Aprendizaje por demostración para la creación de personajes automáticos con estilo de juego humano en videojuegos arcade
Loading...
Download
Official URL
Full text at PDC
Publication date
2024
Defense date
27/06/2023
Authors
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Universidad Complutense de Madrid
Citation
Abstract
En los últimos años la industria del ocio electrónico ha vivido granes cambios motivados por los avances tecnológicos y por las expectativas que los jugadores tienen de los videojuegos que consumen. Por un lado, se ha vivido una gran democratización en el desarrollo, y por otro lado, las expectativas de los consumidores han evolucionado desde un fuerte enfoque en el aspecto visual hacia otros apartados como la implementación de mecánicas de juego más novedosas o el desarrollo de personajes no controlados por el jugador con comportamientos más creíbles. Tradicionalmente, el modelado de comportamientos para estos personajes se realiza con modelos computacionales como máquinas de estados o árboles de comportamiento, y el desarrollo de comportamientos complejos sigue siendo un proceso muy costoso que suele requerir de un gran conocimiento técnico y de largas sesiones de prueba y error. En este contexto, no es de extrañar cómo el aprendizaje por imitación, un paradigma de la IA para el modelado de jugadores cuya idea principal es la de aprender comportamientos en base a la imitación de otros jugadores, principalmente humanos, ha gozado de gran popularidad en los últimos años. La idea principal es la creación de agentes capaces de desplegar un comportamiento complejo asociado a jugadores humanos y no únicamente que sean capaces de jugar bien...
In the recent years, the video game industry has experienced great changes driven by technological advances and the expectations of the players regarding the video games they choose to play. On one hand, there has been a process of democratisation in the context of video games development, and on the other hand, players expectations have evolved from the visual complexity towards otrher aspects like the implementation of modern gameplay mechanics or the develpment of non-player characters (NPCs) with more believable behaviours.Traditionally, the creation of behaviours for NPCs have been done with computational models like finite state machines of bejaviour trees, and the development of complex behaviours is still considered a complex and error prone porcess which usually requires high level of technical knowledge and many hours of trial and error. In this context, it is hardly surprising high learning by imitation, an AI paradigm for player modelling which main idea is to learn behaviours by imitating other players, mainly human, has received great attention in the recent years. The main idea is the develpment of agents capable of deploying complex behaviours, usually associated to human players, and not only to be capable of playing of playing well...
In the recent years, the video game industry has experienced great changes driven by technological advances and the expectations of the players regarding the video games they choose to play. On one hand, there has been a process of democratisation in the context of video games development, and on the other hand, players expectations have evolved from the visual complexity towards otrher aspects like the implementation of modern gameplay mechanics or the develpment of non-player characters (NPCs) with more believable behaviours.Traditionally, the creation of behaviours for NPCs have been done with computational models like finite state machines of bejaviour trees, and the development of complex behaviours is still considered a complex and error prone porcess which usually requires high level of technical knowledge and many hours of trial and error. In this context, it is hardly surprising high learning by imitation, an AI paradigm for player modelling which main idea is to learn behaviours by imitating other players, mainly human, has received great attention in the recent years. The main idea is the develpment of agents capable of deploying complex behaviours, usually associated to human players, and not only to be capable of playing of playing well...
Description
Tesis inédita de la Universidad Complutense de Madrid, Facultad de Informática, Departamento de Ingeniería del Software e Inteligencia Artificial, leída el 27-06-2023.