Pub Date : 2016-07-02DOI: 10.1109/CIG.2016.7860388
Christoffer Holmgård, J. Togelius, L. Henriksen
In this paper, we present the idea that game design, player modeling, and procedural content generation may offer new methods for modern psychological assessment, allowing for daily cognitive assessment in ways previously unseen. We suggest that games often share properties with psychological tests and that the overlap between the two domains might allow for creating games that contain assessment elements and provide examples from the literature that already show this. While approaches like these are typically seen as adding noise to a particular instrument in a psychometric context, research in player modeling demonstrates that it is possible to extract reliable measures corresponding to psychological constructs from in-game behavior and performance. Given these observations, we suggest that the combination of game design, player modeling, and procedural content generation offers new opportunities for conducting psychometric testing with a higher frequency and a higher degree of personalization than has previously been possible. Finally, we describe how we are currently implementing the first version of this vision in the form of an application for mobile devices that will soon be used in upcoming user studies.
{"title":"Computational intelligence and cognitive performance assessment games","authors":"Christoffer Holmgård, J. Togelius, L. Henriksen","doi":"10.1109/CIG.2016.7860388","DOIUrl":"https://doi.org/10.1109/CIG.2016.7860388","url":null,"abstract":"In this paper, we present the idea that game design, player modeling, and procedural content generation may offer new methods for modern psychological assessment, allowing for daily cognitive assessment in ways previously unseen. We suggest that games often share properties with psychological tests and that the overlap between the two domains might allow for creating games that contain assessment elements and provide examples from the literature that already show this. While approaches like these are typically seen as adding noise to a particular instrument in a psychometric context, research in player modeling demonstrates that it is possible to extract reliable measures corresponding to psychological constructs from in-game behavior and performance. Given these observations, we suggest that the combination of game design, player modeling, and procedural content generation offers new opportunities for conducting psychometric testing with a higher frequency and a higher degree of personalization than has previously been possible. Finally, we describe how we are currently implementing the first version of this vision in the form of an application for mobile devices that will soon be used in upcoming user studies.","PeriodicalId":6594,"journal":{"name":"2016 IEEE Conference on Computational Intelligence and Games (CIG)","volume":"1 1","pages":"1-8"},"PeriodicalIF":0.0,"publicationDate":"2016-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89856285","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-07-02DOI: 10.1109/CIG.2016.7860385
J. Togelius, Georgios N. Yannakakis
Arguably the grand goal of artificial intelligence research is to produce machines with general intelligence: the capacity to solve multiple problems, not just one. Artificial intelligence (AI) has investigated the general intelligence capacity of machines within the domain of games more than any other domain given the ideal properties of games for that purpose: controlled yet interesting and computationally hard problems. This line of research, however, has so far focused solely on one specific way of which intelligence can be applied to games: playing them. In this paper, we build on the general game-playing paradigm and expand it to cater for all core AI tasks within a game design process. That includes general player experience and behavior modeling, general non-player character behavior, general AI-assisted tools, general level generation and complete game generation. The new scope for general general game AI beyond game-playing broadens the applicability and capacity of AI algorithms and our understanding of intelligence as tested in a creative domain that interweaves problem solving, art, and engineering.
{"title":"General general game AI","authors":"J. Togelius, Georgios N. Yannakakis","doi":"10.1109/CIG.2016.7860385","DOIUrl":"https://doi.org/10.1109/CIG.2016.7860385","url":null,"abstract":"Arguably the grand goal of artificial intelligence research is to produce machines with general intelligence: the capacity to solve multiple problems, not just one. Artificial intelligence (AI) has investigated the general intelligence capacity of machines within the domain of games more than any other domain given the ideal properties of games for that purpose: controlled yet interesting and computationally hard problems. This line of research, however, has so far focused solely on one specific way of which intelligence can be applied to games: playing them. In this paper, we build on the general game-playing paradigm and expand it to cater for all core AI tasks within a game design process. That includes general player experience and behavior modeling, general non-player character behavior, general AI-assisted tools, general level generation and complete game generation. The new scope for general general game AI beyond game-playing broadens the applicability and capacity of AI algorithms and our understanding of intelligence as tested in a creative domain that interweaves problem solving, art, and engineering.","PeriodicalId":6594,"journal":{"name":"2016 IEEE Conference on Computational Intelligence and Games (CIG)","volume":"3 1","pages":"1-8"},"PeriodicalIF":0.0,"publicationDate":"2016-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"76566279","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-07-02DOI: 10.1109/CIG.2016.7860407
F. Silva, Aaron Isaksen, J. Togelius, Andy Nealen
We consider the problem of generating compact sub-optimal game-playing heuristics that can be understood and easily executed by novices. In particular, we seek to find heuristics that can lead to good play while at the same time be expressed as fast and frugal trees or short decision lists. This has applications in automatically generating tutorials and instructions for playing games, but also in analyzing game design and measuring game depth. We use the classic game Blackjack as a testbed, and compare condition induction with the RIPPER algorithm, exhaustive-greedy search in statement space, genetic programming and axis-aligned search. We find that all of these methods can find compact well-playing heuristics under the given constraints, with axis-aligned search performing particularly well.
