Filters








1,400 Hits in 3.9 sec

Deep Imitation Learning of Sequential Fabric Smoothing From an Algorithmic Supervisor [article]

Daniel Seita, Aditya Ganapathi, Ryan Hoque, Minho Hwang, Edward Cen, Ajay Kumar Tanwani, Ashwin Balakrishna, Brijen Thananjeyan, Jeffrey Ichnowski, Nawid Jamali, Katsu Yamane, Soshi Iba, John Canny (+1 others)
2020 arXiv   pre-print
Due to the complexity of fabric states and dynamics, we apply deep imitation learning to learn policies that, given color (RGB), depth (D), or combined color-depth (RGBD) images of a rectangular fabric  ...  Sequential pulling policies to flatten and smooth fabrics have applications from surgery to manufacturing to home tasks such as bed making and folding clothes.  ...  We train a fabric smoothing policy in simulation using imitation learning on synthetic images.  ... 
arXiv:1910.04854v2 fatcat:zcibrbailjhhdephrdsn4q2v5a

VisuoSpatial Foresight for Physical Sequential Fabric Manipulation [article]

Ryan Hoque, Daniel Seita, Ashwin Balakrishna, Aditya Ganapathi, Ajay Kumar Tanwani, Nawid Jamali, Katsu Yamane, Soshi Iba, Ken Goldberg
2021 arXiv   pre-print
We build upon the Visual Foresight framework to learn fabric dynamics that can be efficiently reused to accomplish different sequential fabric manipulation tasks with a single goal-conditioned policy.  ...  Results suggest that training visual dynamics models using longer, corner-based actions can improve the efficiency of fabric folding by 76% and enable a physical sequential fabric folding task that VSF  ...  [67] , which was shown to be sufficiently accurate for imitation learning and sim-to-real transfer of fabric smoothing policies.  ... 
arXiv:2102.09754v2 fatcat:kyxkizg7ofg2vbsor6q6jclf3y

VisuoSpatial Foresight for Multi-Step, Multi-Task Fabric Manipulation [article]

Ryan Hoque, Daniel Seita, Ashwin Balakrishna, Aditya Ganapathi, Ajay Kumar Tanwani, Nawid Jamali, Katsu Yamane, Soshi Iba, Ken Goldberg
2021 arXiv   pre-print
We extend the Visual Foresight framework to learn fabric dynamics that can be efficiently reused to accomplish different fabric manipulation tasks with a single goal-conditioned policy.  ...  We experimentally evaluate VSF on multi-step fabric smoothing and folding tasks against 5 baseline methods in simulation and on the da Vinci Research Kit (dVRK) surgical robot without any demonstrations  ...  [49] , which was shown to be sufficiently accurate for learning reasonable fabric smoothing policies with imitation learning.  ... 
arXiv:2003.09044v3 fatcat:cgsbhrgtwnblhmugujvc6ckpya

Learning Dense Visual Correspondences in Simulation to Smooth and Fold Real Fabrics [article]

Aditya Ganapathi, Priya Sundaresan, Brijen Thananjeyan, Ashwin Balakrishna, Daniel Seita, Jennifer Grannen, Minho Hwang, Ryan Hoque, Joseph E. Gonzalez, Nawid Jamali, Katsu Yamane, Soshi Iba (+1 others)
2020 arXiv   pre-print
This makes it possible to robustly imitate a broad set of multi-step fabric smoothing and folding tasks on multiple physical robotic systems.  ...  In this paper, we learn visual correspondences for deformable fabrics across different configurations in simulation and show that this representation can be used to design policies for a variety of tasks  ...  [24] learn fabric folding policies by using deep reinforcement learning augmented with task-specific demonstrations.  ... 
arXiv:2003.12698v2 fatcat:ftdd3zxw5vektkcrb5q7zhxi7a

Learning Arbitrary-Goal Fabric Folding with One Hour of Real Robot Experience [article]

Robert Lee, Daniel Ward, Akansel Cosgun, Vibhavari Dasagi, Peter Corke, Jurgen Leitner
2020 arXiv   pre-print
In this paper, we show that it is possible to learn fabric folding skills in only an hour of self-supervised real robot experience, without human supervision or simulation.  ...  Our approach relies on fully convolutional networks and the manipulation of visual inputs to exploit learned features, allowing us to create an expressive goal-conditioned pick and place policy that can  ...  They apply domain randomization to transfer fabric smoothing policies to a real-world surgical robot.  ... 
arXiv:2010.03209v1 fatcat:wzajwisuizhbzla6wbohudqm7u

Deep Reinforcement Learning in Computer Vision: A Comprehensive Survey [article]

Ngan Le, Vidhiwar Singh Rathour, Kashu Yamazaki, Khoa Luu, Marios Savvides
2021 arXiv   pre-print
We start with comprehending the theories of deep learning, reinforcement learning, and deep reinforcement learning.  ...  Deep reinforcement learning augments the reinforcement learning framework and utilizes the powerful representation of deep neural networks.  ...  P-GAIL considers both smoothness and causal entropy in policy update by utilizing Deep P-Network [365] .  ... 
arXiv:2108.11510v1 fatcat:wkkqittwivbx5fpwg3nggcy7cm

Learning to Rearrange Deformable Cables, Fabrics, and Bags with Goal-Conditioned Transporter Networks [article]

