2,808 Hits in 6.2 sec

What Matters in Learning from Offline Human Demonstrations for Robot Manipulation [article]

Ajay Mandlekar, Danfei Xu, Josiah Wong, Soroush Nasiriany, Chen Wang, Rohun Kulkarni, Li Fei-Fei, Silvio Savarese, Yuke Zhu, Roberto Martín-Martín
2021 arXiv   pre-print
Our study analyzes the most critical challenges when learning from offline human data for manipulation.  ...  In this paper, we conduct an extensive study of six offline learning algorithms for robot manipulation on five simulated and three real-world multi-stage manipulation tasks of varying complexity, and with  ...  physical robot tasks.  ... 
arXiv:2108.03298v2 fatcat:p7fmn5qjynftbjaymffoifpf5y

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

Ajay Mandlekar, Fabio Ramos, Byron Boots, Silvio Savarese, Li Fei-Fei, Animesh Garg, Dieter Fox
2020 arXiv   pre-print
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.  ...  We thank Kevin Shih for help with training models on image observations. We thank members of the NVIDIA Seattle Robotics Lab for several helpful discussions and feedback.  ... 
arXiv:1911.05321v2 fatcat:q6kp2tjdrreu3oefw5dyczulj4

Robot learning of industrial assembly task via human demonstrations

Maria Kyrarini, Muhammad Abdul Haseeb, Danijela Ristić-Durrant, Axel Gräser
2018 Autonomous Robots  
The traditional robot programming cannot cope with these challenges of human-robot collaboration. In this paper, a framework for robot learning by multiple human demonstrations is introduced.  ...  Additionally, the robot learns every path as needed for object manipulation (low-level learning).  ...  The authors thank MeRoSy industrial project partner pi4 robotics GmbH for the technical support in assembly scenario.  ... 
doi:10.1007/s10514-018-9725-6 fatcat:pgxk6qhfffguvpjxxbrkbvfdjq

Benchmark for Bimanual Robotic Manipulation of Semi-deformable Objects

Konstantinos Chatzilygeroudis, Bernardo Fichera, Ilaria Lauzana, Fanjun Bu, Kunpeng Yao, Farshad Khadivar, Aude Billard
2020 IEEE Robotics and Automation Letters  
We propose a new benchmarking protocol to evaluate algorithms for bimanual robotic manipulation semi-deformable objects.  ...  Index Terms-Performance evaluation and benchmarking, dual arm manipulation, model learning for control, dexterous manipulation.  ...  Zhu et al. survey in [27] the approaches to learn the assembly task using Learning from Demonstration, from either kinesthetic, motion-sensor or teleoperated demonstrations.  ... 
doi:10.1109/lra.2020.2972837 fatcat:bdchalm6wfhm5euyizpc5w4c3q

Batch Exploration with Examples for Scalable Robotic Reinforcement Learning [article]

Annie S. Chen, HyunJi Nam, Suraj Nair, Chelsea Finn
2020 arXiv   pre-print
Learning from diverse offline datasets is a promising path towards learning general purpose robotic agents.  ...  However, a core challenge in this paradigm lies in collecting large amounts of meaningful data, while not depending on a human in the loop for data collection.  ...  By minimizing the need for a human in the loop, this framework enables scalably collecting and learning from robotic data. provides some weak supervision to the agent to indicate what regions of the state  ... 
arXiv:2010.11917v1 fatcat:xu2kw5ffhvdulbs3rrsek2bf3y

Research Trends in Social Robots for Learning

Wafa Johal
2020 Current Robotics Reports  
A trend analysis is then proposed demonstrating the potential for educational context to nest impactful research from human-robot interaction.  ...  Secondly, we explore the potential for education to be the ideal context to investigate central human-robot interaction research questions.  ...  What can be taught with a social robot? What is the role of social robots in education? & Social adaptation and personalized learning: Can educational paradigm be applied to robots for learning?  ... 
doi:10.1007/s43154-020-00008-3 fatcat:ydoinppcwrh27ejewwyas42vhq

The Surprising Effectiveness of Representation Learning for Visual Imitation [article]

Jyothish Pari, Nur Muhammad Shafiullah, Sridhar Pandian Arunachalam, Lerrel Pinto
2021 arXiv   pre-print
Such joint learning causes an interdependence between these two problems, which often results in needing large amounts of demonstrations for learning.  ...  We experimentally show that this simple decoupling improves the performance of visual imitation models on both offline demonstration datasets and real-robot door opening compared to prior work in visual  ...  lightweight mobile manipulator for indoor human envi- Improved baselines with momentum contrastive learning. ronments. arXiv preprint arXiv:2109.10892, 2021.  ... 
arXiv:2112.01511v2 fatcat:qsysmd56mbdipkx2jt5wbkzrru

Quantifying Hypothesis Space Misspecification in Learning From Human–Robot Demonstrations and Physical Corrections

