Meta-Evolve: Continuous Robot Evolution for One-to-many Policy Transfer
Not on the map. Why?
Abstract (arXiv)
We investigate the problem of transferring an expert policy from a source robot to multiple different robots. To solve this problem, we propose a method named Meta-Evolve that uses continuous robot evolution to efficiently transfer the policy to each target robot through a set of tree-structured evolutionary robot sequences. The robot evolution tree allows the robot evolution paths to be shared, so our approach can significantly outperform naive one-to-one policy transfer. We present a heuristic approach to determine an optimized robot evolution tree. Experiments have shown that our method is able to improve the efficiency of one-to-three transfer of manipulation policy by up to 3.2× and one-to-six transfer of agile locomotion policy by 2.4× in terms of simulation cost over the baseline of launching multiple independent one-to-one policy transfers.
Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Builds on 31 of its 53 references that are in the map. First posted on arXiv 6 May 2024.
Builds on
31 of 53 references.
References · 31
- 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
- 2017Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networkscited by 4,861
- 2016Generative Adversarial Imitation Learningcited by 1,576
- 2021Ego4D: Around the World in 3,000 Hours of Egocentric Videocited by 1,530
- 2017Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real Worldcited by 1,401
- 2021Isaac Gym: High Performance GPU-Based Physics Simulation For Robot Learningcited by 882
- 2018Scaling Egocentric Vision: The EPIC-KITCHENS Datasetcited by 789
- 2017Learning Complex Dexterous Manipulation with Deep Reinforcement Learning and Demonstrationscited by 685
- 2017Sim-to-Real Transfer of Robotic Control with Dynamics Randomizationcited by 616
- 2019Meta-Learning with Implicit Gradientscited by 523
- 2020robosuite: A Modular Simulation Framework and Benchmark for Robot Learningcited by 484
- 2015Policy Distillationcited by 414
19 more
- 2021DexYCB: A Benchmark for Capturing Hand Grasping of Objectscited by 320
- 2018Learning to Adapt in Dynamic, Real-World Environments Through Meta-Reinforcement Learningcited by 270
- 2018Probabilistic Model-Agnostic Meta-Learningcited by 254
- 2023Orbit: A Unified Simulation Framework for Interactive Robot Learning Environmentscited by 239
- 2017One-Shot Imitation Learningcited by 232
- 2020One Policy to Control Them All: Shared Modular Policies for Agent-Agnostic Controlcited by 114
- 2020State-Only Imitation Learning for Dexterous Manipulationcited by 90
- 2017Learning Invariant Feature Spaces to Transfer Skills with Reinforcement Learningcited by 84
- 2022MetaMorph: Learning Universal Controllers with Transformerscited by 78
- 2019State Alignment-based Imitation Learningcited by 73
- 2019Learning to Control Self-Assembling Morphologies: A Study of Generalization via Modularitycited by 68
- 2017Towards Generalization and Simplicity in Continuous Controlcited by 65
- 2021Meta-Learning with Fewer Tasks through Task Interpolationcited by 45
- 2018Meta Reinforcement Learning with Latent Variable Gaussian Processescited by 44
- 2020Hierarchically Decoupled Imitation for Morphological Transfercited by 30
- 2020Meta-Reinforcement Learning for Robotic Industrial Insertion Taskscited by 28
- 2022AnyMorph: Learning Transferable Polices By Inferring Agent Morphologycited by 25
- 2022REvolveR: Continuous Evolutionary Models for Robot-to-robot Policy Transfercited by 12
- 2022HERD: Continuous Human-to-Robot Evolution for Learning from Human Demonstrationcited by 8
Built on by
Citing papers per year: 2024: 1; 2025: 1; 2026: 1.
3 papers.