Human intellect, by and large, is independent and, therefore, decentralized;

Intelligence is the strategic use of relatively small-world information, with variety and velocity, on specific subjects, as opposed to “whole of internet” to start with, that are:

Combinatorial (Combinatorial information would aggregate information over many variables, by allowing bets on all variable value combinations. Reminds Seinfeld episodes) – “Do you want to know about A? …But I was talking about B”;

Asymmetric (or imperfect) – “I don’t know, what you know and you don’t know, what I know”; and

Incomplete information – “I can’t recall” and “I don’t know, never heard”;

A self-organized learning method, in accordance with the dynamic proactive-retroactive learning method is executed to segment a graph network data based on bounded diffusion of collective individual information interactions. The nearest-neighbor or group data is determined from grouping of individual transactional data for a group where individual is a member. After a certain upper-bound number of groups, the system applies a diffusion-limited aggregation (“DLA”) – a formation process whereby individuals in a group, as particles, and their signals – defined as change or the first derivative in an individual’s data – of a subject matter undergo a stochastic process for clustering together to different aggregates (“clusters”) of such individuals. These signals and their changes – defined as the second derivative in an individual’s data – are used for predicting the group’s current state, as described above, and applied sheafing method, (or group theory) for “grouping” mechanism – depending on the geometry of the growth, for example, whether it be from a single point radially outward or from a plane or line – of clusters where the individual is a member, to determine the state.

The bounded diffusion in DLA, for example, may have one additional parameter, the position of the decision bound, say A. If at time t of the state data of the individual (or subject matter e.g., search for an item) is x, the distribution of the state at a future time may be s > t, hence the term “forward” diffusion. The backward diffusion, on the other hand, may be useful when the individual at a future time s has a particular behavior due to past decision, the distribution at time is t < s. This may impose a terminal condition, which is integrated backward in time, from s to t (hence the term “backward” is associated with this).