Err... interesting, maybe -- useful? I don't see why. There are a near infinite number of function approximators to choose from, and theory-building science ha

Err... interesting, maybe -- useful? I don't see why. There are a near infinite number of function approximators to choose from, and theory-building science has little use for any. Empirical, or heuristic, science already uses much faster and arbitrarily accurate ones.