search

RobKlock / tddm-heuristic-model

Model of the TDDM heuristic currently under development.
0 stars 0 forks source link

2/22/22 #13

Open RobKlock opened 2 years ago

RobKlock commented 2 years ago

After the update rule is selectively applied, ignore the weights that weren't updated on the next trial (they aren't associated with the terminating event)

RobKlock commented 2 years ago

Need to break down ramps further after terminating event. (If we have A/A, when we have A/C, don't update the ramps that learned A/A)

RobKlock commented 2 years ago

When a stimulus occurs, look at when the House Light goes off next, calculate all the activations between the onset event and each event between them. update free timers accordingly (A/A, A/B, A/A/B, etc)