Creating An AI That Can Play a Tactical Intelligence Game Alongside Human Players

Robert Harris

Supervised by Alun D Preece, Moderated by Federico Cerutti

The aim of this project is to determine whether a virtual system can be a viable asset in contributing to crowd sourced knowledge bases. The project aims to answer this by implementing a system that can play the crowd sourced knowledge game SHERLOCK. Using image recognition, data reasoning and communication it is hoped the system can act as ???one of the crowd??? and contribute at the same level a human player can when playing the game. The project will use a mix of technologies such as; controlled natural languages, CENode and image recognition. The design and implementation will aim to have the attributes commonly found in a cognitive architecture. The primary deliverable will see the AI player's contribution interact with the SHERLOCK dashboard alongside a human player's contribution. This would signify that a human and virtual player can work together to produce crowdsourced knowledge.

Initial Plan

Final Report

Publication Form