Monday, March 30, 2009

Neural Networks Follow the Golden Rule

We can understand the functioning of neural networks best when we see them as communities of neurons. Like communities or societies in the more conventional sense, their members behave in ways that reflect concern for the well-being of their neighbors.

Recalling Marshall McLuhan, we are able to do only what some portion of our body specializes in doing. Our various organs have cells that specialize in doing the work of that organ. We can be members of communities concerned about one another perhaps only because we have brain cells that try to help their neighbors.

Neural networks within organisms emerge when entities (cells) within the organism develop the ability to read or discern the state of other entities like themselves which they are in contact with, in communication with, with the "intention" or aim of helping their neighbors approach their more ideal state.

Neural networks function most effectively to the extent that each neuron, each member of the community of neurons, is "trying" to bring itself toward its more ideal state (that is, either a state of being active at a steady pace, OR a state of rest); while at the same time "trying" to help bring its neighbors toward their more ideal state (of steady activity OR rest).

For neurons that are molecular machines within biological organisms, the question of whether a neuron is at or near its ideal state depends on the levels of activity and patterns of connections among the various members of the community. These patterns of connection are exceedingly variable, since each synapse tip can grow or shrink slightly to form or break a connection with a neighbor. Each neuron has about 10,000 synapses. Ten thousand neighbors that it chooses to be or to not be in communication with.

Particular patterns of connectedness can result in a particular neuron community's having most of its members settled firmly into either a resting or a steadily active state (the ideal); or, conversely, with a different pattern of connectedness, most of the members of a community could be in an in-between, somewhat active state (a less-efficient state).

When in the less-efficient state, each member will try to adjust its connections with its neighbors so that moderately active elements will become fully active, while moderately quiet members will become more completely quiet. They do this by forming connections with some neighbors while breaking connections with others.

Each member of the community seeks to adjust its connections with its neighbors so as to approach a state of steady activity, OR (if it is near a resting state) to approach a state of being fully at rest. But it makes these adjustments while also "discerning" and responding to what changes would most aid its neighbors in their movement toward either a state of activity or rest.

The golden rule among members of a neuron community is to increase signaling to neighbors that appear to be tending toward more activity, while decreasing signaling to those who are inclined to become more quiet. In other words, the golden rule is to help your neighbors reach their more ideal state.

Without this concern for the 'other', any adjustments that a neuron might make in seeking a pattern of connectedness that results in a more nearly ideal state for itself would very likely frustrate the attempts of neighboring neurons to reach their more ideal state.

(Within each neuron there are countless microtubules that are able to affect one another's states, which are defined by how many electrons or ions are trapped in them, if any. When these microtubules lengthen or shorten or bend slightly, they change which of their neighbors they are able to communicate with.)

If we think of a society of human beings as a neural network, then we will see that we are more likely to bring ourselves AND the larger community (one another) toward our ideal state when we are trying to help others and respect the golden rule while also trying to take care of ourselves.




Natural Law Requires Respect of Public Property Rights, Too


Biodiversityt as a public good

No comments: