3 ms·
TD-Gammon from 1992 was an early success of neural network based reinforcement learning. Descendants of the technique remain standard today in RL.
by currymj 6y ago
TD-Gammon from 1992 was an early success of neural network based reinforcement learning. Descendants of the technique remain standard today in RL.