The purpose of this week's lecture is to explore a new class of associative networks, called autoassociative memories. These are networks that involve a single set of processors that are connected to one another. As a result, the processors are both inputs and outputs. These systems are dynamic, changing their activity over time until they stablize into a state that represents an answer to an input problem.

The activity for this week involves using Microsoft Excel to explore the operation of a particular type of autoassociative network, the Hopfield net. One spreadsheet let's you play with letter recognition, while the other demonstrates how one might use a Hopfield net to get response latencies 'for free'.