Sample-efficient learning of soft task priorities through bayesian optimization

We use the Bayesian Optimization procedure to regulate the RBFNs corresponding to different tasks based on performances indexes that are extracted for a fixed episode. In comparison with tuning the weights using another stochastic optimizationtechnique, i.e., CMA-ES, we find that the proposed approachrequires much less samples to evaluate.

Download paper here

Recommended citation:

@inproceedings{su2018humanoids,
  author={Y. {Su} and Y. {Wang} and A. {Kheddar}},
  Booktitle={IEEE-RAS Int. Conf. on Humanoid Robots}, 
  title={Sample-Efficient Learning of Soft Task Priorities Through Bayesian Optimization}, 
  Address = {Beijing, China},
  year={2018},
  Month = {15-17 October},
}