I'm Dino Ratcliffe. A PhD student researching Artificial Intelligence at the University of Essex
Started a Msci in computer science at the University of Essex in 2011. During the next 4 years focused on modules that involved improving technical skills and Artificial Intelligence. Winner of the K.F Bowden Memorial prize in two separate years. Worked at the London startup Signal Media during the summer of 2014 and continued to work for them part time during my masters year. Graduated with a 1st class degree.
Teaching AI agents transferable skills for game playing
My research focuses on the ability of an AI agent to be able to evaluate the various skills it would need to master a game, such as in an FPS (first person shooter) like doom. If the agent can learn to cluster actions that may split into strategies such as attacking enemies, gathering ammo/health and avoiding enemy fire this information could then be used in similar games. This information would also provide a base for being to evaluate players on a skill level, giving a much more granular view of their strengths and weaknesses in any of these games. This could then be used for better matchmaking in team games, placing players into teams whose skill sets complement each other. Other applications include being able to guide the player into situations that give them more experience in the areas they are weakest.
Competent in many programming languages including functional languages such as Clojure and object oriented languages such as C++. Experience implementing neural networks in C++ and using neural network frameworks such as TensorFlow. Light experiance in using Unity and Unreal Engine 4.