Ensemble-CIO: Full-body dynamic motion planning that transfers to physical humanoids

By Igor Mordatch, Kendall Lowrey, E. Todorov. 2015

171 citations recorded by the source corpus.

Subfields: Biomedical Engineering, Computer Vision and Pattern Recognition, Control and Systems Engineering, Reinforcement Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Biomedical Engineering

In Computer Vision and Pattern Recognition

In Control and Systems Engineering

In Reinforcement Learning

Citations in the corpus

Domain randomization for transferring deep neural networks from simulation to the real world (2017, background): “In the context of physics adaptation, Mordatch and collaborators [30] show that training a policy on an ensemble of dynamics models can make the controller robust to modeling error and improve transfer to a real robot.”
Closing the Sim-to-Real Loop: Adapting Simulation Randomization with Real World Experience (2018, methodology): “resentations of real and synthetic data [16], using generative models to make synthetic images more realistic [17], fine-tuning the policies trained in simulation in the real world [18], learning inverse dynamics models [19], multiobjective optimization of task fitness and transferability to the rea”
Sim-to-Real: Learning Agile Locomotion For Quadruped Robots (2018, methodology): “A robust controller can be learned by injecting noise to the simulation [19], adding perturbations to the simulated robot [37], leveraging multiple simulators [3], using domain randomization [48] and dynamics randomization [29, 40, 34].”
Learning Complex Dexterous Manipulation with Deep Reinforcement Learning and Demonstrations (2017, positive): “By training policies to succeed on the entire ensemble, we can acquire policies that are explicitly trained for robustness, similar in spirit to previously proposed model ensemble methods [30], [39].”
Towards Generalization and Simplicity in Continuous Control (2017, background): “Even though the interactive tasks that we study are not as complex as those in Mordatch et al. (2015), to our knowledge, this is the first work that considers model-free approaches for continuous control tasks where a user can interact with the learned policy.”
Preparing for the Unknown: Learning a Universal Policy with Online System Identification (2017, background): “[21] who optimized the motion trajectory for an ensemble of dynamic models perturbed from the assumed one.”
EPOpt: Learning Robust Neural Network Policies Using Model Ensembles (2016, methodology): “Mordatch et al. (2015a) use a trajectory optimization approach and an ensemble with small finite set of models; whereas we follow a sampling based direct policy search approach over a continuous distribution of uncertain parameters, and also show domain adaptation.”
Transfer from Simulation to Real World through Learning Deep Inverse Dynamics Model (2016, methodology): “Our approach is agnostic to how the policy πsource was obtained.”
Learning Powerful Policies by Using Consistent Dynamics Model (2019, background): “Consider the humanoid robot from the MuJoCo environment (Mordatch et al., 2015).”
Sim-To-Real via Sim-To-Sim: Data-Efficient Robotic Grasping via Randomized-To-Canonical Adaptation Networks (2018, background): “There have been more robotics works that do not use vision, but that apply domain randomization on physical properties of the simulator to aid transferability [40, 46, 1, 68, 43].”
Simulation-based design of dynamic controllers for humanoid balancing (2016, methodology): “With the aid of these simulators, model-based approaches can effectively control humanoid robots in contact rich scenarios [26], [27].”
Reinforcement learning for non-prehensile manipulation: Transfer from simulation to physical system (2018, unknown): “The goal with ensemble approaches [6], [7],”
Robust Visual Domain Randomization for Reinforcement Learning. (2019, background): “…of each episode during training, has been shown to lead to improved generalization and transfer to the real world with little or no real world data (Mordatch et al., 2015; Rajeswaran et al., 2017; Tobin et al., 2017; Sadeghi & Levine, 2017; Antonova et al., 2017; Peng et al., 2018; OpenAI et al.,…”
NoRML: No-Reward Meta Learning (2019, unknown): “Othermethods haveleveragedpriorexperiencetolearnasinglepolicythatis robusttomanydifferentdynamics[20–22 ,”
Reasoning About Liquids via Closed-Loop Simulation (2017, background): “For rigid-body models, there has been some work connecting the modeled state of the system to the state of the real environment [16, 27].”
Efficient Bimanual Manipulation Using Learned Task Schemas (2019, methodology): “More recent techniques include learning from model ensembles [16] and utilizing domain randomization [17], [18], [19], in which physical properties of a simulated environment are randomized to allow learned policies to be robust.”