Decoupling Gradient-Like Learning Rules from Representations.
Philip S. ThomasChristoph DannEmma BrunskillPublished in: ICML (2018)
Keyphrases
- learning rules
- biologically inspired
- neural network
- hybrid neural
- artificial neural networks
- input output
- learning scheme
- back propagation
- artificial intelligence
- learning algorithm
- hopfield neural network
- edge detection
- image gradient
- knowledge base
- image processing
- optical flow
- multiple representations
- learned rules
- data sets