Zobrazeno 1 - 1
of 1
pro vyhledávání: '"Gautam, Simanta"'
Learning robust value functions given raw observations and rewards is now possible with model-free and model-based deep reinforcement learning algorithms. There is a third alternative, called Successor Representations (SR), which decomposes the value
Externí odkaz:
http://arxiv.org/abs/1606.02396
