Comparing and Combining Curriculum Learning and Behaviour Cloning to train First person Shooter Agents
Autor(a) principal: | |
---|---|
Data de Publicação: | 2024 |
Tipo de documento: | Dissertação |
Idioma: | eng |
Título da fonte: | Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
Texto Completo: | http://hdl.handle.net/11110/2800 |
Resumo: | Artificial Intelligence bots are extensively used in multiplayer First-Person Shooter games. By using Machine Learning techniques, we can improve their performance and bring them to human skill levels. In this work, we focused on comparing and combining two Reinforcement Learning training architectures, Curriculum Learning and Behaviour Cloning, applied to a First-Person Shooter developed in the Unity Engine. We have created four team of three agents each: one team for Curriculum Learning, one for Behaviour Cloning, and another two for two different methods of combining Curriculum Learning and Behaviour Cloning. After completing the training, each agent was matched to battle against each agent of the other teams until each pairing had five wins or ten timeouts. In the end, results showed us that the agents trained with Curriculum Learning achieve better performance than the ones trained with Behaviour Cloning by a matter of 23.67% more average victories in one case. In terms of the combination attempts, not only did the agents trained with both devised methods have problems during training, but they also achieved insufficient results in battle, with an average of 0 wins. |
id |
RCAP_7f7f3d44a9bf1c95f89a30a585b8707a |
---|---|
oai_identifier_str |
oai:ciencipca.ipca.pt:11110/2800 |
network_acronym_str |
RCAP |
network_name_str |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
repository_id_str |
7160 |
spelling |
Comparing and Combining Curriculum Learning and Behaviour Cloning to train First person Shooter AgentsReinforcement LearningUnityFirst-Person Shooter GamesCurriculum LearningBehaviour CloningArtificial Intelligence bots are extensively used in multiplayer First-Person Shooter games. By using Machine Learning techniques, we can improve their performance and bring them to human skill levels. In this work, we focused on comparing and combining two Reinforcement Learning training architectures, Curriculum Learning and Behaviour Cloning, applied to a First-Person Shooter developed in the Unity Engine. We have created four team of three agents each: one team for Curriculum Learning, one for Behaviour Cloning, and another two for two different methods of combining Curriculum Learning and Behaviour Cloning. After completing the training, each agent was matched to battle against each agent of the other teams until each pairing had five wins or ten timeouts. In the end, results showed us that the agents trained with Curriculum Learning achieve better performance than the ones trained with Behaviour Cloning by a matter of 23.67% more average victories in one case. In terms of the combination attempts, not only did the agents trained with both devised methods have problems during training, but they also achieved insufficient results in battle, with an average of 0 wins.2024-01-26T17:39:22Z2024-01-262024-01-26T00:00:00Zinfo:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/masterThesishttp://hdl.handle.net/11110/2800http://hdl.handle.net/11110/2800TID:203486048engAlmeida de Neves Marta, Pedroinfo:eu-repo/semantics/openAccessreponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãoinstacron:RCAAP2024-02-01T07:33:36Zoai:ciencipca.ipca.pt:11110/2800Portal AgregadorONGhttps://www.rcaap.pt/oai/openaireopendoar:71602024-03-20T01:59:22.948577Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãofalse |
dc.title.none.fl_str_mv |
Comparing and Combining Curriculum Learning and Behaviour Cloning to train First person Shooter Agents |
title |
Comparing and Combining Curriculum Learning and Behaviour Cloning to train First person Shooter Agents |
spellingShingle |
Comparing and Combining Curriculum Learning and Behaviour Cloning to train First person Shooter Agents Almeida de Neves Marta, Pedro Reinforcement Learning Unity First-Person Shooter Games Curriculum Learning Behaviour Cloning |
title_short |
Comparing and Combining Curriculum Learning and Behaviour Cloning to train First person Shooter Agents |
title_full |
Comparing and Combining Curriculum Learning and Behaviour Cloning to train First person Shooter Agents |
title_fullStr |
Comparing and Combining Curriculum Learning and Behaviour Cloning to train First person Shooter Agents |
title_full_unstemmed |
Comparing and Combining Curriculum Learning and Behaviour Cloning to train First person Shooter Agents |
title_sort |
Comparing and Combining Curriculum Learning and Behaviour Cloning to train First person Shooter Agents |
author |
Almeida de Neves Marta, Pedro |
author_facet |
Almeida de Neves Marta, Pedro |
author_role |
author |
dc.contributor.author.fl_str_mv |
Almeida de Neves Marta, Pedro |
dc.subject.por.fl_str_mv |
Reinforcement Learning Unity First-Person Shooter Games Curriculum Learning Behaviour Cloning |
topic |
Reinforcement Learning Unity First-Person Shooter Games Curriculum Learning Behaviour Cloning |
description |
Artificial Intelligence bots are extensively used in multiplayer First-Person Shooter games. By using Machine Learning techniques, we can improve their performance and bring them to human skill levels. In this work, we focused on comparing and combining two Reinforcement Learning training architectures, Curriculum Learning and Behaviour Cloning, applied to a First-Person Shooter developed in the Unity Engine. We have created four team of three agents each: one team for Curriculum Learning, one for Behaviour Cloning, and another two for two different methods of combining Curriculum Learning and Behaviour Cloning. After completing the training, each agent was matched to battle against each agent of the other teams until each pairing had five wins or ten timeouts. In the end, results showed us that the agents trained with Curriculum Learning achieve better performance than the ones trained with Behaviour Cloning by a matter of 23.67% more average victories in one case. In terms of the combination attempts, not only did the agents trained with both devised methods have problems during training, but they also achieved insufficient results in battle, with an average of 0 wins. |
publishDate |
2024 |
dc.date.none.fl_str_mv |
2024-01-26T17:39:22Z 2024-01-26 2024-01-26T00:00:00Z |
dc.type.status.fl_str_mv |
info:eu-repo/semantics/publishedVersion |
dc.type.driver.fl_str_mv |
info:eu-repo/semantics/masterThesis |
format |
masterThesis |
status_str |
publishedVersion |
dc.identifier.uri.fl_str_mv |
http://hdl.handle.net/11110/2800 http://hdl.handle.net/11110/2800 TID:203486048 |
url |
http://hdl.handle.net/11110/2800 |
identifier_str_mv |
TID:203486048 |
dc.language.iso.fl_str_mv |
eng |
language |
eng |
dc.rights.driver.fl_str_mv |
info:eu-repo/semantics/openAccess |
eu_rights_str_mv |
openAccess |
dc.source.none.fl_str_mv |
reponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação instacron:RCAAP |
instname_str |
Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação |
instacron_str |
RCAAP |
institution |
RCAAP |
reponame_str |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
collection |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) |
repository.name.fl_str_mv |
Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação |
repository.mail.fl_str_mv |
|
_version_ |
1799137077622734848 |