{"title":"Generating heuristics for novice players","authors":"F. Silva, Aaron Isaksen, J. Togelius, Andy Nealen","doi":"10.1109/CIG.2016.7860407","DOIUrl":"https://doi.org/10.1109/CIG.2016.7860407","url":null,"abstract":"We consider the problem of generating compact sub-optimal game-playing heuristics that can be understood and easily executed by novices. In particular, we seek to find heuristics that can lead to good play while at the same time be expressed as fast and frugal trees or short decision lists. This has applications in automatically generating tutorials and instructions for playing games, but also in analyzing game design and measuring game depth. We use the classic game Blackjack as a testbed, and compare condition induction with the RIPPER algorithm, exhaustive-greedy search in statement space, genetic programming and axis-aligned search. We find that all of these methods can find compact well-playing heuristics under the given constraints, with axis-aligned search performing particularly well.","PeriodicalId":6594,"journal":{"name":"2016 IEEE Conference on Computational Intelligence and Games (CIG)","volume":"14 1","pages":"1-8"},"PeriodicalIF":0.0,"publicationDate":"2016-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86709243","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-07-02DOI: 10.1109/CIG.2016.7860431
M. Tamassia, W. Raffe, R. Sifa, Anders Drachen, Fabio Zambetta, M. Hitchens
Destiny is, to date, the most expensive digital game ever released with a total operating budget of over half a billion US dollars. It stands as one of the main examples of AAA titles, the term used for the largest and most heavily marketed game productions in the games industry. Destiny is a blend of a shooter game and massively multi-player online game, and has attracted dozens of millions of players. As a persistent game title, predicting retention and churn in Destiny is crucial to the running operations of the game, but prediction has not been attempted for this type of game in the past. In this paper, we present a discussion of the challenge of predicting churn in Destiny, evaluate the area under curve (ROC) of behavioral features, and use Hidden Markov Models to develop a churn prediction model for the game.
{"title":"Predicting player churn in destiny: A Hidden Markov models approach to predicting player departure in a major online game","authors":"M. Tamassia, W. Raffe, R. Sifa, Anders Drachen, Fabio Zambetta, M. Hitchens","doi":"10.1109/CIG.2016.7860431","DOIUrl":"https://doi.org/10.1109/CIG.2016.7860431","url":null,"abstract":"Destiny is, to date, the most expensive digital game ever released with a total operating budget of over half a billion US dollars. It stands as one of the main examples of AAA titles, the term used for the largest and most heavily marketed game productions in the games industry. Destiny is a blend of a shooter game and massively multi-player online game, and has attracted dozens of millions of players. As a persistent game title, predicting retention and churn in Destiny is crucial to the running operations of the game, but prediction has not been attempted for this type of game in the past. In this paper, we present a discussion of the challenge of predicting churn in Destiny, evaluate the area under curve (ROC) of behavioral features, and use Hidden Markov Models to develop a churn prediction model for the game.","PeriodicalId":6594,"journal":{"name":"2016 IEEE Conference on Computational Intelligence and Games (CIG)","volume":"9 1","pages":"1-8"},"PeriodicalIF":0.0,"publicationDate":"2016-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80633053","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-07-02DOI: 10.1109/CIG.2016.7860423
Anders Drachen, James Green, Chester Gray, Elie Harik, Patty Lu, R. Sifa, D. Klabjan
Behavioral profiling in digital games with persistent online worlds are vital for a variety of tasks ranging from understanding the player community to informing design and business decisions. In this paper behavioral profiles are developed for the online multiplayer shooter/role-playing game Destiny, the most expensive game to be launched to date and a unique hybrid incorporating designs from multiple traditional genres. The profiles are based on playstyle features covering a total of 41 features and over 4,800 randomly selected players at the highest level in the game. Four clustering models were applied (k-means, Gaussian mixture models, k-maxoids and Archetype Analysis) across the two primary game modes in Destiny: Player-versus-Player and Player-versus-Environment. The performance of each model is described and cross-model analysis is used to identify four to five distinct playstyles across each method, using a variety of similarity metrics. Discussion on which model to use in different circumstances is provided. The profiles are translated into design language and the insights they provide into the behavior of Destiny's player base described.
