A VIEW ON DEEP REINFORCEMENT LEARNING IN IMPERFECT INFORMATION GAMES
Many real-world applications can be described as large-scale games of imperfect information. This kind of games is particularly harder than the deterministic one as the search space is even more sizeable. In this paper, I want to explore the power of reinforcement learning in such an environment; t...
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Babes-Bolyai University, Cluj-Napoca
2020-10-01
|
| Series: | Studia Universitatis Babes-Bolyai: Series Informatica |
| Subjects: | |
| Online Access: | http://193.231.18.162/index.php/subbinformatica/article/view/3888 |
| _version_ | 1827356526823604224 |
|---|---|
| author | Tidor-Vlad PRICOPE |
| author_facet | Tidor-Vlad PRICOPE |
| author_sort | Tidor-Vlad PRICOPE |
| collection | DOAJ |
| description |
Many real-world applications can be described as large-scale games of imperfect information. This kind of games is particularly harder than the deterministic one as the search space is even more sizeable. In this paper, I want to explore the power of reinforcement learning in such an environment; that is why I take a look at one of the most popular game of such type, no limit Texas Hold’em Poker, yet unsolved, developing multiple agents with different learning paradigms and techniques and then comparing their respective performances. When applied to no-limit Hold’em Poker, deep reinforcement learning agents clearly outperform agents with a more traditional approach. Moreover, if these last agents rival a human beginner level of play, the ones based on reinforcement learning compare to an amateur human player. The main algorithm uses Fictitious Play in combination with ANNs and some handcrafted metrics. We also applied the main algorithm to another game of imperfect information, less complex than Poker, in order to show the scalability of this solution and the increase in performance when put neck in neck with established classical approaches from the reinforcement learning literature.
|
| first_indexed | 2024-03-08T05:11:01Z |
| format | Article |
| id | doaj.art-4b97a325d39940c78f139b1a0617e43d |
| institution | Directory Open Access Journal |
| issn | 2065-9601 |
| language | English |
| last_indexed | 2024-03-08T05:11:01Z |
| publishDate | 2020-10-01 |
| publisher | Babes-Bolyai University, Cluj-Napoca |
| record_format | Article |
| series | Studia Universitatis Babes-Bolyai: Series Informatica |
| spelling | doaj.art-4b97a325d39940c78f139b1a0617e43d2024-02-07T10:03:36ZengBabes-Bolyai University, Cluj-NapocaStudia Universitatis Babes-Bolyai: Series Informatica2065-96012020-10-0165210.24193/subbi.2020.2.03A VIEW ON DEEP REINFORCEMENT LEARNING IN IMPERFECT INFORMATION GAMESTidor-Vlad PRICOPE0University of Edinburgh. Email address: T.V.Pricope@sms.ed.ac.uk Many real-world applications can be described as large-scale games of imperfect information. This kind of games is particularly harder than the deterministic one as the search space is even more sizeable. In this paper, I want to explore the power of reinforcement learning in such an environment; that is why I take a look at one of the most popular game of such type, no limit Texas Hold’em Poker, yet unsolved, developing multiple agents with different learning paradigms and techniques and then comparing their respective performances. When applied to no-limit Hold’em Poker, deep reinforcement learning agents clearly outperform agents with a more traditional approach. Moreover, if these last agents rival a human beginner level of play, the ones based on reinforcement learning compare to an amateur human player. The main algorithm uses Fictitious Play in combination with ANNs and some handcrafted metrics. We also applied the main algorithm to another game of imperfect information, less complex than Poker, in order to show the scalability of this solution and the increase in performance when put neck in neck with established classical approaches from the reinforcement learning literature. http://193.231.18.162/index.php/subbinformatica/article/view/3888Artificial Intelligence, Computer Poker, Adaptive Learning, Fictitious Play, Deep Reinforcement Learning, Neural Networks. |
| spellingShingle | Tidor-Vlad PRICOPE A VIEW ON DEEP REINFORCEMENT LEARNING IN IMPERFECT INFORMATION GAMES Studia Universitatis Babes-Bolyai: Series Informatica Artificial Intelligence, Computer Poker, Adaptive Learning, Fictitious Play, Deep Reinforcement Learning, Neural Networks. |
| title | A VIEW ON DEEP REINFORCEMENT LEARNING IN IMPERFECT INFORMATION GAMES |
| title_full | A VIEW ON DEEP REINFORCEMENT LEARNING IN IMPERFECT INFORMATION GAMES |
| title_fullStr | A VIEW ON DEEP REINFORCEMENT LEARNING IN IMPERFECT INFORMATION GAMES |
| title_full_unstemmed | A VIEW ON DEEP REINFORCEMENT LEARNING IN IMPERFECT INFORMATION GAMES |
| title_short | A VIEW ON DEEP REINFORCEMENT LEARNING IN IMPERFECT INFORMATION GAMES |
| title_sort | view on deep reinforcement learning in imperfect information games |
| topic | Artificial Intelligence, Computer Poker, Adaptive Learning, Fictitious Play, Deep Reinforcement Learning, Neural Networks. |
| url | http://193.231.18.162/index.php/subbinformatica/article/view/3888 |
| work_keys_str_mv | AT tidorvladpricope aviewondeepreinforcementlearninginimperfectinformationgames AT tidorvladpricope viewondeepreinforcementlearninginimperfectinformationgames |