Autoassociative Memory

Autoassociative memory, also known as auto-association memory or an autoassociation network, is often misunderstood to be only a form of backpropagation or other neural networks. It is actually a more generic term that refers to all memories that enable one to retrieve a piece of data from only a tiny sample of itself.

Traditional memory stores data at a unique address and can recall the data upon presentation of the complete unique address. Autoassociative memories are capable of retrieving a piece of data upon presentation of only partial information from that piece of data. Heteroassociative memories, on the other hand, can recall an associated piece of datum from one category upon presentation of data from another category. Hopfield networks have been shown to act as autoassociative memory since they are capable of remembering data by observing a portion of that data. Biological neural networks, on the other hand, are heteroassociative memories since they can remember a completely different item to the one presented as input. Bidirectional Associative Memories (BAM) are Artificial Neural Networks that have long been used for performing heteroassociative recall.

For example, the fragments presented below should be all that's necessary to retrieve the appropriate memory:

1. "A day that will live in ______"
2. "To be or not to be"
3. "I came, I saw, I conquered"

The first example will make the reader fill in the blank with the word "infamy", while making him or her think of Franklin D. Roosevelt. The second example is only a tiny phrase from William Shakespeare's Hamlet, yet readers will be able to associate it with the play. And finally, most people will be quick to translate Caesar's quote over to "Veni, Vidi, Vici". The conclusion to be drawn is that Autoassociation networks can recreate the whole from merely its small parts.