Reinforcement learning: an introduction
Автор: Rotova O.M., Shibanova A.D.
Журнал: Теория и практика современной науки @modern-j
Рубрика: Математика, информатика и инженерия
Статья в выпуске: 1 (55), 2020 года.
Бесплатный доступ
This article discusses one of the main provisions of the machine learning algorithm with reinforcement. Models of reinforcement learning are formulated. Three fundamental problems that reinforcement learning must solve are also considered - the trade-off between research and exploitation, the problem of delayed reward and the need to generalization.
Artificial intelligence, machine learning, reinforcement learning, markov decision process, state transitions, reward
Короткий адрес: https://sciup.org/140275028
IDR: 140275028