Материалы международной научной конференции «Уфимская осенняя математическая школа» (г. Уфа, 6-9 октября 2021 г.). Том 2 / отв. редактор З.Ю. Фазуллин. - Уфа: РИЦ БашГУ, 2021. - 272 с.
Applying probability theory to neural networks
Ашимов И.
Using mathematical models to create an image of the human brain reproduction of machines by artificial intelligence.
An artificial neural network consists of three components: input layer, hidden layers, output layer.
Also, in the center of the neural network is the loss function. Which minimizes errors known as:
quadratic, cross-entropy, AdaBoost, Kullback distance. Loss function should not depend on the activation values of the neural network.
Applying probability theory to neural networks
Using mathematical models to create an image of the human brain reproduction of machines by artificial intelligence.
An artificial neural network consists of three components: input layer, hidden layers, output layer.
Also, in the center of the neural network is the loss function. Which minimizes errors known as:
quadratic, cross-entropy, AdaBoost, Kullback distance. Loss function should not depend on the activation values of the neural network.