ArticlePDF Available

Review and analysis of research on Video Games and Artificial Intelligence: a look back and a step forward

Authors:
João Paulo Pereira de Sousa at Instituto Politécnico de Bragança - Research Centre in Digitalization and Intelligent Robotics (CeDRI)
João Paulo Pereira de Sousa
  • Instituto Politécnico de Bragança - Research Centre in Digitalization and Intelligent Robotics (CeDRI)
Rogério Tavares
Rogério Tavares
Rogério Tavares
  • This person is not on ResearchGate, or hasn't claimed this research yet.
João Pedro Gomes at Polytechnic Institute of Bragança
João Pedro Gomes
Vitor Mendonça at Polytechnic Institute of Bragança
Vitor Mendonça
Download full-text PDF
Read full-text
Download citation

Abstract

This article shows the intimate relationship between Artificial Intelligence (AI) and video games research in 13 categories of analysis based on a bibliometric survey carried out in the Scopus database. We first briefly reviewed the relation between video games and AI. Then, we introduced the methodology of literature collection, presented and discussed the query, as well the flow of data treatment in the applications and plugins used. Since the article is concerned with a historical point of view of the relationship between digital games and AI the results were many and, therefore, we focused on the top 10 of each ranking, and discussed these results separately. Finally, we discuss the limitations of our review, proposing future research directions for scholars.
Content uploaded by João Paulo Pereira de Sousa
Author content
All content in this area was uploaded by João Paulo Pereira de Sousa on Sep 19, 2022
Content may be subject to copyright.
ScienceDirect
Available online at www.sciencedirect.com
Procedia Computer Science 204 (2022) 315–323
1877-0509 © 2022 The Authors. Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0)
Peer-review under responsibility of the scientific committee of the International Conference on Industry Sciences and Computer Sciences
Innovation
10.1016/j.procs.2022.08.038
10.1016/j.procs.2022.08.038 1877-0509
© 2022 The Authors. Published by Elsevier B.V.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0)
Peer-review under responsibility of the scientic committee of the International Conference on Industry Sciences and Computer
Sciences Innovation
Available online at www.sciencedirect.com
ScienceDirect
Procedia Computer Science 00 (2021) 000–000
www.elsevier.com/locate/procedia
1877-0509 © 2022 The Authors. Published by ELSEVIER B.V.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0)
Peer-review under responsibility of the scientific committee of the International Conference on Industry Sciences and Computer Sciences
Innovation
International Conference on Industry Sciences and Computer Science Innovation
Review and analysis of research on Video Games and Artificial
Intelligence: a look back and a step forward
João Paulo Sousaa,b* , Rogério Tavaresa,c, João Pedro Gomesa, Vitor Mendonçaa
aInstituto Politécnico de Bragança, Bragança, Portugal
bResearch Centre in Digitalization and Intelligent Robotics, Bragança, Portugal
cCenter for Informatics and systems of the University of Coimbra, Coimbra, Portugal
Abstract
This article shows the intimate relationship between Artificial Intelligence (AI) and video games research in 13 categories of
analysis based on a bibliometric survey carried out in the Scopus database. We first briefly reviewed the relation between video
games and AI. Then, we introduced the methodology of literature collection, presented and discussed the query, as well the flow
of data treatment in the applications and plugins used. Since the article is concerned with a historical point of view of the relationship
between digital games and AI the results were many and, therefore, we focused on the top 10 of each ranking, and discussed these
results separately. Finally, we discuss the limitations of our review, proposing future research directions for scholars.
© 2022 The Authors. Published by ELSEVIER B.V.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0)
Peer-review under responsibility of the scientific committee of the International Conference on Industry Sciences and Computer
Sciences Innovation
Keywords: Computer games; AI; Games, Bibliometric analysis, Virtual Reality.
1. Introduction
Since the beginning of AI, it has been related to games [1]. These games were fundamentally used to measure the
ability and performance of the AI. In 1950 Shannon shared the main aspects of computational routine in the article
* Corresponding author. Tel.: +351300029900; fax: +351278201340.
E-mail address: address: jpaulo@ipb.pt
Available online at www.sciencedirect.com
ScienceDirect
Procedia Computer Science 00 (2021) 000–000
www.elsevier.com/locate/procedia
1877-0509 © 2022 The Authors. Published by ELSEVIER B.V.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0)
Peer-review under responsibility of the scientific committee of the International Conference on Industry Sciences and Computer Sciences
Innovation
International Conference on Industry Sciences and Computer Science Innovation
Review and analysis of research on Video Games and Artificial
Intelligence: a look back and a step forward
João Paulo Sousaa,b* , Rogério Tavaresa,c, João Pedro Gomesa, Vitor Mendonçaa
aInstituto Politécnico de Bragança, Bragança, Portugal
bResearch Centre in Digitalization and Intelligent Robotics, Bragança, Portugal
cCenter for Informatics and systems of the University of Coimbra, Coimbra, Portugal
Abstract
This article shows the intimate relationship between Artificial Intelligence (AI) and video games research in 13 categories of
analysis based on a bibliometric survey carried out in the Scopus database. We first briefly reviewed the relation between video
games and AI. Then, we introduced the methodology of literature collection, presented and discussed the query, as well the flow
of data treatment in the applications and plugins used. Since the article is concerned with a historical point of view of the relationship
between digital games and AI the results were many and, therefore, we focused on the top 10 of each ranking, and discussed these
results separately. Finally, we discuss the limitations of our review, proposing future research directions for scholars.
