1A1-O02 Effect of Cognitive Bias in Robotic Motion Learning : Application of Loosely Symmetric Model to Q-learning(Evolution and Learning for Robotics)
Daisuke Uragami, Hisham Alsubeheen, Tatsuji Takahashi, Akinori SEKIGUCHI, Yoshiki Matsuo · The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec) · 2011
The Loosely Symmetric model (LS model) is a model of human illogical decision. A previous study revealed that agents implemented LS model achieved high performance for the N-arm bandit problem. In this study, we evaluate the performance of LS model applied in Q-learning, using the Acrobot example.