RUS  ENG
Full version
VIDEO LIBRARY

International Conference Dedicated to the 100th Anniversary of the Birthday of V. S. Vladimirov (Vladimirov-100)
January 13, 2023 15:30, Moscow, Steklov Mathematical Institute, room 430 (Gubkina 8) + Zoom


Learning by population genetics and matrix Riccati equation

S. V. Kozyrev

Steklov Mathematical Institute of Russian Academy of Sciences, Moscow



Abstract: A model of learning as generalization of the Eigen's quasispecies model in population genetics is introduced. Eigen's model is considered as a matrix Riccati equation, the error catastrophe in the Eigen's model (when the purifying selection becomes ineffective) is discussed as the divergence of Perron–Frobenius eigenvalue of the Riccati model in the limit of large matrices. Known estimate for Perron–Frobenius eigenvalue provides explanation for observed patterns of genomic evolution. We propose to consider the error catastrophe in Eigen's model as analogue of overfitting in learning theory, this gives a criterion for presence of overfitting in learning.

Language: English


© Steklov Math. Inst. of RAS, 2024