© 2022 The Authors. Published by ELSEVIER B.V.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0)
Peer-review under responsibility of the scientific committee of the International Conference on Industry Sciences and Computer
Sciences Innovation
Keywords: Computer games; AI; Games, Bibliometric analysis, Virtual Reality.
1. Introduction
Since the beginning of AI, it has been related to games [1]. These games were fundamentally used to measure the
ability and performance of the AI. In 1950 Shannon shared the main aspects of computational routine in the article
* Corresponding author. Tel.: +351300029900; fax: +351278201340.
E-mail address: address: jpaulo@ipb.pt
316 João Paulo Sousa et al. / Procedia Computer Science 204 (2022) 315–323
2 Author name / Procedia Computer Science 00 (2019) 000–000
“Programming a computer for playing chess” [1]. The paper introduces a process for the computer to be able to decide
which move to perform next. For this, the minimax procedure was used, based on an evaluation function of a certain
chess position.
In 1952 a British professor of computer science, Alexander Douglass, created the OXO game also known as tic-
tac-toe, as part of his Ph.D. research on human-computer interaction for the University of Cambridge. OXO was a
single-player game in which one player plays against the computer. The first computer game is generally assumed to
be "Spacewar!", developed in 1962 at MIT. Pac-Man, a simple arcade game from 1981 introduced different AI
heuristics to each one of the four ghosts of the game, creating personalities for these enemies. The first sequel of Pac-
Man, Ms. Pac-Man, released in 1982, is more challenging and according to [2] the version Ms. Pac-Man vs. Ghosts
is a preference for the current AI's competitions since it supports the two main approaches to train videogame bots:
Genetic Algorithms and Reinforcement Learning. There is a common criticism about game AI, based on its objective,
that this kind of AI needs only to look intelligent from a player's perspective, while AI, on an industrial approach, for
example, must focus on real-world problems [3]. But, since the '90s, when games evolve to be more complex, mixing
different genres, it is possible to see the uses of genetic algorithms, neural networks, evolutionary approaches [4], and
Deep Learning (a subset of Machine Learning) in video games, thereby decreasing the difference between AI and
game AI. Nowadays game critics include the performance of AI in a game in the same way they assess its graphics,
settings, physics, narrative, and more [5]. The Holy Grail for AI is to surpass the best human players in complex games
such as StarCraft (Blizzard Entertainment, 1998), a sci-fi military-strategic game with intense resource management,
where players assume three different roles using three different races. According to [6] the mastering of StarCraft is a
challenge for artificial agents to compete and coordinate with other agents inside complex environments, so this game
has emerged as a real target for artificial intelligence research. Since 1998, its release date, the StarCraft agents never
come close to matching the best StarCraft players abilities, but in 2019 the StarCraft's agent AlphaStar surpasses
99,8% of the best human players in all 3 races of the game Protoss, Terrans, and Zergs, reaching a new milestone for
the AI.
Despite the number of scientific papers related to this topic, and to the best of our knowledge, there is no
bibliometric, scientometric, or informetric analysis of the existing scientific literature on the use of AI in video games.
A search on Scopus and Web of Science (WoS) databases, with no proper results, seems to confirm that. Therefore,
we intend to make a first contribution to filling this gap, by conducting a bibliometric study on the distribution and
interest of publications relating to research between video games and AI, as well to identify the sources and authors
with more scientific production. Thus, the research questions that this paper attempt to answer are the following: What
are the most influential published articles? What are the main publication sources? Who are the most prolific authors
from de search? What are the most frequently used keywords in articles published? To respond to these research
questions, the major purpose of this study is to provide a holistic review of video games and AI research and to identify
the challenges and gaps that are needed to be addressed by future research. The rest of the paper is structured as
follows: first, the methodology used to obtain the dataset is described. Then, in Section 3, the results and quantitative
analysis are presented, and finally, the discussion of the results and future work are addressed in Section 4.
