tasks, and their effectiveness on long-horizon, multi-step robotic manipulation remains limit... | [
"VLA; Manipulation"
] | null | 348 | 2508.19958 | title_snapshot |
5htQM8jqOe | D-Cubed: Latent Diffusion Trajectory Optimisation for Dexterous Deformable Manipulation | https://openreview.net/forum?id=5htQM8jqOe | [
"Jun Yamada",
"Shaohong Zhong",
"Jack Collins",
"Ingmar Posner"
] | Poster | null | Mastering deformable object manipulation often necessitates the use of anthropomorphic, high-degree-of-freedom robot hands capable of precise, contact-rich control. However, current trajectory optimisation methods often struggle in these settings due to the large search space and the sparse task information available f... | [
"Trajectory Optimisation",
"Dexterous Deformable Object Manipulation",
"Latent Diffusion Model",
"Gradient-Free Guidance"
] | We present a trajectory optimisation method using a gradient-free guided sampling within the reverse process of skill-latent diffusion models to solve dexterous deformable object manipulation tasks. | 355 | 2403.12861 | title_snapshot |
xpEjjGC82v | COMBO-Grasp: Learning Constraint-Based Manipulation for Bimanual Occluded Grasping | https://openreview.net/forum?id=xpEjjGC82v | [
"Jun Yamada",
"Alexander Luis Mitchell",
"Jack Collins",
"Ingmar Posner"
] | Poster | null | This paper addresses the challenge of occluded robot grasping, i.e. grasping in situations where the desired grasp poses are kinematically infeasible due to environmental constraints such as surface collisions. Existing RL methods struggle with task complexity, and collecting expert demonstrations is often impractical.... | [
"Occluded Grasping",
"Bimanual Manipulation",
"Reinforcement Learning"
] | We propose COMBO-Grasp, a bimanual learning-based framework combining self-supervised and RL policies to tackle occluded robotic grasping tasks. | 356 | 2502.08054 | title_snapshot |
End of preview. Expand in Data Studio
README.md exists but content is empty.
- Downloads last month
- 59