Our Papers

Engaging the Player

Aneurin Hodgson-Roch

This paper will be exploring the methods used to engage the player and why they are effective. It will also look at three key examples of a game with a good player engagement. These players shall be broken down to the components that make them engaging. Afterwards, the methods off engagement will be view for how they could apply to AIs Defense.

Artificial Intelligence implementations for Games

Dylan McAdam

The purpose of this research is to explore the various implementations of Artificial Intelligence (AI) used in games to create realistic virtual worlds and environments. The focus of our research will be on methods used in AAA titles, as it is typically the larger, more prominent development companies that are the ‘market leaders’ in cutting-edge AI for games. We will discuss an area of research being conducted by Ubisoft to quickly generate landscapes and cities using machine learning techniques, including their partnership with the Universities of Lyon and Purdue. We also discuss a bespoke system, created by Monolith Productions, called the ‘Nemesis System’, which they use to create completely unique enemies with their own stats, traits, strengths and weaknesses.

Neural Networks and Machine Learning used to train Artificial Intelligence systems for Virtual World and Terrain Generation

Dylan McAdam

The purpose of this research is to study and analyse the technologies and methodologies used to train Artificial Intelligence (AI) systems to generate realistic and unique virtual worlds or terrains. We specifically look at the use of Neural Networks and Machine Learning techniques to train the AI systems, as well as the methodologies involved in implementing them for certain applications. Our aim is to analyse and identify to most appropriate methods and techniques available to train AI-based systems for specific purposes that will save in overall project development time and costs.