Researched
Operant conditioning
B. F. Skinner showed in 1938 that behaviour is shaped by its consequences: rewards make a repeat more likely, punishments less likely.
Open in the interactive tree →Building on Thorndike's 1898 law of effect, Skinner's work underlies behavioural therapy and programmed teaching. Sutton and Barto's reinforcement learning in AI grew from the same idea of trial-and-error learning.
Prerequisites
Unlocks
- Learning by reward: TD-Gammon1992Reinforcement learning formalises Skinner's idea that behaviour is shaped by reward
- Dopamine and reward prediction1997