2. Methodology
Before performing this search, topic keywords were applied to Scopus and WoS databases. As the Scopus took a
significantly larger number than WoS, it was chosen. Our bibliometric analysis was carried out on publications
published in peer-reviewed journals and conference proceedings; other documents such as books, reviews were
excluded from this bibliometric analysis. Only publications written in English were considered, there were no time
restrictions, and the search returned articles from 1971 to 2021. The resulting query is as follows:
(TITLE-ABS-KEY("computer game" OR "video game" OR videogame) AND TITLE-ABS-KEY("artificial intelligence")) AND (
LIMIT-TO ( DOCTYPE,"cp" ) OR LIMIT-TO ( DOCTYPE,"ar" ) ) AND ( LIMIT-TO ( LANGUAGE,"English" ))
A set of 2604 publications was returned. This data was then downloaded in RIS format. Later, this dataset was
imported to the Biblioshiny for Bibliometrix in R 4.1.1. R is an open-source environment for statistical computing
and graphics, with a collection of several packages such as Bibliometrix, developed specifically for bibliometric and
scientometric studies [7]. Biblioshiny is a user-friendly web interface for Bibliometrix, to perform comprehensive
João Paulo Sousa et al. / Procedia Computer Science 204 (2022) 315–323 317
Author name / Procedia Computer Science 00 (2019) 000–000 3
science mapping analysis. Using those tools were carried out an analysis of a publication dataset and built matrices to
perform network analysis for conceptual structure, intellectual structure , and social structure.
3. Analysis
Table 1 shows key information about data and document types retrieved from the Scopus database query on
November 12, 2021. A total of 2604 publications from 1971 to 2022 were retrieved, of which 456 were journal articles ,
and 2148 articles from conference proceedings. A total of 5865 authors were identified, the average of authors per
publication was 2.25 and 88% of the documents were written by more than one author.
Table 1. Main information about the search.
Description Results
Total publications 2604
Articles 456
Proceedings papers 2148
Period 1971-2022
Date of query 12-10-2021
Authors 5864
Authors per document 2.25
Co-Authors per documents 3.19
Single-authored documents 351
Sources (Journals,proceedings) 810
Countries 77
3.1. Documents
The chart in Fig. 1(a) shows the growth of the number of publications published annually and the mean of total
citations by year. The first publication returned was from 1971. Before 1999, the number of publications on AI and
video games was on a slow-growth trend, ranging from 0 to 5. Starting from 1998, the number of publications has
been increasing rapidly. Specifically, from 2004 to 2006 that duplicated the number of publications, and also between
2012 and 2016 the number of articles tripled, reaching its maximum of 316 publications in 2006. This trend is also
consistent with the findings reported in other studies related to bibliometric analysis in the video games field [8] and
[9]. But not in line with bibliometric analysis in the AI field ([10], [11]) where the biggest growth curve was in other
periods. The underlying reason that video games and AI research has grown rapidly after 2004 could be attributed to
the increase of interest in using AI solutions to solve classical problems of video games AI as pathfinding, decision
making, strategy, and procedural generation [4].
Fig. 1. (a) Annual Scientific Production and citation and (b) documents by subject area
In Fig. 1(a) it is possible to identify two citation peaks: the first in 1995 and the second in 2015. Two of the three
318 João Paulo Sousa et al. / Procedia Computer Science 204 (2022) 315–323
4 Author name / Procedia Computer Science 00 (2019) 000–000
works that the research returned this year contributed to the first peak ([12], [13]) with 847 and 527 citations
respectively. For the peak of 2015, [14] contributed a lot with 8905 citations, completely distancing itself from the
other publications this year. Looking at the graph in Fig. 1(b), it is clear that the most published field of research of
Computer Sciences, followed by Mathematics and Engineering. This is no surprise given that video games and AI are
inherently related to those fields. Table 2 shows the 10 most influential publications in video games in the AI field,
based on the number of citations. As previously mentioned, [14] is the most cited publication with 8905 citations. The
paper presents the first deep learning model capable of achieving the performance of a professional player in 49 games,
having as input the pixels and the score of the game. In second place with 939 citations appears [15]. This publication
unveils the smart technologies coded and used in Kinect for sensor calibration, human skeleton tracking, and facial
expression tracking. With 833 citations, [12] presents the game-learning program using neural networks to play the
game Gammon, coming in third place. In fourth position comes [16] with 654 citations. This paper presents the
algorithm which would later defeat the best GO player in the world, and register one of the most important moments
in the history of AI. In [13] is used the video game Tetris to study interactive processes of how agents configure their
workplace for specific tasks and how they can continuously manage that workplace. In [17] the authors reflect on
what is necessary for current AI systems to think like humans. For this, they review the progress of cognitive science
and point out paths for what they learn and how they learn. In [6] a new breakthrough in AI and video games is
announced, the multi-agent reinforcement learning algorithm (AlphaStar) is presented. After training, reaches the level
of the grandmaster level in the real-time strategy game (StarCraft II), being able to beat 99.8 percent of all human
players in the competition. In [18] cooperative algorithms based on the A* algorithm are provided that address specific
cooperative pathfinding problems. [19] presents the algorithm for imperfect information settings adapted for games,
where players have perfect information, and it was tested in a professional poker tournament. In [20] an algorithm
based on the A* algorithm is presented, for finding paths in grids with locked and unlocked cells, used to represent
3D terrain in video games. Finally, [21] presents a program for playing poker, which uses learning techniques to build
statistical models of each.
