J. Audiffren (CMLA): Equivalence between regularized learning algorithms




Date(s) : 03/10/2014   iCal
14 h 00 min - 15 h 00 min

Equivalence between regularized learning algorithms by J. Audiffren (CMLA)\n\nThis talk is about the concept of equivalence equivalence between machine learning algorithms\, and more particularly equivalence between algorithms defined with a minimization problem. We will discuss the notion of equivalence\, its definition and its consequences. We will introduce two notions of algorithmic equivalence\, namely\, weak and strong equivalence\, and we will study their strengths and their limit.

Catégories Pas de Catégories



Retour en haut