Corpus ID: 216036080

Efficient Adaptation for End-to-End Vision-Based Robotic Manipulation

@article{Julian2020EfficientAF,
  title={Efficient Adaptation for End-to-End Vision-Based Robotic Manipulation},
  author={Ryan C. Julian and Benjamin Swanson and Gaurav S. Sukhatme and Sergey Levine and Chelsea Finn and Karol Hausman},
  journal={ArXiv},
  year={2020},
  volume={abs/2004.10190}
}
One of the great promises of robot learning systems is that they will be able to learn from their mistakes and continuously adapt to ever-changing environments. Despite this potential, most of the robot learning systems today are deployed as a fixed policy and they are not being adapted after their deployment. Can we efficiently adapt previously learned behaviors to new environments, objects and percepts in the real world? In this paper, we present a method and empirical evidence towards a… Expand
Legged Robots that Keep on Learning: Fine-Tuning Locomotion Policies in the Real World
TLDR
It is demonstrated that a modest amount of real-world training can substantially improve performance during deployment, and this enables a real A1 quadrupedal robot to autonomously fine-tune multiple locomotion skills in a range of environments, including an outdoor lawn and a variety of indoor terrains. Expand
Grasp Adaptation for Robotic Manipulation Description
In this project, we will mainly follow a similar approach that has been studied in [3] where they use a simple meta-learning technique that pre-trains a joint policy on a very large real dataset.Expand
COG: Connecting New Skills to Past Experience with Offline Reinforcement Learning
TLDR
It is shown that even when the prior data does not actually succeed at solving the new task, it can still be utilized for learning a better policy, by providing the agent with a broader understanding of the mechanics of its environment. Expand
Bayesian Meta-Learning for Few-Shot Policy Adaptation Across Robotic Platforms
TLDR
The proposed method can successfully adapt a trained policy to different robotic platforms with novel physical parameters and the superiority of the metalearning algorithm compared to state-of-the-art methods for the introduced few-shot policy adaptation problem is demonstrated. Expand
Lifelong Robotic Reinforcement Learning by Retaining Experiences
TLDR
This work study a practical sequential multi-task RL problem that is motivated by the practical constraints of physical robotic systems, and derive an approach that effectively leverages the data and policies learned for previous tasks to cumulatively grow the robot’s skillset. Expand
Transfer Reinforcement Learning Across Homotopy Classes
TLDR
A novel fine-tuning algorithm is proposed that consists of a relaxing stage and a curriculum learning stage to enable transfer learning across homotopy classes and proves that the Ease-In-Ease-Out fine- Tuning method can successfully fine-tune in a sample-efficient way compared to existing baselines. Expand
Continual Learning for Real-World Autonomous Systems: Algorithms, Challenges and Frameworks
TLDR
The purpose of this study is to review the state-of-the-art methods that allow continuous learning of computational models over time and critically analyze the key challenges associated with continual learning for autonomous real-world systems. Expand
A Workflow for Offline Model-Free Robotic Reinforcement Learning
TLDR
This paper develops a practical workflow for using offline RL analogous to the relatively well-understood workflows for supervised learning problems, and devise a set of metrics and conditions that can be tracked over the course of offline training, and can inform the practitioner about how the algorithm and model architecture should be adjusted to improve final performance. Expand
Policy Transfer across Visual and Dynamics Domain Gaps via Iterative Grounding
TLDR
A novel policy transfer method with iterative “environment grounding”, IDAPT, that alternates between directly minimizing both visual and dynamics domain gaps by grounding the source environment in the target environment domains, and training a policy on the grounded source environment. Expand
Improved Reinforcement Learning Coordinated Control of a Mobile Manipulator using Joint Clamping
TLDR
This work studies joint penalties and joint limits of a state-of-the-art mobile manipulator whole-body controller that uses LIDAR sensing for obstacle collision avoidance and proposes directions to improve the reinforcement learning method. Expand
...
1
2
...

References

SHOWING 1-10 OF 80 REFERENCES
End-to-End Training of Deep Visuomotor Policies
TLDR
This paper develops a method that can be used to learn policies that map raw image observations directly to torques at the robot's motors, trained using a partially observed guided policy search method, with supervision provided by a simple trajectory-centric reinforcement learning method. Expand
Sim-to-Real Robot Learning from Pixels with Progressive Nets
TLDR
This work proposes using progressive networks to bridge the reality gap and transfer learned policies from simulation to the real world, and presents an early demonstration of this approach with a number of experiments in the domain of robot manipulation that focus on bridging thereality gap. Expand
QT-Opt: Scalable Deep Reinforcement Learning for Vision-Based Robotic Manipulation
TLDR
QT-Opt is introduced, a scalable self-supervised vision-based reinforcement learning framework that can leverage over 580k real-world grasp attempts to train a deep neural network Q-function with over 1.2M parameters to perform closed-loop, real- world grasping that generalizes to 96% grasp success on unseen objects. Expand
Visual Reinforcement Learning with Imagined Goals
TLDR
An algorithm is proposed that acquires general-purpose skills by combining unsupervised representation learning and reinforcement learning of goal-conditioned policies, efficient enough to learn policies that operate on raw image observations and goals for a real-world robotic system, and substantially outperforms prior techniques. Expand
Task-Embedded Control Networks for Few-Shot Imitation Learning
TLDR
Task-Embedded Control Networks are introduced, which employ ideas from metric learning in order to create a task embedding that can be used by a robot to learn new tasks from one or more demonstrations, and which surpass the performance of a state-of-the-art method when using only visual information from each demonstration. Expand
Adaptive Prior Selection for Repertoire-Based Online Adaptation in Robotics
TLDR
An algorithm, APROL (Adaptive Prior selection for Repertoire-based Online Learning) to plan the next action by incorporating these priors when the robot has no information about the current situation is presented. Expand
Learning to Adapt in Dynamic, Real-World Environments through Meta-Reinforcement Learning
TLDR
This work uses meta-learning to train a dynamics model prior such that, when combined with recent data, this prior can be rapidly adapted to the local context and demonstrates the importance of incorporating online adaptation into autonomous agents that operate in the real world. Expand
Unsupervised Visuomotor Control through Distributional Planning Networks
TLDR
This work aims to learn an unsupervised embedding space under which the robot can measure progress towards a goal for itself, and enables learning effective and control-centric representations that lead to more autonomous reinforcement learning algorithms. Expand
Deep visual foresight for planning robot motion
TLDR
This work develops a method for combining deep action-conditioned video prediction models with model-predictive control that uses entirely unlabeled training data and enables a real robot to perform nonprehensile manipulation — pushing objects — and can handle novel objects not seen during training. Expand
RoboNet: Large-Scale Multi-Robot Learning
TLDR
This paper proposes RoboNet, an open database for sharing robotic experience, which provides an initial pool of 15 million video frames, from 7 different robot platforms, and studies how it can be used to learn generalizable models for vision-based robotic manipulation. Expand
...
1
2
3
4
5
...