Following the tutorials from aymericdamien's Tensorflow Examples

Reading material from Neural Networks and deep Learning

Implement these with a basic knowledge of Data Analytics and Machine Learning

1. Linear regression

(tutorial)

2. Logistic Regression

(tutorial)

3. kNN

(tutorial)

4. Random Forests (Needs a fix)

(tutorial)

5. CNN

(tutorial)

Also read the awesome tutorial on backpropogation from the same book (Chapter 2) and why it works so well for neural nets

6. CNN with Abstraction using Tensorflow (Needs a fix) (tutorial)

Read what happens when the network has a slow learning rate due to the L2 cost function from Chapter 3

1. Cross Entropy
2. SoftMax

Read about Regularization and why it works. Simple is better, but not neccesarily.

3. Over Fitting

(textbook)

4. No Free Lunches

L1 Regularization - Makes the network smaller with lesser number of connections. L2 Regularization - Makes sure the weights are not too big. Dropout - Works similar to averaging multiple nets.

5. Data Augmentation: Makes the network more susceptible to changes

7. Convolution Neural Networks (textbook chapter 4)

8. Cats vs Dogs - Here's an example using a vanilla CNN

9. Example code for using tensorboard

10. RNN(LSTM)

The timesteps are added to the LSTM cell in the RNN example. For more info, read this blog on Understanding LSTMs.

11. Saving and Restoring a Model Here's an excellent blog post on saving and restoring models

Special Thanks to Naresh for his awesome backpropogation for my errors!