Neural Networks can Learn Representations with Gradient Descent.

Alexandru Damian Jason D. Lee Mahdi Soltanolkotabi

Published in: COLT (2022)

Keyphrases

neural network
learning rules
pattern recognition
artificial neural networks
function approximators
back propagation

genetic algorithm
distributed representations
loss function
higher level
cost function
fuzzy logic

objective function
symbolic knowledge
support vector machine
self organizing maps
fuzzy systems

multiscale
efficient learning
cellular neural networks
image segmentation
decision trees