Table 2 - Most Global Cited Documents
Paper
DOI
Total Citations
MNIH V, 2015, NATURE
10.1038/nature14236
8905
HAN J, 2013, IEEE TRANS CYBERN
10.1109/TCYB.2013.2265378
939
TESAU C, 1995, COMMUN ACM
10.1145/203330.203343
833
SILVER D, 2018, SCI
10.1126/science.aar6404
654
KIRSH D, 1995, ARTIF INTELL
10.1016/0004-3702(94)00017-U
517
LAKE BM, 2017, BEHAV BRAIN SCI
10.1017/S0140525X16001837
502
VINYALS O, 2019, NATURE
10.1038/s41586-019-1724-z
437
SILVER D, 2005, PROC ART. INT. INTER. DIG. ENTERT. CONF., AIIDE
NA
291
MORAVČÍK M, 2017, SCIENCE
10.1126/science.aam6960
280
BILLINGS D, 2002, ARTIF INTELL
10.1016/S0004-3702(01)00130-8
191
3.2. Sources
Figure 2 shows an exaggerated concentration on few actors in Most Local Cited Sources, in which the first three
sources account for 46% of the citations, while the other 7 share the rest.
João Paulo Sousa et al. / Procedia Computer Science 204 (2022) 315–323 319
Author name / Procedia Computer Science 00 (2019) 000–000 5
Fig. 2. Top 10 most local cited sources
In Table 3 it is clear the predominance of two main sources. From the top ten publications, the first one alone holds
40%, and the first and the second together hold 58%. This also points to a preponderance of Bioinformatics, as the
source that holds 41% of publications focuses on this area.
Table 3 - Top 10 sources and number of papers published during the period.
Number of papers
417
183
80
65
58
57
47
46
45
43
However, in Table 4, the Source Local Impact presents a more balanced distribution. Even with the first four places
concentrating just over half of the local impact, the distribution is more homogeneous, considering the H-index, since
the difference from first to last place is little more than double, 2.2x (20 vs 9), and not 9.7 times as much (471 vs 43)
as in Most Relevant Sources. In the G-index this difference is greater, 3.5 times (45 vs 10) but it also maintains a more
homogeneous distribution.
Table 4 - Top 10 most impact sources.
Element
h_index
g_index
IJCAI International Joint Conference On Artificial Intelligence
20
29
Lecture Notes In Computer Science (Including Subseries Lecture Notes In AI And In Bioinformatics)
20
31
Proceedings Of The National Conference On Artificial Intelligence
19
35
AAAI Workshop - Technical Report
15
28
IEEE Conference On Computational Intelligence And Games, CIG
13
25
Artificial Intelligence
12
13
IEEE Transactions On Cybernetics
12
14
IEEE Transactions On Computational Intelligence And AI In Games
11
26
2008 Ieee Symposium On Computational Intelligence And Games, CIG 2008
9
10
Proceedings Of The 4th Artificial Intelligence And Interactive Digital Entertainment Conference, AIIDE
2008
9
18
320 João Paulo Sousa et al. / Procedia Computer Science 204 (2022) 315–323
6 Author name / Procedia Computer Science 00 (2019) 000–000
3.3. Authors
The documents analyzed were authored by 5864 different researchers, with an aver age of 0.444 documents per
author. The large majority of those authors, 5555 (95%), produced documents in co-authorship, while 309 (5%)
authors produced their documents without collaboration. There were 351 (13%) single-author documents.
Relatively to co-authorship indices, the Authors per Documents index is 2.25 and The Co-Authors per Documents
is 3.19, the former being the ratio between the total number of authors and the total number of articles and the latter
being the average number of co-authors per article. The collaboration index, that is, the ratio between the number of
authors of multi-authored documents and the number of multi-authored documents is 2.47. Fig. 3(a) shows the most
relevant authors per number of authored documents. The two most productive ones are Vadim Bulitko (University of
Alberta, Canada) with 35 documents, an average of 2.5 per year published continuously from 2005 to 2019, and Julian
Togelius (New York University, USA) with 31 documents, with the first publication happening only in 2014 but with
an average of 3.875 documents a year since then. There are several measures used to quantify the impact of individual
authors. The table in Fig. 3(b) shows the 10 authors with more local impact, by h-index. This is the most common
measure used, although studies suggest that the use of the h-index in ranking scientists should be reconsidered [22].
Fig. 3. (a) The 10 most relevant authors by the number of documents; (b) The 10 most relevant authors by local impact.
3.4. Most Frequent Words
The analysis of the most cited words must be careful, due to the occurrence of synonyms and specializations. For
example, the most cited term (Fig. 4), Artificial Intelligence, is a broad term that involves several specializations, such
as machine learning, deep learning, and neural networks, which appear in 2nd, 14th, and 17th respectively. As for
synonyms, Artificial Intelligence's abbreviation, AI, appears in 10th place, which would increase the initial count of
329 appearances for the keyword to 362, approximately 10% more. Regarding synonyms, we have the broad concept
of digital games, which appears as computer games, or video games, usually, a difference in the hardware used, games,
an abbreviation, computer game and video game, singular forms of the aforementioned plurals, serious game, a game
genre, videogames, a synonym, and AI games and game design, digital games specialization areas. If added, these
terms together come in at 427, that is, more than Artificial Intelligence and its abbreviation together (showing that
these people know how to have fun). When used to analyze trends, as we will see later in Fig. 5, the separation of
synonyms and technologies is important, as it allows us to observe how keywords evolve and become less general
over the years, but when it comes to analyzing only the amount of words, without the need of a timeline, they can be
put together without problems of interpretation.
