Computational Neuroscience and Reinforcement Learning

Models of learning as prediction and reward: temporal-difference learning, dopamine as prediction error, and the split between model-based and model-free control. Relevance to our work: a term collision. 'Model learning' names an established idea here, so Levent's working label is written as template learning in the text. Founding texts (Sutton and Barto 1998; Schultz, Dayan and Montague 1997) not yet in the register; ruling requested before adding them.