Crossing Levels: Meta-induction and the Problem of Induction.

Gerhard Schurz claims to have a solution to Hume's problem of induction based on results from machine learning concerning meta-induction. His argument has two steps. The first is to establish a justification for following a certain meta-inductive strategy based on its predictive optimality. The seco...

Descripción completa

Detalles Bibliográficos
Publicado en:Philosophy of Science Vol. 90; no. 4; pp. 985 - 994
Autor principal: Henderson, Leah
Formato: Artículo
Publicado: Cambridge University Press Oct2023
Materias:
Acceso en línea:Ver este registro en EBSCOhost
Descripción
Sumario:Gerhard Schurz claims to have a solution to Hume's problem of induction based on results from machine learning concerning meta-induction. His argument has two steps. The first is to establish a justification for following a certain meta-inductive strategy based on its predictive optimality. The second step is to show how this justification can be transferred to object-induction. I unpack the second step and fail to find a convincing argument supporting the transfer of justification from meta-induction to object-induction. My conclusion is that the problem of induction has not yet been solved by appeal to meta-induction.