Successor Uncertainties: Exploration and Uncertainty in Temporal Difference Learning
1 188
18
Microsoft Research334 тыс
Следующее
24.05.19 – 92748:23Популярные
Опубликовано 22 мая 2019, 21:17
Probabilistic Q-learning is a promising approach balancing exploration and exploitation in reinforcement learning.
However, existing implementations have significant limitations: they either fail to incorporate uncertainty about long-term consequences of actions or ignore fundamental dependencies in state-action values implied by the~Bellman equation. These problems result in sub-optimal exploration. As a solution, we develop Successor Uncertainties (SU), a probabilistic Q-learning method free of the aforementioned problems. SU outperforms existing baselines on tabular problems and on the Atari benchmark benchmark suite. Overall, SU is an improved and scalable probabilistic Q-learning method with better properties than its predecessors at no extra cost.
See more at microsoft.com/en-us/research/v...
However, existing implementations have significant limitations: they either fail to incorporate uncertainty about long-term consequences of actions or ignore fundamental dependencies in state-action values implied by the~Bellman equation. These problems result in sub-optimal exploration. As a solution, we develop Successor Uncertainties (SU), a probabilistic Q-learning method free of the aforementioned problems. SU outperforms existing baselines on tabular problems and on the Atari benchmark benchmark suite. Overall, SU is an improved and scalable probabilistic Q-learning method with better properties than its predecessors at no extra cost.
See more at microsoft.com/en-us/research/v...
Свежие видео
2 дня – 2 8731:05:03
4 дня – 6 2192:08
6 дней – 9373:57
8 дней – 8000:34
40 дней – 24 11125:53Случайные видео
89 дней – 6854:36
19.03.23 – 63 7390:57
18.12.22 – 3 227 23812:27
30.06.22 – 2266:45
02.09.10 – 1172:13
1 день – 141 1670:27
1 день – 102 2893:33
4 дня – 1 5240:53
4 дня – 4334:50
5 дней – 7010:50
8 дней – 1 0470:58
