Ural School of Pattern Recognition: Majoritarian Approach to Ensemble Learning

Abstract

This article provides an overview of the significant achievements of the Ural School of Pattern Recognition. The focus is on majoritarian generalized solutions for algebraic equations and inequalities that may not always adhere to standard properties. The paper also delves into the broader applications of these findings in collective machine learning techniques. In the literature, these generalized solutions are frequently referred to as committee generalized solutions or simply committees, leading to the derived learning methods being called committee machines. Our discussion primarily centers on the foundational theorems confirming the existence of such solutions, the intricacies of combinatorial optimization during their exploration, and the subsequent emergence of collective machine learning algorithms.