Direct Code Access in Self-Organizing Neural Networks for Reinforcement Learning.
Ah‐Hwee Tan · 2007
TD-FALCON is a self-organizing neural network that incorporates Temporal Difference (TD) meth-ods for reinforcement learning. Despite the advan-tages of fast and stable learning, TD-FALCON still relies on an iterative process to evaluate each avail-able action in a decision cycle. To remove this defi-ciency, this paper presents a direct code access pro-cedure whereby TD-FALCON conducts instanta-neous searches for cognitive nodes that match with the current states and at the same time providemax-imal reward values. Our comparative experiments show that TD-FALCON with direct code access produces comparable performance with the origi-nal TD-FALCON while improving significantly in computation efficiency and network complexity. 1