{"title":"Guns and guardians: Comparative cluster analysis and behavioral profiling in destiny","authors":"Anders Drachen, James Green, Chester Gray, Elie Harik, Patty Lu, R. Sifa, D. Klabjan","doi":"10.1109/CIG.2016.7860423","DOIUrl":"https://doi.org/10.1109/CIG.2016.7860423","url":null,"abstract":"Behavioral profiling in digital games with persistent online worlds are vital for a variety of tasks ranging from understanding the player community to informing design and business decisions. In this paper behavioral profiles are developed for the online multiplayer shooter/role-playing game Destiny, the most expensive game to be launched to date and a unique hybrid incorporating designs from multiple traditional genres. The profiles are based on playstyle features covering a total of 41 features and over 4,800 randomly selected players at the highest level in the game. Four clustering models were applied (k-means, Gaussian mixture models, k-maxoids and Archetype Analysis) across the two primary game modes in Destiny: Player-versus-Player and Player-versus-Environment. The performance of each model is described and cross-model analysis is used to identify four to five distinct playstyles across each method, using a variety of similarity metrics. Discussion on which model to use in different circumstances is provided. The profiles are translated into design language and the insights they provide into the behavior of Destiny's player base described.","PeriodicalId":6594,"journal":{"name":"2016 IEEE Conference on Computational Intelligence and Games (CIG)","volume":"140 1","pages":"1-8"},"PeriodicalIF":0.0,"publicationDate":"2016-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89918466","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-07-02DOI: 10.1109/CIG.2016.7860398
André Mendes, J. Togelius, Andy Nealen
In general video game playing, the challenge is to create agents that play unseen games proficiently. Stochastic tree search algorithms, like Monte Carlo Tree Search, perform relatively well on this task. However, performance is non-transitive: different agents perform best in different games, which means that there is not a single agent that is the best in all the games. Rather, some types of games are dominated by a few agents whereas other different agents dominate other types of games. Thus, it should be possible to construct a hyper-agent that selects from a portfolio, in which constituent sub-agents will play a new game best. Since there is no knowledge about the games, the agent needs to use available features to predict the most suitable algorithm. This work constructs such a hyper-agent using the General Video Game Playing Framework (GVGAI). The proposed method achieves promising results that show the applicability of hyper-heuristics in general video game playing and related tasks.
{"title":"Hyper-heuristic general video game playing","authors":"André Mendes, J. Togelius, Andy Nealen","doi":"10.1109/CIG.2016.7860398","DOIUrl":"https://doi.org/10.1109/CIG.2016.7860398","url":null,"abstract":"In general video game playing, the challenge is to create agents that play unseen games proficiently. Stochastic tree search algorithms, like Monte Carlo Tree Search, perform relatively well on this task. However, performance is non-transitive: different agents perform best in different games, which means that there is not a single agent that is the best in all the games. Rather, some types of games are dominated by a few agents whereas other different agents dominate other types of games. Thus, it should be possible to construct a hyper-agent that selects from a portfolio, in which constituent sub-agents will play a new game best. Since there is no knowledge about the games, the agent needs to use available features to predict the most suitable algorithm. This work constructs such a hyper-agent using the General Video Game Playing Framework (GVGAI). The proposed method achieves promising results that show the applicability of hyper-heuristics in general video game playing and related tasks.","PeriodicalId":6594,"journal":{"name":"2016 IEEE Conference on Computational Intelligence and Games (CIG)","volume":"16 1","pages":"1-8"},"PeriodicalIF":0.0,"publicationDate":"2016-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78802315","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-06-14DOI: 10.1109/CIG.2016.7860424
Michael Cook, J. Gow, S. Colton
Procedural generation is important to modern game development as both an artistic implement and an engineering tool. However, developing procedural generators and understanding how they work are both difficult tasks, and even more so for novice developers. In this paper we describe Danesh, a tool to help in analysing, changing and exploring procedural content generators. In particular, we describe several features in Danesh which help a user optimise their procedural generator towards a certain kind of output by automatically changing parameters and evaluating the effect it has on the generator. We compare different approaches to these tasks and describe our future intentions for Danesh's automated features.