Daniel Seita, Pete Florence, Jonathan Tompson, Erwin Coumans, Vikas Sindhwani, Ken Goldberg, Andy Zeng
2021 arXiv   pre-print
We propose embedding goal-conditioning into Transporter Networks, a recently proposed model architecture for learning robotic manipulation that rearranges deep features to infer displacements that can  ...  Rearranging and manipulating deformable objects such as cables, fabrics, and bags is a long-standing challenge in robotic manipulation.  ...  Tools used often involve either Imitation Learning (IL) [3] or Reinforcement Learning (RL) [58] , and in recent years, such techniques have been applied for manipulation of deformables.  ... 
arXiv:2012.03385v3 fatcat:ern2rrdeznbdpcnwm7u2r5vqai

A Survey on Interpretable Reinforcement Learning [article]

Claire Glanois, Paul Weng, Matthieu Zimmer, Dong Li, Tianpei Yang, Jianye Hao, Wulong Liu
2022 arXiv   pre-print
Although deep reinforcement learning has become a promising machine learning approach for sequential decision-making problems, it is still not mature enough for high-stake domains such as autonomous driving  ...  In such contexts, a learned policy needs for instance to be interpretable, so that it can be inspected before any deployment (e.g., for safety and verifiability reasons).  ...  in the space of programmatic policies via imitation learning.  ... 
arXiv:2112.13112v2 fatcat:wixrobyiwfabnfle2ftjfk64ki

Diluted Near-Optimal Expert Demonstrations for Guiding Dialogue Stochastic Policy Optimisation [article]

Thibault Cordier, Tanguy Urvoy, Lina M. Rojas-Barahona, Fabrice Lefèvre
2020 arXiv   pre-print
We present in this paper several imitation learning strategies for dialogue policy where the guiding expert is a near-optimal handcrafted policy.  ...  We notably propose a randomised exploration policy which allows for a seamless hybridisation of the learned policy and the expert.  ...  We think that exploring with smooth stochastic policy should improve the stability of learning in spite of exploitation strength.  ... 
arXiv:2012.04687v1 fatcat:uxj4bsznvfdhvfod2npdcehz24

Untangling Dense Knots by Learning Task-Relevant Keypoints [article]

Jennifer Grannen, Priya Sundaresan, Brijen Thananjeyan, Jeffrey Ichnowski, Ashwin Balakrishna, Minho Hwang, Vainavi Viswanath, Michael Laskey, Joseph E. Gonzalez, Ken Goldberg
2020 arXiv   pre-print
We instantiate this into an algorithm, HULK: Hierarchical Untangling from Learned Keypoints, which combines learning-based perception with a geometric planner into a policy that guides a bilateral robot  ...  HULK successfully untangles a cable from a dense initial configuration containing up to two overhand and figure-eight knots in 97.9% of 378 simulation experiments with an average of 12.1 actions per trial  ...  Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the sponsors.  ... 
arXiv:2011.04999v1 fatcat:kqclqjw7nfdqxiaaerladnuywi

Modeling, learning, perception, and control methods for deformable object manipulation

Hang Yin, Anastasia Varava, Danica Kragic
2021 Science Robotics  
of deformable objects.  ...  We adopt a learning perspective to unify the discussion over analytical and data-driven approaches, addressing how to use and integrate model priors and task data in perceiving and manipulating a variety  ...  In (37) , the authors train a deep RL policy to fold a cloth or to place the cloth on a hanger.  ... 
doi:10.1126/scirobotics.abd8803 pmid:34043538 fatcat:3q54vpiprrcsnpffmbduj3fjrm

2020 Index IEEE Robotics and Automation Letters Vol. 5

2020 IEEE Robotics and Automation Letters  
., +, LRA Oct. 2020 5448-5455 Reinforcement Learning for POMDP: Partitioned Rollout and Policy Iteration With Application to Autonomous Sequential Repair Problems.  ...  ., +, LRA Oct. 2020 5494-5501 Reinforcement Learning for POMDP: Partitioned Rollout and Policy Iteration With Application to Autonomous Sequential Repair Problems.  ... 
doi:10.1109/lra.2020.3032821 fatcat:qrnouccm7jb47ipq6w3erf3cja

Table of Contents

2021 IEEE Robotics and Automation Letters  
Gaze-Based Dual Resolution Deep Imitation Learning for High-Precision Dexterous Robot Manipulation . . . J. H. Ma, S. Sefati, R. H. Taylor, and M.  ...  Phee 2469 Imitation Learning of Hierarchical Driving Model: From Continuous Intention to Continuous Trajectory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  ... 
doi:10.1109/lra.2021.3072707 fatcat:qyphyzqxfrgg7dxdol4qamrdqu

A Systematic Survey on Deep Generative Models for Graph Generation [article]

Xiaojie Guo, Liang Zhao
2020 arXiv   pre-print
As one of a critical problem in this area, graph generation considers learning the distributions of given graphs and generating more novel graphs.  ...  This article provides an extensive overview of the literature in the field of deep generative models for the graph generation.  ...  Reinforcement Learning and Deep Q-Network Reinforcement learning (RL) is a commonly used framework for learning controlling policies by a computer algorithm, the so-called agent, through interacting with  ... 
arXiv:2007.06686v2 fatcat:xox7apwdvbfhlgnsgrr3w3rv5m

Learning-based Feedback Controller for Deformable Object Manipulation [article]

Biao Jia and Zhe Hu and Zherong Pan and Dinesh Manocha and Jia Pan
2018 arXiv   pre-print
An offline imitation learning framework is also proposed to achieve a control policy that is robust to large perturbations in the human-robot interaction.  ...  A feedback control policy is then optimized to push the object toward a desired featured status efficiently. The feedback policy can be learned either online or offline.  ...  imitation learning algorithm.  ... 
arXiv:1806.09618v2 fatcat:nekes2brfjcere3zz5w4j7ttmi
« Previous Showing results 1 — 15 out of 1,400 results