What is the main theoretical limitation of the backpropagation algorithm in deep neural networks ?
- the vanishing and exploding gradient problem in deep layers
- lack of GPU compability
- it’s inability to learn nonlinear relationships
- the inability to update weights with out direct supervisons