Do Wide and Deep Networks Learn the Same Things? Uncovering How Neural Network Representations Vary with Width and Depth.
Thao NguyenMaithra RaghuSimon KornblithPublished in: ICLR (2021)
Keyphrases
- neural network
- network structure
- wide range
- recurrent networks
- artificial neural networks
- network model
- genetic algorithm
- learning rules
- pattern recognition
- complex networks
- self organizing maps
- back propagation
- data sets
- social networks
- fault diagnosis
- spiking neurons
- computer networks
- neural nets
- neural network ensemble
- function approximators
- neural network is trained
- distributed representations
- topological features
- network size
- deep learning
- activation function
- network analysis
- radial basis function
- higher level
- fuzzy logic
- probabilistic model