Andreea Bobu, Andrea Bajcsy, Jaime F. Fisac, Sampada Deglurkar, Anca D. Dragan
2020 IEEE Transactions on robotics  
We demonstrate our method on a 7 degree-of-freedom robot manipulator in learning from two important types of human input: demonstrations of manipulation tasks, and physical corrections during the robot's  ...  Recent work focuses on how robots can use such input - like demonstrations or corrections - to learn intended objectives.  ...  However, longer offline computation is possible in our learning-from-demonstrations scenario as the inference happens offline, after providing the robot with human demonstrations.  ... 
doi:10.1109/tro.2020.2971415 fatcat:btlfyed7ibdqddogkmtwcigxuu

Vision for Autonomous Vehicles and Probes (Dagstuhl Seminar 15461)

André Bruhn, Atsushi Imiya, Ales Leonardis, Tomas Pajdla, Marc Herbstritt
2016 Dagstuhl Reports  
Computer vision plays a key role in advanced driver assistance systems (ADAS) as well as in exploratory and service robotics.  ...  In 2013, Daimler succeeded autonomous driving on a public drive way. Today, the Curiosity mars rover is sending video views from Mars to Earth.  ...  Who Where and what is included in the environment maps? we should know limitations of online/offline mapping! what is the difference between online/offline mapping?  ... 
doi:10.4230/dagrep.5.11.36 dblp:journals/dagstuhl-reports/BruhnILP15 fatcat:l2nqd45tnrabpdqmwex6enkxei

Learning from Humans [chapter]

Aude G. Billard, Sylvain Calinon, Rüdiger Dillmann
2016 Springer Handbook of Robotics  
The field is best known as robot programming by demonstration, robot learning from/by demonstration, apprenticeship learning and imitation learning.  ...  This chapter surveys the main approaches developed to date to endow robots with the ability to learn from human guidance.  ...  Reinforcement learning (RL) appeared particularly suitable for type of problem. Indeed, imitation learning is limiting in that it requires the robot to learn only from what has been demonstrated.  ... 
doi:10.1007/978-3-319-32552-1_74 fatcat:wtcftkgkwveexpfbmnkcebi5wu

PLATO: Predicting Latent Affordances Through Object-Centric Play [article]

Suneel Belkhale, Dorsa Sadigh
2022 arXiv   pre-print
Due to this diverse coverage, existing approaches for learning from play are more robust to online policy deviations from the offline data distribution.  ...  Constructing a diverse repertoire of manipulation skills in a scalable fashion remains an unsolved challenge in robotics.  ...  Imitation Learning is a common method for learning robot policies from human demonstrations, where a policy is trained to mimic the demonstrated actions [22] .  ... 
arXiv:2203.05630v1 fatcat:zmhd4h5qlzfylhd7zkr6sezkiu

What Matters in Language Conditioned Robotic Imitation Learning [article]

Oier Mees, Lukas Hermann, Wolfram Burgard
2022 arXiv   pre-print
In this paper, we conduct an extensive study of the most critical challenges in learning language conditioned policies from offline free-form imitation datasets.  ...  While recently substantial advances have been achieved in language-driven robotics by leveraging end-to-end learning from pixels, there is no clear and well-understood process for making various design  ...  , it is difficult to asses what matters for language conditioned policy learning.  ... 
arXiv:2204.06252v1 fatcat:xnq35sp7xvelpfubinjpraec7a

Offline Reinforcement Learning as Anti-Exploration [article]

Shideh Rezaeifar, Robert Dadashi, Nino Vieillard, Léonard Hussenot, Olivier Bachem, Olivier Pietquin, Matthieu Geist
2021 arXiv   pre-print
Offline Reinforcement Learning (RL) aims at learning an optimal control from a fixed dataset, without interactions with the system.  ...  We thus take inspiration from the literature on bonus-based exploration to design a new offline RL agent.  ...  Learning complex dexterous manipulation with deep reinforcement learning and demonstrations. Proceedings of Robotics: Science and Systems (RSS), 2017. [43] D. J. Rezende, S. Mohamed, and D.  ... 
arXiv:2106.06431v1 fatcat:crppp6covnc7plgf6uttkwyili

Affordances in Robotic Tasks – A Survey [article]

Paola Ardón, Èric Pairet, Katrin S. Lohan, Subramanian Ramamoorthy, Ronald P. A. Petrick
2020 arXiv   pre-print
Affordances are key attributes of what must be perceived by an autonomous robotic agent in order to effectively interact with novel objects.  ...  specifications useful for robots.  ...  [54] improve the classification rate by learning from human demonstration to grasp objects in a household task.  ... 
arXiv:2004.07400v1 fatcat:hiexg6suzvg6ld7cwjtzyz4qka

Planning, Learning and Reasoning Framework for Robot Truck Unloading [article]

Fahad Islam, Anirudh Vemula, Sung-Kyun Kim, Andrew Dornbush, Oren Salzman, Maxim Likhachev
2020 arXiv   pre-print
We consider the task of autonomously unloading boxes from trucks using an industrial manipulator robot.  ...  In particular, motion planning and execution modules are evaluated in simulation and on a real robot, while offline learning and online decision-making are evaluated in simulated real-world scenarios.  ...  In addition to the real robot, we also employ a simulation of the robot, trailer, and boxes to estimate the outcomes of actions, and perform offline learning.  ... 
arXiv:1910.09453v2 fatcat:xrry6dfwdbgw5pu6rpvej6hd3m
« Previous Showing results 1 — 15 out of 2,808 results