A Theoretical Study Of Causality In Machine Learning

Curiosity is human nature.From the beginning of human conscious historical activities,human exploration of everything in the world has not stopped for a moment.Aristotle,an ancient Greek philosopher,also said that "philosophy originated from surprise",while philosophy is the wisdom generated by the study of everything in nature.Each of us has been full of curiosity about things around us since we were young.We always ask "why" about everything we have experienced.This ability to ask "why" just reflects human being's unique ability to think about causation.David Hume,a modern empiric philosopher,made the first in-depth analysis of the scientific meaning of causation.It is also because Hume's argument that causation is attributed to a psychological customary association.Since then,there has been a lack of uniform understanding of the definition of causation in academia,and various logical attempts to explain causation have encountered certain difficulties,even the metaphysical concept that causation is regarded by some philosophers as supposed to be eliminated.However,in recent years,causation has gradually become a hot topic in the field of machine learning.In view of this,the author traces and examines the new development of contemporary causation theory in machine learning in this paper.The first chapter of this article first introduces and discusses the relevant concepts and theoretical development of machine learning,as well as the internal relationship between machine learning development and causation research.The second chapter focuses on the theory of causation in machine learning,and analyzes the three basic models of "seeing","intervention" and "counterfactual" in causation reasoning.The third chapter summarizes and evaluates the theoretical thought of causation in machine learning,and conceives the possibility of future development of causation reasoning in machine learning.By summarizing the current causation thought in the field of machine learning,we hope it can help deepen our understanding of the development of contemporary causation theory.