Economics of Neural Networks

Any “general rule of arbitrary precision” must include a limit (time delineation) in order to categorize and test an outcome(consequence), since we may categorize consequences at any point in the time line in which actionable or deducible constant relations are identifiable. In other words, searches for prediction of futures are change (state) dependent.

This may be heavy but it means that your prediction of future events from any state may vary by the utility you prefer.

We must operate by general rules (categories) because that is all we can act upon (a concentration of constant relations during which we can effect a change in state.)

We all bias our utility (judgements) on similar timelines if not only due to ability, but also on commensurability. Ergo, we develop out of necessity time preferences and the more expertise we develop in any time frame the more related (dependent) associations we develop in concert.

This isn’t just choice it’s the economics of neural networks, and that economics is no different from the ‘economics’ of physics, biology, and sentience.

(for Andy Curzon)
Apr 18, 2018 9:59am

Leave a Reply