Bio-Inspired Plastic Neural Networks for Zero-Shot Out-of-Distribution Generalization in Complex Animal-Inspired Robots
Not on the map. Why?
Abstract (arXiv)
Artificial neural networks can be used to solve a variety of robotic tasks. However, they risk failing catastrophically when faced with out-of-distribution (OOD) situations. Several approaches have employed a type of synaptic plasticity known as Hebbian learning that can dynamically adjust weights based on local neural activities. Research has shown that synaptic plasticity can make policies more robust and help them adapt to unforeseen changes in the environment. However, networks augmented with Hebbian learning can lead to weight divergence, resulting in network instability. Furthermore, such Hebbian networks have not yet been applied to solve legged locomotion in complex real robots with many degrees of freedom. In this work, we improve the Hebbian network with a weight normalization mechanism for preventing weight divergence, analyze the principal components of the Hebbian's weights, and perform a thorough evaluation of network performance in locomotion control for real 18-DOF dung beetle-like and 16-DOF gecko-like robots. We find that the Hebbian-based plastic network can execute zero-shot sim-to-real adaptation locomotion and generalize to unseen conditions, such as uneven terrain and morphological damage.
Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 25 of its 50 references that are in the map. First posted on arXiv 16 Mar 2025.
Builds on
25 of 50 references.
References · 25
- 2017Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real Worldcited by 1,498
- 2021Generalizing to Unseen Domains: A Survey on Domain Generalizationcited by 766
- 2019Solving Rubik's Cube with a Robot Handcited by 734
- 2017Sim-to-Real Transfer of Robotic Control with Dynamics Randomizationcited by 672
- 2017Evolution Strategies as a Scalable Alternative to Reinforcement Learningcited by 574
- 2021Learning to Walk in Minutes Using Massively Parallel Deep Reinforcement Learningcited by 512
- 2021RMA: Rapid Motor Adaptation for Legged Robotscited by 508
- 2020Sim-to-Real Transfer in Deep Reinforcement Learning for Robotics: a Surveycited by 413
- 2014Robots that can adapt like animalscited by 204
- 2018A Study on Overfitting in Deep Reinforcement Learningcited by 174
- 2022Walk These Ways: Tuning Robot Control for Generalization with Multiplicity of Behaviorcited by 138
- 2020Multi-expert learning of adaptive legged locomotioncited by 67
13 more
- 2021Robot Learning from Randomized Simulations: A Reviewcited by 63
- 2019Policies Modulating Trajectory Generatorscited by 49
- 2020Meta Learning Backpropagation And Improving Itcited by 49
- 2020Meta-Learning through Hebbian Plasticity in Random Networkscited by 47
- 2017Born to Learn: the Inspiration, Progress, and Future of Evolved Plastic Artificial Neural Networkscited by 29
- 2019Investigating Generalisation in Continuous Deep Reinforcement Learningcited by 18
- 2023Towards Self-Assembling Artificial Neural Networks through Neural Developmental Programscited by 17
- 2019Efficacy of Modern Neuro-Evolutionary Strategies for Continuous Control Optimizationcited by 15
- 2021Evolving and Merging Hebbian Learning Rules: Increasing Generalization by Decreasing the Number of Rulescited by 12
- 2023Learning Robust, Agile, Natural Legged Locomotion Skills in the Wildcited by 8
- 2022Meta-Reinforcement Learning with Self-Modifying Networkscited by 4
- 2020Evolving Inborn Knowledge For Fast Adaptation in Dynamic POMDP Problemscited by 3
- 2020Testing the Genomic Bottleneck Hypothesis in Hebbian Meta-Learningcited by 3
Built on by
Citing papers per year: 2025: 1.
1 paper.