SKINNER BOX
Behavioral Algorithms:
■ Acquisition (Reinforcement):
Every rewarded trial strengthens the behavior asymptotically. Rescorla-Wagner Model:
Learning rate slows as it approaches max association.
■ Extinction:
When the CS (or Lever) is presented without the US (Reward), the association exponentially decays.
■ Spontaneous Recovery:
After a rest period following extinction, the CR briefly rebounds before dying out again if unreinforced.
■ Acquisition (Reinforcement):
Every rewarded trial strengthens the behavior asymptotically. Rescorla-Wagner Model:
ΔV = αβ(λ - ΣV)Learning rate slows as it approaches max association.
■ Extinction:
When the CS (or Lever) is presented without the US (Reward), the association exponentially decays.
■ Spontaneous Recovery:
After a rest period following extinction, the CR briefly rebounds before dying out again if unreinforced.