Final Year Project

Project Title

Adaptive AI for Interactive Environments

Description

Built in Unity using ML-Agents, this project replaces traditional state-machine logic with a trained neural network to create a dynamic adversarial entity. The goal is to engineer an intelligent system that continuously analyzes user behavior to dynamically adjust its strategies and difficulty in real time. The research demonstrates the viability of applying reinforcement learning within real-time games or other interactive environments and to practically evaluate both the benefits and the technical limitations of this approach.