Title :
A Temporal Difference GNG-Based Algorithm That Can Learn to Control in Reinforcement Learning Environments
Author :
Vieira, Davi C. L. ; Adeodato, Paulo J. L. ; Goncalves, Paulo M.
Author_Institution :
Inst. Fed. de Sergipe - IFS, Aracaju, Brazil
Abstract :
This paper proposes a new reinforcement learning algorithm called TD-GNG that uses the Growing Neural Gas (GNG) network to deal with environments of large domains. The proposed algorithm is capable to reduce the dimensionality of the problem by aggregating similar states. In experimental comparison against tile-coding in mountain car and puddle world, the TD-GNG showed an increase in the generalization without loosing quality in the policy obtained.
Keywords :
generalisation (artificial intelligence); learning (artificial intelligence); neural nets; GNG network; TD-GNG; dimensionality reduction; generalization; growing neural gas network; reinforcement learning environments; temporal difference GNG-based algorithm; Convergence; Electronic mail; Equations; Learning (artificial intelligence); Mathematical model; Neural networks; Prediction algorithms; Adaptive State Space Partitioning; Growing Neural Gas; Reinforcement Learning; Temporal Difference;
Conference_Titel :
Machine Learning and Applications (ICMLA), 2013 12th International Conference on
Conference_Location :
Miami, FL
DOI :
10.1109/ICMLA.2013.67