João Paulo Sousa et al. / Procedia Computer Science 204 (2022) 315–323 321
Author name / Procedia Computer Science 00 (2019) 000–000 7
Fig. 4. Most relevant words by occurrences
Although the term Monte Carlo Tree Search (MCTS) first appeared in 2006 by Remí Coulom [23], it ends up
standing out in 2016 when it is successfully combined with neural networks at work [16] and the era of deep learning
is created. Procedural content generation in video games has a long history, but Fig. 5 shows that in recent years there
has been an increase in interest around challenges posed by game content generation due to the emergence of deep
learning tools. This search trend will likely remain a search for the foreseeable future.
Fig. 5. Trend topics since 2007.
4. Discussion
Through the analysis of articles returned by the query, it was possible to identify two relationships between AI and
video games: video games can use AI algorithms to incorporate into their gameplay. For example, the case of using
an AI algorithm to create certain behaviors in an NPC, controlling the real-time strategy (RTS) game AI, or the
procedural level generation. The use of video games to test and assess the ability of the algorithm to solve problems.
Video games are usually designed to challenge humans and allow you to recreate more or less complex virtual
environments. This makes video games excellent tools to test various cognitive problems, including reasoning,
planning, strategy, coordination, perception, behavior, kinesthetics, among others. One example is the use of video
games in learning algorithms ([6], [16], [19]). In the top 10 of the most cited articles, and through co-occurrence, it is
322 João Paulo Sousa et al. / Procedia Computer Science 204 (2022) 315–323
8 Author name / Procedia Computer Science 00 (2019) 000–000
possible to observe that researchers around the world are using video games to create environments to be used in
learning algorithms. The doubling of the number of articles published in 2016, can be due to the convergence of some
factors that promoted the rapid growth of deep learning, including the availability of a great amount of data and more
processing power given by GPUs evolution, new algorithms like convolutional neural networks, in 2012, Generative
Adversarial Networks, in 2014, and the availability of more user-friendly machine learning frameworks, like
Tensorflow, introduced by Google in 2015.
Future work could deal with some shortcomings of this study: Incorporated other databases such as WoS, and
others. To perform a screening strategy, analyze the content of the publication to verify if its focus is related to AI and
videogames, and thus obtain a better sample. Lastly, considering the limitations of bibliometrics, a deeper content
analysis of papers can be conducted to gain a deeper understanding of the AI and videogames relations.
Acknowledgments
This work has been supported by FCT---Fundação para a Ciência e Tecnologia within the Project Scope: UIDB/05757/2020.
References
[1] C. E. Shannon, ‘Programming a Computer for Playing Chess’, in Computer Chess Compendium, D. Levy, Ed. New York, NY: Springer New
York, 1988, pp. 2–13. doi: 10.1007/978-1-4757-1968-0_1.
[2] F. Domınguez-Estevez, A. A. Sanchez-Ruiz, and P. Pablo, ‘Training Pac-Man bots using Reinforcement Learning and Case-based Reasoning’,
p. 12, 2017.
[3] G. N. Yannakakis, ‘Game AI revisited’, in Proceedings of the 9th conference on Computing Frontiers - CF ’12, Cagliari, Italy, 2012, p. 285.
doi: 10.1145/2212908.2212954.
[4] M. Straeubig, ‘Games, AI and Systems’, Eludamos J. Comput. Game Cult., vol. 10, no. 1, Art. no. 1, Apr. 2020.
[5] ‘Red Dead Redemption 2: A deep dive into Rockstar’s game design’, VentureBeat, Dec. 05, 2018. https://venturebeat.com/2018/12/05/red-
dead-redemption-a-deep-dive-into-rockstars-game-design/ (accessed Nov. 22, 2021).
[6] O. Vinyals et al., ‘Grandmaster level in StarCraft II using multi-agent reinforcement learning’, Nature, vol. 575, no. 7782, pp. 350–354, Nov.
2019, doi: 10.1038/s41586-019-1724-z.
[7] M. Aria and C. Cuccurullo, ‘bibliometrix: An R-tool for comprehensive science mapping analysis’, J. Informetr., vol. 11, no. 4, pp. 959–975,
Nov. 2017, doi: 10.1016/j.joi.2017.08.007.
[8] P. García-Sánchez, A. M. Mora, P. A. Castillo, and I. J. Pérez, ‘A bibliometric study of the research area of videogames using Dimensions.ai
database’, Procedia Comput. Sci., vol. 162, pp. 737–744, Jan. 2019, doi: 10.1016/j.procs.2019.12.045.
