Parallel learning: Overview and perspective for computational learning across Syn2Real and Sim2Real
Q Miao, Y Lv, M Huang, X Wang… - IEEE/CAA Journal of …, 2023 - ieeexplore.ieee.org
The virtual-to-real paradigm, ie, training models on virtual data and then applying them to
solve real-world problems, has attracted more and more attention from various domains by …
solve real-world problems, has attracted more and more attention from various domains by …
Robot learning from randomized simulations: A review
The rise of deep learning has caused a paradigm shift in robotics research, favoring
methods that require large amounts of data. Unfortunately, it is prohibitively expensive to …
methods that require large amounts of data. Unfortunately, it is prohibitively expensive to …
A walk in the park: Learning to walk in 20 minutes with model-free reinforcement learning
Deep reinforcement learning is a promising approach to learning policies in uncontrolled
environments that do not require domain knowledge. Unfortunately, due to sample …
environments that do not require domain knowledge. Unfortunately, due to sample …
Legged robots that keep on learning: Fine-tuning locomotion policies in the real world
Legged robots are physically capable of traversing a wide range of challenging
environments, but designing controllers that are sufficiently robust to handle this diversity …
environments, but designing controllers that are sufficiently robust to handle this diversity …
Towards zero domain gap: A comprehensive study of realistic lidar simulation for autonomy testing
Testing the full autonomy system in simulation is the safest and most scalable way to
evaluate autonomous vehicle performance before deployment. This requires simulating …
evaluate autonomous vehicle performance before deployment. This requires simulating …
What went wrong? closing the sim-to-real gap via differentiable causal discovery
Training control policies in simulation is more appealing than on real robots directly, as it
allows for exploring diverse states in an efficient manner. Yet, robot simulators inevitably …
allows for exploring diverse states in an efficient manner. Yet, robot simulators inevitably …
Real2sim2real: Self-supervised learning of physical single-step dynamic actions for planar robot casting
This paper introduces the task of Planar Robot Casting (PRC): where one planar motion of a
robot arm holding one end of a cable causes the other end to slide across the plane toward …
robot arm holding one end of a cable causes the other end to slide across the plane toward …
Visual reinforcement learning with self-supervised 3d representations
A prominent approach to visual Reinforcement Learning (RL) is to learn an internal state
representation using self-supervised methods, which has the potential benefit of improved …
representation using self-supervised methods, which has the potential benefit of improved …
Cross-domain policy adaptation via value-guided data filtering
Generalizing policies across different domains with dynamics mismatch poses a significant
challenge in reinforcement learning. For example, a robot learns the policy in a simulator …
challenge in reinforcement learning. For example, a robot learns the policy in a simulator …
Sim2real transfer learning for point cloud segmentation: An industrial application case on autonomous disassembly
C Wu, X Bi, J Pfrommer, A Cebulla… - Proceedings of the …, 2023 - openaccess.thecvf.com
On robotics computer vision tasks, generating and annotating large amounts of data from
real-world for the use of deep learning-based approaches is often difficult or even …
real-world for the use of deep learning-based approaches is often difficult or even …