Simple-NN

This project is the Neural Network course project in which I tried to implement a Neural Network from scratch and try to solve some image problems with it. Then, in the next assignment, I have tried to solve the Fashion-MNIST and the CIFAR-10 challenges using CNN and Keras library.

The report of the codes are available for download as pdf at here.

Assignment #1 (MLP)

• Day-1: Implementing neural network with backpropagation and try to solve XOR problem.
• Day-2: Clean the code and make it more general in terms of using with different functions.
• Day-3: Try to solve MNIST challenge with this network. As a result, some additional functionality are also added to the network including batch normalization and dropout layers. [This attempt was not very successfull]

Assignment #2 (CNN)

• Day-5: Solve Fashion-MNIST challenge. Implement both an MLP and a CNN to solve this problem and compare the results.
• Day-6: Solve CIFAR-10 with CNN implemented in Keras.

Assignment #3 (RNN)

• Day-7: Solve text sentiment detection using RNN implementation in Keras.