[9] A. R. Martins, S. Monteiro, J. C. Duarte, and L. R. Oliveira, ‘Games and learning: a bibliometric analysis of the scientific production’, 2015.
Accessed: Nov. 19, 2021. [Online]. Available: http://repositorium.sdum.uminho.pt/
[10] J. M. Alonso, C. Castiello, and C. Mencar, ‘A Bibliometric Analysis of the Explainable Artificial Intelligence Research Field’, in Information
Processing and Management of Uncertainty in Knowledge-Based Systems. Theory and Foundations, Cham, 2018, pp. 3–15. doi: 10.1007/978-3-
319-91473-2_1.
[11] W. Tu, J. Fang, Fangyu, Jiao, and Y. Ci, ‘Bibliometric research in the field of artificial intelligence’, J. Phys. Conf. Ser., vol. 1883, no. 1, p.
012165, Apr. 2021, doi: 10.1088/1742-6596/1883/1/012165.
[12] C. Tesau and G. Tesau, ‘Temporal Difference Learning and TD-Gammon’, Commun. ACM, vol. 38, no. 3, pp. 58–68, 1995, doi:
10.1145/203330.203343.
[13] D. Kirsh, ‘The intelligent use of space’, Artif. Intell., vol. 73, no. 1–2, pp. 31–68, 1995, doi: 10.1016/0004 -3702(94)00017-U.
[14] V. Mnih et al., ‘Human-level control through deep reinforcement learning’, Nature, vol. 518, no. 7540, pp. 529–533, 2015, doi:
10.1038/nature14236.
[15] J. Han, L. Shao, D. Xu, and J. Shotton, ‘Enhanced Computer Vision With Microsoft Kinect Sensor: A Review’, IEEE Trans. Cybern ., vol. 43,
no. 5, pp. 1318–1334, Oct. 2013, doi: 10.1109/TCYB.2013.2265378.
[16] D. Silver et al., ‘Mastering the game of Go with deep neural networks and tree search’, Nature, vol. 529, no. 7587, pp. 484–489, 2016, doi:
10.1038/nature16961.
[17] B. M. Lake, T. D. Ullman, J. B. Tenenbaum, and S. J. Gershman, ‘Building Machines That Learn and Think Like People’, ArXiv160400289
Cs Stat, Nov. 2016, Accessed: Dec. 04, 2021. [Online]. Available: http://arxiv.org/abs/1604.00289
[18] D. Silver, ‘Cooperative pathfinding’, in Proceedings of the First AAAI Conference on Artificial Intelligence and Interactive Digital
Entertainment, Marina del Rey, California, Jun. 2005, pp. 117–122.
[19] M. Moravčík et al., ‘DeepStack: Expert-level artificial intelligence in heads-up no-limit poker’, Science, vol. 356, no. 6337, pp. 50 8–513, 2017,
doi: 10.1126/science.aam6960.
[20] K. Daniel, A. Nash, S. Koenig, and A. Felner, ‘Theta*: Any-Angle Path Planning on Grids’, J Artif Intell Res JAIR, vol. 39, Jan. 2014, doi:
João Paulo Sousa et al. / Procedia Computer Science 204 (2022) 315–323 323
Author name / Procedia Computer Science 00 (2019) 000–000 9
10.1613/jair.2994.
[21] D. Billings, A. Davidson, J. Schaeffer, and D. Szafron, ‘The challenge of poker’, Artif. Intell., vol. 134, no. 1, pp. 201–240, Jan. 2002, doi:
10.1016/S0004-3702(01)00130-8.
[22] V. Koltun and D. Hafner, ‘The h-index is no longer an effective correlate of scientific reputation’, PLOS ONE, vol. 16, no. 6, p. e0253397,
Jun. 2021, doi: 10.1371/journal.pone.0253397.
[23] R. Coulom, ‘Efficient Selectivity and Backup Operators in Monte-Carlo Tree Search’, in 5th International Conference on Computer and Games,
Turin, Italy, May 2006. Accessed: Dec. 17, 2021. [Online]. Available: https://hal.inria.fr/inria-00116992
... Notably, from 2004 to 2006, the number of publications fluctuated significantly. In both 2012 and 2016, the publication rate doubled, reaching a maximum of 316 in 2006.The fundamental reason for the rapid growth of video games and artificial intelligence research after 2004 can be attributed to people's increasing interest in using artificial intelligence solutions to solve classic problems in video games -artificial intelligence as pathfinding, decision-making, strategy, and procedural generation [1]. This paper investigates the use of the Fine-Tuning Ring Toss Game Optimization Adaptive Artificial Neural Network (FRTGO -AANN) method for artificial intelligence in game software. ...
... However, when there are multiple actions with the same Q -value, the DQN algorithm will overestimate the true Q -value. [1] The Q -Learning algorithm maintains a Q -table, using a table to store the rewards obtained by taking an action a in each state s, that is, the state -value function Q(s,a), . This algorithm has great limitations. ...
