10 Days of Deep Learning

I took a challenge upon myself to fast pace my Deep Learning skills using the python data stack. I will take on a real life practical challenge everyday and try to solve it using Deep Learning (Starting from the basics)

Day 1

Text classification using Multi Layered Perceptron and comparing with NN's

TODO:

1. Use learning rate decay
2. Batch normalization
3. More number of epochs

Day 2

Text classification using TFIDF feature matrix and MLP

TODO:

1. Use LRD, Batch Normalisation
2. Use Dropout
3. Further process the TFIDF matrix using TruncatedSVD

Day 3

Text classification using word2vec and MLP

TODO:

1. Use LRD, Batch Normalisation, Dropout
2. Tweaking word2vec hyperparams

Day 4

Text classification using Convolutional Neural Networks

TODO:

1. Use LRD, Batch Normalisation, Dropout
2. Try a multi-layer Convolutional Network

Day 5

Text classification using Recurrent Neural Networks

TODO:

1. Use LRD, Batch Normalisation, Dropout
2. Use a combination of CNN and RNN