{"title":"Towards the automatic optimisation of procedural content generators","authors":"Michael Cook, J. Gow, S. Colton","doi":"10.1109/CIG.2016.7860424","DOIUrl":"https://doi.org/10.1109/CIG.2016.7860424","url":null,"abstract":"Procedural generation is important to modern game development as both an artistic implement and an engineering tool. However, developing procedural generators and understanding how they work are both difficult tasks, and even more so for novice developers. In this paper we describe Danesh, a tool to help in analysing, changing and exploring procedural content generators. In particular, we describe several features in Danesh which help a user optimise their procedural generator towards a certain kind of output by automatically changing parameters and evaluating the effect it has on the generator. We compare different approaches to these tasks and describe our future intentions for Danesh's automated features.","PeriodicalId":6594,"journal":{"name":"2016 IEEE Conference on Computational Intelligence and Games (CIG)","volume":"33 1","pages":"1-8"},"PeriodicalIF":0.0,"publicationDate":"2016-06-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85095058","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2016-05-06DOI: 10.1109/CIG.2016.7860433
Michal Kempka, Marek Wydmuch, Grzegorz Runc, Jakub Toczek, Wojciech Jaśkowski
The recent advances in deep neural networks have led to effective vision-based reinforcement learning methods that have been employed to obtain human-level controllers in Atari 2600 games from pixel data. Atari 2600 games, however, do not resemble real-world tasks since they involve non-realistic 2D environments and the third-person perspective. Here, we propose a novel test-bed platform for reinforcement learning research from raw visual information which employs the first-person perspective in a semi-realistic 3D world. The software, called ViZDoom, is based on the classical first-person shooter video game, Doom. It allows developing bots that play the game using the screen buffer. ViZDoom is lightweight, fast, and highly customizable via a convenient mechanism of user scenarios. In the experimental part, we test the environment by trying to learn bots for two scenarios: a basic move-and-shoot task and a more complex maze-navigation problem. Using convolutional deep neural networks with Q-learning and experience replay, for both scenarios, we were able to train competent bots, which exhibit human-like behaviors. The results confirm the utility of ViZDoom as an AI research platform and imply that visual reinforcement learning in 3D realistic first-person perspective environments is feasible.
{"title":"ViZDoom: A Doom-based AI research platform for visual reinforcement learning","authors":"Michal Kempka, Marek Wydmuch, Grzegorz Runc, Jakub Toczek, Wojciech Jaśkowski","doi":"10.1109/CIG.2016.7860433","DOIUrl":"https://doi.org/10.1109/CIG.2016.7860433","url":null,"abstract":"The recent advances in deep neural networks have led to effective vision-based reinforcement learning methods that have been employed to obtain human-level controllers in Atari 2600 games from pixel data. Atari 2600 games, however, do not resemble real-world tasks since they involve non-realistic 2D environments and the third-person perspective. Here, we propose a novel test-bed platform for reinforcement learning research from raw visual information which employs the first-person perspective in a semi-realistic 3D world. The software, called ViZDoom, is based on the classical first-person shooter video game, Doom. It allows developing bots that play the game using the screen buffer. ViZDoom is lightweight, fast, and highly customizable via a convenient mechanism of user scenarios. In the experimental part, we test the environment by trying to learn bots for two scenarios: a basic move-and-shoot task and a more complex maze-navigation problem. Using convolutional deep neural networks with Q-learning and experience replay, for both scenarios, we were able to train competent bots, which exhibit human-like behaviors. The results confirm the utility of ViZDoom as an AI research platform and imply that visual reinforcement learning in 3D realistic first-person perspective environments is feasible.","PeriodicalId":6594,"journal":{"name":"2016 IEEE Conference on Computational Intelligence and Games (CIG)","volume":"97 1","pages":"1-8"},"PeriodicalIF":0.0,"publicationDate":"2016-05-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85752075","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-08-01DOI: 10.1109/CIG.2015.7317660
G. Kendall
Hyper-heuristics have been successfully applied in solving a variety of computational search problems. We discuss how a hyper-heuristic can be used to generate adaptive strategies for games. Based on a set of low-level heuristics (or strategies), a hyper-heuristic game player can generate strategies which adapt to both the behaviour of the co-players and the game dynamics. By using a simple heuristic selection mechanism, a number of existing heuristics for specialised games can be integrated into an automated game player. We have developed hyper-heuristics for three games: iterated prisoner's dilemma, repeated Goofspiel and the competitive traveling salesmen problem. The results demonstrate that a hyper-heuristic game player outperforms the low-level heuristics, when used individually in game playing and it can generate adaptive strategies even if the low-level heuristics are deterministic. This methodology provides an efficient way to develop new strategies for games based on existing strategies.
