Jonathan Mugan and Benjamin Kuipers. 2009.
A comparison of strategies for developmental action acquisition in QLAP.
International Conference on Epigenetic Robotics (EpiRob-09).
An important part of development is ac- quiring actions to interact with the envi- ronment. We have developed a computa- tional model of autonomous action acquisi- tion, called QLAP. In this paper we investi- gate different strategies for developmental ac- tion acquisition within this model. In par- ticular, we introduce a way to actively learn actions and we compare this active action ac- quisition with passive learning of actions. We also compare curiosity based exploration with random exploration. And finally, we exam- ine the effects of resource restrictions on the agent's ability to learn actions.