Natural and Robust Walking using Reinforcement Learning without Demonstrations in High-Dimensional Musculoskeletal Models.
Pierre SchumacherThomas GeijtenbeekVittorio CaggianoVikash KumarSyn SchmittGeorg MartiusDaniel F. B. HaeuflePublished in: CoRR (2023)
Keyphrases
- reinforcement learning
- high dimensional
- probabilistic model
- dimensionality reduction
- function approximation
- statistical models
- prior knowledge
- machine learning algorithms
- discrete data
- neural network
- statistical model
- computationally efficient
- nearest neighbor
- state space
- dynamic programming
- hidden markov models
- feature space
- similarity measure
- learning algorithm