PHD Code All This git contains all the PHD code for Reluctant learning algorithm. This was submitted to UWS in July 2019

Code contains

ImageClassification_CIFAR10_ReluctantLearning Python code using reluctant learning algorithm to perform classification on CIFAR 10 dataset requires python, keras, tensorflow to execute

imageClassificatio_CIFAR10_Standard Python standard example of training for CIFAR 10 dataset (Control for above code)

Iris c#/iris-classification C# Iris training dataset using both standard backpropagation algorithm and reinforcement learning algorithm adaptation (gui controls input and data results are plotted on x,y output

iris Python Python Tensorflow, keras implementation of c# iris dataset example code contains one standard approach and one reinforcement learning approach

Racing_XNA_source Code/Figureof8Race this is a XNA c# code, required XNA plugin. this code was originally written in 2012 as an interesting look at games and neural networking. building a neural network without a framework from scratch to understand how they are built and work. code is more difficult to run now as XNA does not run on modern version of visual studio. XNA code was used then as it enabled this code (version of this code) to run on a Xbox360. This was useful at the time to as run on a standard PC was very long and the xbox360 enabled faster calculations.

prisoner This is a simple example of using a neural network to run the Prisoner's Dilemma. it can be compared against standard neural network updates, reinforcement learning algorithm, tit for tat and random. all can be done by slight adaptations to the code.

the code generates a number of games, outputting training data to csv and then running some code to aggregate the results.

Chris Jones chris.jones@mavoco.com