6. Using AI in Computer Games (Learning Techniques)

NWC603COM – Using AI in Computer Games                                                        Rachel Osho                                                                                                                                              ID: 309740
Assignment 1 – Theory One

Learning Techniques

Q - Learning

Q-Learning is a method that was introduced in 1989 by Watkins. (Wikipedia, n.d)  Q learning which is a reinforcement learning algorithm, this the algorithm was used via Google in order to beat humans whilst playing atari games. As a reinforcment learing algorithm this is one of the types machine learning, this algorithm learns to react to an environment through positive feedback. The term ‘reinforcement learning algorithms’ derives from a set of learning algorithms that is stimulated through the use of behavioural physycology. This is the idea that the algorithm is instructed to carry out specific actions due to previous experiences by means of either rewarding or punishing actions, this can be thought to be similar to a teacher praising a child for he/she’s good behaviour (D, 2017)

Artificial Neural Networks

Artificial neural networks is a learning technique which was biologically inspired. Over the years as many developers have worked with artificial neural networks for large scale behaviour control and have found a flaws in the learning technique as they found weaknesses and disadvantages within the approach. On the other hand this approach has left individuals that are artificial intelligence enthusiasts lost and confused, as they can’t seem to comprehend why exactly they are not being used across and all over within the industry. Where as developers do not share the same viewpoint on ANN’s* as they are often disappointed in regards to the approach’s weaknesses as they view the approach as useless.