- Abstract: Learning control policies in robotic tasks requires a large number of interactions due to small learning rates, bounds on the updates or unknown constraints. In contrast humans can infer protective and safe solutions after a single failure or unexpected observation. In order to reach similar performance, we developed a hierarchical Bayesian optimization algorithm that replicates the cognitive inference and memorization process for avoiding failures in motor control tasks. A Gaussian Process implements the modeling and the sampling of the acquisition function. This enables rapid learning with large learning rates while a mental replay phase ensures that policy regions that led to failures are inhibited during the sampling process. The features of the hierarchical Bayesian optimization method are evaluated in a simulated and physiological humanoid postural balancing task. We quantitatively compare the human learning performance to our learning approach by evaluating the deviations of the center of mass during training. Our results show that we can reproduce the efficient learning of human subjects in postural control tasks which provides a testable model for future physiological motor control tasks. In these postural control tasks, our method outperforms standard Bayesian Optimization in the number of interactions to solve the task, in the computational demands and in the frequency of observed failures.
- Keywords: Human Postural Control Model, Hierarchical Bayesian Optimization, Acquisition Function
- TL;DR: This paper presents a computational model for efficient human postural control adaptation based on hierarchical acquisition functions with well-known features.