5.4.1 How to deal with multiple-trial learning

We start by assuming that we are only dealing with 'raw' information, i.e. as it comes from perception and as it is required for bodily action.

We need to account for how multiple experiences in similar circumstances could possibly result in an action that is a blended composite of the previous actions.

• Previously, for our simplistic model of one-shot learning, we just repeated the most recent action

To use the same video-tape-like memory model, we have to search through all relevant experiences

• We then have to integrate or blend the recorded actions
• We might also follow other strategies, such as selecting a single action sequences where the circumstances (initial perception) matches most closely.
• Alternatively we can use the perceptual similarities to set up a weighting scheme for integrating (averaging) the actions.