Research on the Application of Generative Artificial Intelligence in the Video Game Field
Article
Full-text available
  • Dec 2024
In recent years, with the rapid progress of hardware and computer technology, generative artificial intelligence has achieved unprecedented development. At the same time, video games have become the so-called "ninth art" after painting, architecture and other art categories. In the field of video games, players' requirements for game experience are constantly increasing, and they expect games to provide a richer and more immersive experience. The development of generative artificial intelligence provides new ideas for the game field. Developers can use AI for innovative experiments, explore new game mechanisms and gameplay, and open up infinite possibilities for game design, content creation, and the way players interact. This paper studies the application of generative artificial intelligence in electronic games. It first investigates the current development needs of games and the existing condition of generative artificial intelligence at the current time. This paper discusses and studies three popular related algorithms and lists several popular games that have used generative artificial intelligence in recent years for elaboration.
View
... However, improving the resources available for AI will benefit games besides RTS. With the rise of modern machine learning approaches such as deep learning, and due to an increase in GPU power and capabilities, there is an increase of interest in the literature to integrate deep learning approaches into video game systems [46]. From improved versions of traditional algorithms with neural networks to entirely new deep learning approaches. ...
Hybrid Architecture for AI-Based RTS Games
Article
Full-text available
  • Jan 2025
Video games have evolved into a key part of modern culture and a major economic force, with the global market projected to reach {\}$ 282.30 billion in 2024. As technology advances, video games increasingly demand high computing power, often requiring specialized hardware for optimal performance. Real-time strategy games, in particular, are computationally intensive, with complex artificial intelligence algorithms that simulate numerous units and behaviors in real-time. Specialized gaming PCs are use a dedicated GPU to run video games. Due to the usefulness of GPUs besides gaming, modern processors usually include an integrated GPU, specially in the laptop market. We propose a hybrid architecture that utilizes both the dedicated GPU and the integrated GPU simultaneously, to accelerate AI and physics simulations in video games. The hybrid approach aims to maximize the utilization of all available resources. The AI and physics computations are offloaded from the dedicated GPU to the integrated GPU. Therefore, the dedicated GPU can be used exclusively for rendering, resulting in improved performance. We implemented this architecture in a custom-built game engine using OpenGL for graphics rendering and OpenCL for general-purpose GPU computations. Experimental results highlight the performance characteristics of the hybrid architecture, including the challenges of working with the two devices and multi-tenant GPU interference.
View
... Wstęp W dzisiejszych czasach obserwuje się dynamiczny rozwój branży gier komputerowych, która stała się nie tylko ważnym elementem rynku rozrywki, ale także istotnym obszarem rozwoju zaawansowanych technologii takich jak sztuczna inteligencja [1][2][3]. Pandemia koronawirusa znacznie zwiększyła aktywność online, dzięki której gry komputerowe stały się nie tylko źródłem rozrywki, ale także czasem relaksu po ciężkim dniu pracy i rywalizacją międzynarodową co znacznie przyczyniło się do rozwoju e-sportu [4,5]. ...
Comparision of Artificial Inteligence models used in computer games on the Unity platform
Article
Full-text available
  • Dec 2024
Comparative analysis of artificial intelligence (AI) models used in computer games encompasses various approaches and techniques employed to create intelligent agents within gaming environments. Contemporary computer games utilize advanced AI systems to provide a more realistic and demanding experience for players. Three models were compared: Finite State Machine, AI Utility, and GOAP. The aim was to examine their effectiveness and applications in the context of creating intelligent agents in gaming environments. Modern games increasingly employ sophisticated AI models to ensure player immersion and maintain realism. During the research, each model was implemented, and the obtained results were subjected to detailed analysis.
View
... While its origins lie in the realms of sciences and humanities (Windsor 1976), this analytical approach quickly found its footing in diverse sectors like economics and management (Leeebvre and Fetterman 1985, 1977-1983. Nowadays, its versatility extends across an array of research areas (Sousa et al. 2022;Morais, Cunha, and Sousa 2021), offering profound insights into current trends, future directions, and author collaborations within the dynamic world of tourism. ...
Literature Review on the XR Technologies in Tourism: One Decade of Virtual Continuum in Tourism
Chapter
  • Jun 2024
This chapter begins with a concise overview of recent studies and reports on how the tourism industry is embracing digital transformation. This initial section aims to provide context for the current significance of technology adoption in tourism and an insight into the prevailing strategies and challenges faced by the industry. In light of this context, it becomes essential to delve into the scientific literature to comprehend how it reflects the evolution of technology adoption within tourism, including themes under investigation and potential trends. Hence, the following section delves into a bibliometric analysis of scientific production related to the use of technologies in tourism over the past decade. This study enables us to gain a comprehensive understanding of the evolving scientific literature on technology adoption in tourism and it sets forth an agenda for future research, paving the way for advancements in the realm of new digital technologies in the tourism sector.
