bdhammel / machine-learning-models

Tensorflow scripts of different machine learning algorithms

Basic machine learning models for reference

Algorithms benchmarked to the MNIST dataset

Principle Component Analysis (PCA)

Naive Bayes accuracy: 56%

Artificial Neural Network (ANN)

Simple fully-connected network

MNIST accuracy: 96%

Convolutional Neural Network (CNN)

Simple convolutional neural network

MNIST accuracy: 98%

Notes: Slow, only ran for 2 epoch.

Convolutional Neural Network (VGG16)

Implementation of the VGG16 architecture with pretrained MNIST data

cite: https://www.cs.toronto.edu/~frossard/post/vgg16/

Note: Because MNIST is 28x28, first 10 layers omitted and layer depth made a factor of 3 smaller (as MNIST has only 1 color layer)

MNIST accuracy:

Stacked Autoencoder (X-wing AE)

A symmetric autoencoder for unsupervised learning

Notes: Success is largely dependent on layer sizes, seems to have a small window of convergence

Convolutional Autoencoder (CAE)

A symmetric convolutional autoencoder

Variational Autoencoder (VAE)

Generative Adversarial Network (DCGAN)

A Deep Convolutional GAN

• "Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks”

Recurrent Neural Network (RNN)

TODO: Add in additional hidden layers

Basic RNN cell

MNIST accuracy: 98%
epochs: 100

LSTM cell

MNIST accuracy: 95%
epochs: 10