7 points | by walterbell 20 hours ago
3 comments
Technically, we could say?
(1) Single-loop: fixes actions within fixed rules, like Reinforcement Learning.
(2) Double-loop: questions and adapts the rules, somewhat like Meta-Reinforcement Learning.
Technically, we could say?
(1) Single-loop: fixes actions within fixed rules, like Reinforcement Learning.
(2) Double-loop: questions and adapts the rules, somewhat like Meta-Reinforcement Learning.