Investigating the Behaviour of Deep Neural Networks for Classification

Phil Kim's 'Deep Learning with Matlab' book in Python

A FEED-FORWARD DEEP NEURAL NETWORK with:

• Batch Normalization
• Activation Functions
• Delta Error Rule
• Backward Error Propagation
• Weight Update Optimization Techniques
• Drop-Outs

Some Results:

Experiment 1:

Experiment 2:

Experiment 3:

Experiment 4:

Project Hypothesis

Email: [email protected]

Somefun, O.A.

(c) 2017