Warning: this page is under construction, and may not make much sense.

The Connectionist Navigational Map (CNM)

The Connectionist Navigational Map (CNM) is an architecture designed to enable an autonomus robot to learn about the spatial structure of its environment. So far, the CNM has only been tested on simple, simulated "grid-worlds":
Stand in the place where you live
Now face North
Think about direction
Wonder why you haven't before
Now stand in the place where you were
Now face West
Think about the place where you live
Wonder why you haven't before
 
If you are confused check with the sun
Carry a compass to help you along
Your feet are going to be on the ground
Your head is there to move you around
REM, "Stand"

Cluster diagram of the CNM

Here's a cluster analysis of the spatial representations the Connectionist Navigational Map (CNM) developed on its own. A full explanation is under construction, but the main point is that the CNM developed, on its own, similar representations for its different perspectives on the same locations (as can be seen by the fact that all the ways of getting to, e.g., the location (4,4) are clustered together).

This is an example of moving from a perspective-dependent understanding of a domain to a more objective, more perspective-independent understanding. I take this type of transition (and its converse) to be central to cognition.

The output for this diagram was generated using Andy Holland's implementation of the CNM.

A few references on the CNM are:


Ron Chrisley