Percorrer por autor "Albino, Pedro Miguel Santos"
A mostrar 1 - 1 de 1
Resultados por página
Opções de ordenação
- Integrating GPT-based language models into NPCs for realistic dialogue and contextual awareness in unity gamesPublication . Albino, Pedro Miguel Santos; Rodrigues, PedroThis thesis explores the integration of large language models (LLMs), specifically the LLama family, to enhance the realism and interactivity of non-player characters (NPC) dialogue in video games developed with Unity. Traditional NPC interactions are often constrained by scripted dialogue trees, which limit immersion and adaptability. The main objective of this work is to demonstrate how GPT-like models can be embedded into game environments to generate dynamic, context-aware responses that elevate player engagement. The methodology involves using open-source LLMs to suit game-specific dialogue styles, then deploying these models within Unity using a local server to maintain performance. Custom environments were created in Blender to build an 3D prototype showcasing the real-time NPC conversations. Key implementation challenges, such as latency, prompt management, and dialogue coherence, were addressed through model tuning techniques. The results indicate that integrating LLMs into Unity significantly enhances dialogue variety, responsiveness, and player immersion compared to traditional systems. Furthermore, the flexible nature of LLMs opens new possibilities for adaptive storytelling and personalized gameplay.