Principles of deep learning theory, the: an effective theory approach to understanding neural networks
Auteur :
Roberts, Daniel A. / Yaida, Sho
Éditeur :
Cambridge University Press
ISBN :
9781316519332
Date de publication :
26 mai 2022
Dimensions :
26,1 x 18,4 x 2,6 cm
Poids :
1060 g
Langue :
Anglais
Pays d'origine :
Grande Bretagne
This is the first book focused entirely on deep learning theory. Tools from theoretical physics are borrowed and adapted to explain, from first principles, how realistic deep neural networks work, benefiting practitioners looking to build better AI models and theorists looking for a unifying framework for understanding intelligence.