Gradient

Mechanism

The directional rate of change of a scalar field across a differentiable space. For any sufficiently smooth field f over a differentiable space, the gradient ∇f at a point is the vector whose direction is the one of steepest ascent of f and whose magnitude is the rate of increase in that direction. Substrate property: gradients exist wherever fields exist, whether or not anyone measures them. Specific cases include Entropy gradients (the direction of maximum disorder increase), attention gradients, credibility gradients, information-density gradients. Higher patterns use gradients for hill-climbing search, attention allocation, credit assignment, and flow routing.