Chris Pollett > Students

Simple multi layer AI using Keras (Built on top of TensorFlow and THEANO)

Description:

In order to be able to play the game, the machine needs to control it. Deliverable 2 comprises of a program in python that controls the keyboard and mouse functions of the laptop. This would further help in controlling the game and then take screenshots of it.

About the program: (cnn.py)

The first hidden layer is a convolutional layer - Convolution2D. It haa 30 feature maps, with the size of 5 by 5 and a rectifier activation function. This is the input layer, expecting images with the structure outline above [pixels][width][height].
Next is a pooling layer that takes the max called MaxPooling2D. ( configured with a pool size of 2 by 2.)
The third hidden layer is a convolutional layer - Convolution2D. It haa 15 feature maps, with the size of 3 by 3.
Next is a pooling layer that takes the max called MaxPooling2D. ( configured with a pool size of 2 by 2.)
The next layer is a regularization layer using dropout called Dropout. It is configured to randomly exclude 20% of neurons in the layer in order to reduce overfitting.
Next is a layer that converts the 2D matrix data to a vector called Flatten. It allows the output to be processed by standard fully connected layers.
Next a fully connected layer with 128 neurons and rectifier activation function.
The layer after that is another fully connected layer with 50 neurons and rectifier activation function.
Finally, the output layer has 10 neurons for the 10 classes and a softmax activation function to output probability-like predictions for each class.

The first hidden layer is a convolutional layer - Convolution2D. It haa 30 feature maps, with the size of 5 by 5 and a rectifier activation function. This is the input layer, expecting images with the structure outline above [pixels][width][height].
Next is a pooling layer that takes the max called MaxPooling2D. ( configured with a pool size of 2 by 2.)
The third hidden layer is a convolutional layer - Convolution2D. It haa 15 feature maps, with the size of 3 by 3.
Next is a pooling layer that takes the max called MaxPooling2D. ( configured with a pool size of 2 by 2.)
The next layer is a regularization layer using dropout called Dropout. It is configured to randomly exclude 20% of neurons in the layer in order to reduce overfitting.
Next is a layer that converts the 2D matrix data to a vector called Flatten. It allows the output to be processed by standard fully connected layers.
Next a fully connected layer with 128 neurons and rectifier activation function.
The layer after that is another fully connected layer with 50 neurons and rectifier activation function.
Finally, the output layer has 10 neurons for the 10 classes and a softmax activation function to output probability-like predictions for each class.

Download the program to train on MNIST from here: cnn.py
Download the library THEANO from here OR (if using linux), on your Terminal, type: "pip install THEANO"
Download the library Tensorflow from here OR (if using linux), on your Terminal, type: "pip install tensorflow"
Download the library Keras from here OR (if using linux), on your Terminal, type: "pip install keras"
To run the program, on your terminal, type: "python cnn.py"
After the program runs successfully, an h5 file will be generated, called: "my_model.h5"
Once the model is generated
Download the program to predict the labels from here: cnn_test.py
To run the program, on your terminal, type: "python cnn_test.py"
You shall be able to see the labels of the predicted images!

Note: cnn.py might take a long time to run. To decrease the time, decrease the number of epochs.