{"title":"Keynote speech IV: Where games meet hyper-heuristics","authors":"G. Kendall","doi":"10.1109/CIG.2015.7317660","DOIUrl":"https://doi.org/10.1109/CIG.2015.7317660","url":null,"abstract":"Hyper-heuristics have been successfully applied in solving a variety of computational search problems. We discuss how a hyper-heuristic can be used to generate adaptive strategies for games. Based on a set of low-level heuristics (or strategies), a hyper-heuristic game player can generate strategies which adapt to both the behaviour of the co-players and the game dynamics. By using a simple heuristic selection mechanism, a number of existing heuristics for specialised games can be integrated into an automated game player. We have developed hyper-heuristics for three games: iterated prisoner's dilemma, repeated Goofspiel and the competitive traveling salesmen problem. The results demonstrate that a hyper-heuristic game player outperforms the low-level heuristics, when used individually in game playing and it can generate adaptive strategies even if the low-level heuristics are deterministic. This methodology provides an efficient way to develop new strategies for games based on existing strategies.","PeriodicalId":6594,"journal":{"name":"2016 IEEE Conference on Computational Intelligence and Games (CIG)","volume":"1 1","pages":"19"},"PeriodicalIF":0.0,"publicationDate":"2015-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90726271","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-08-01DOI: 10.1109/CIG.2015.7317657
X. Yao
Co-evolution has been used widely in automatic learning of game-playing strategies, e.g., for iterated prisoner's dilemma games, backgammon, chess, etc. It is a very interesting form of learning because it learns by interactions only, without any explicit target output information. In other words, the correct choices or moves were not provided as teacher information in learning. Yet co-evolutionary learning is still able to learn high-performance, in comparison to average human performance, game-playing strategies. Interestingly, the research of co-evolutionary learning has not focused on its generalisation ability, in sharp contrast to machine learning in general, where generalisation is at the heart of learning of any form. This talk presents one of the few generic frameworks that are available for measuring generalisation of coevolutionary learning. It enables us to discuss and study generalisation of different co-evolutionary algorithms more objectively and quantitatively. As a result, it enables us to draw more appropriate conclusions about the abilities of our learned game-playing strategies in dealing with totally new and unseens environments (including opponents). The iterated prisoner's dilemma game will be used as an example in this talk to illustrate our theoretical framework and performance improvements we could gain by following this more principled approach to co-evolutionary learning.
{"title":"Keynote speech I: Co-evolutionary learning in game-playing","authors":"X. Yao","doi":"10.1109/CIG.2015.7317657","DOIUrl":"https://doi.org/10.1109/CIG.2015.7317657","url":null,"abstract":"Co-evolution has been used widely in automatic learning of game-playing strategies, e.g., for iterated prisoner's dilemma games, backgammon, chess, etc. It is a very interesting form of learning because it learns by interactions only, without any explicit target output information. In other words, the correct choices or moves were not provided as teacher information in learning. Yet co-evolutionary learning is still able to learn high-performance, in comparison to average human performance, game-playing strategies. Interestingly, the research of co-evolutionary learning has not focused on its generalisation ability, in sharp contrast to machine learning in general, where generalisation is at the heart of learning of any form. This talk presents one of the few generic frameworks that are available for measuring generalisation of coevolutionary learning. It enables us to discuss and study generalisation of different co-evolutionary algorithms more objectively and quantitatively. As a result, it enables us to draw more appropriate conclusions about the abilities of our learned game-playing strategies in dealing with totally new and unseens environments (including opponents). The iterated prisoner's dilemma game will be used as an example in this talk to illustrate our theoretical framework and performance improvements we could gain by following this more principled approach to co-evolutionary learning.","PeriodicalId":6594,"journal":{"name":"2016 IEEE Conference on Computational Intelligence and Games (CIG)","volume":"52 360 1","pages":"16"},"PeriodicalIF":0.0,"publicationDate":"2015-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"83736200","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：