Can Bayesian Networks provide answers when Machine Learning comes up short? It’s a question of probabilities
April 1, 2021
Judea Pearl is the author of The Book of Why: The New Science of Cause and Effect lead Causal Revolution. He is also the recipient of the 2020 World Leader in AI World Society and leads the section Modern Causal Inference of AIWS.net.
It is common sense that modern science and technology and the Industrial Revolutions, involving technological, socioeconomic, political, psychological, and cultural changes, are causally interconnected as complex entangled heterogeneous dynamic causal networks.
We have a series of scientific, technological, cultural, and industrial revolutions, while ignoring the causal revolution in our mentality, sciences, technologies and industries.
Our very existence depends on causality as a master principle and universal law, but we still debate if causality exists, dividing in various groups, believers or non-believers, agnostics and gnostics, realists, conceptualists and nominalists, theists and atheists, etc.
In all, causality and its causation has not advanced much since Aristotle’s ideas of four causes, especially in terms of complexity and nonlinearity. What we have now, is more formalized and much narrower simplistic statistical linear causal models, as presented in Judea Pearl and Dana Mackenzie’s The Book of Why. The New Science of Cause and Effect
Unlike the simple user-friendly artificial models, real causality is hyper-complex, interactive, productive, determinate and stochastic, nonlinear and multi-causal, emergent and omni-directional, top-down and bottom-up, reversed, inverse, inverted, reciprocal, reflexive and symmetrical.
Causality has the absolute priority of ontological existence, even recognized by religion, the Creator is the Great First Cause of all things, but we collectively pretend to understand it devising all sorts of simplified naive linear causal models.
Causality involves the most critical categories, as all infinite universes or world or reality with its thing, entity, substance, state, change, relation, space, time, or agents, causes, processes, effects, and forces, that together embrace everything existing and predictable.
The main features involved in the Causal Revolution are disruptive scientific and technological changes, as in:
A comprehensive, consistent and coherent causal model of the world for humans and computers;
The Master Algorithm of Reality as Descriptive, Deductive, Intuitive, Inductive, Exploratory, Explainable, Predictive and Prescriptive (DDIIEEPP) Platform;
Real AI, Causal Machine Intelligence and Learning, as The Next Big Thing in Technology: Causal/Real AI (RAI) as Leibniz’s Superscientist or Laplace’s Demon