A correspondence between normalization strategies in artificial and biological neural networks

A correspondence between normalization strategies in artificial and biological neural networks

A fundamental challenge at the interface of machine learning and neuroscience is to uncover computational principles that are shared between artificial and biological neural networks. In deep learning, normalization methods such as batch normalization, weight normalization, and their many variants h...

Full description

Bibliographic Details
Main Authors: Navlakha, S. (Author), Shen, Y. (Author), Wang, J. (Author)
Format: Article
Language:English
Published: MIT Press Journals 2021
Subjects:
algorithm
Algorithms
Biological neural networks
brain
Brain
Brain functions
Deep learning
Economic and social effects
Hidden units
machine learning
Machine Learning
nerve cell
Network activities
Network training
Neural networks
Neural Networks, Computer
Neurons
Normalisation
Normalization methods
Normalization strategies
Scale normalization
Spatial scale
Online Access:View Fulltext in Publisher

Internet

View Fulltext in Publisher

Similar Items