View
... Additive Manufacturing: Additive manufacturing is a new breakthrough in the manufacturing industry by utilizing 3D printing machines or often known as 3D printing. (Mah et al., 2022); (Khurana et al., 2023); (Callister, 2020); (Nijhawan et al., 2022); Natural language processing (Khan & Javaid, 2022); (Prijs et al., 2022); (Mahadevkar et al., 2022) Computer vision (El Khatib et al., 2019); (Mohammed et al., 2023) Cloud computing (Schijven & Kikkawa, 2022); (Sousa et al., 2022); (Zhang et al., 2022) Game playing From the table above it can be explained each of them namely expert System is a software that saves knowledge of an expert thus the computer will have expertise like an expert. With this Natural Language Processing, it is hoped that in the future humans can communicate with computers using everyday language. ...
ChatGPT Artificial Intelligence Integration in Science Learning Media: Systematic Literature Review
Article
Full-text available
  • Jul 2023
Artificial Intelligence technology enables organizations to support and improve knowledge management practices. From an AI perspective, knowledge representation defines organizational processes that are familiar with this knowledge. In this era of digitalization, the presence of ChatGPT technology opens opportunities to utilize AI chatbots for education in Indonesia, especially in developing student competencies (skills) needed in the 21st century. Where the purpose of the research is to explain ChatGPT Artificial Intelligence Integration in Science Learning Media. A review is conducted on the state-of-the-art methods using the preferred reporting items for reviews and meta-analyses (PRISMA) guidelines. We review literature from several publications and analyze the Era of Education 4.0 with the help of computers to improve quality in the industrial sector and also in the field of computer-assisted education. Many innovations can be proposed to improve the quality of learning implementation. GPT chat is a search engine like Google with a form of chat or dialogue. GPT Chat is an information search tool and scientific communication tool.
View
View
9781032715957 previewpdf
Chapter
  • Oct 2024
Future Tech Startups and Innovation in the Age of AI Our book, Future of Tech Startups and Innovations in the Age of AI, mainly focuses on artificial intelligence (AI) tools, AI-based startups, AI-enabled innovations, autonomous AI Agents (Auto-GPT), AI-based marketing startups, machine learning for organizations, AI-internet of things (IoT) for new tech companies, AI-enabled drones for agriculture industry, machine learning (ML)/deep learning (DL)-based drip farming, AI-based driverless cars, AI-based weather prediction startups, AI tools for personal branding, AI-based teaching, AI-based doctor/hospital startups, AI for game companies , AI-based finance tools, AI for human resource management, AI-powered management tools, AI tools for future pandemics, AI/ML-based transportation companies, AI for media, AI for carrier counseling, AI for customer care, AI for next generation businesses, and many more applications. AI tools and techniques will revolutionize startups all over the world. Entrepreneurs, engineers, and practitioners have already moved toward AI-based solutions to reshape businesses. AI/ML will create possibilities and opportunities for improving human lifestyles. AI-enabled startups will work on cost-effective solutions to solve difficult problems. Recently, many research companies are interested in providing solutions and investing a lot in AI-based startups. AI-driven products will revolutionize the "smart world. " AI computing tech companies will help to model human speech recognition systems. Also, AI-based startups will focus on perception and reasoning of autonomous robotic systems. AI/ML-based tech startups will introduce smart online education systems for future pandemics. More interestingly, people are also moving for online job opportunities and trying to work from home. Future innovation needs closer relations between academia and industry. Therefore, online platforms need to be introduced that will only focus on academia and industry linkage. Future AI tech-based startups will focus more on research and development to introduce novel products to the market. Accordingly, engineers and many other people should be trained on AI tools and techniques to introduce innovative solutions for the smart world. In addition, integration of many new technologies with AI will be made possible. AI with IoT, smart cities, unmanned aerial vehicles (UAVs), wireless sensor networks, software-defined networks, network management, vehicular ad hoc networks , flying ad hoc networks, wireless communication technologies, ML, reinforcement learning, federated learning and other mechanisms will introduce new technological products.
View
View
Future Tech Startups and Innovation in the Age of AI
Book
  • Sep 2024
Future Tech Startups and Innovation in the Age of AI Our book, Future of Tech Startups and Innovations in the Age of AI, mainly focuses on artificial intelligence (AI) tools, AI-based startups, AI-enabled innovations, autonomous AI Agents (Auto-GPT), AI-based marketing startups, machine learning for organizations, AI-internet of things (IoT) for new tech companies, AI-enabled drones for agriculture industry, machine learning (ML)/deep learning (DL)-based drip farming, AI-based driverless cars, AI-based weather prediction startups, AI tools for personal branding, AI-based teaching, AI-based doctor/hospital startups, AI for game companies , AI-based finance tools, AI for human resource management, AI-powered management tools, AI tools for future pandemics, AI/ML-based transportation companies, AI for media, AI for carrier counseling, AI for customer care, AI for next generation businesses, and many more applications. AI tools and techniques will revolutionize startups all over the world. Entrepreneurs, engineers, and practitioners have already moved toward AI-based solutions to reshape businesses. AI/ML will create possibilities and opportunities for improving human lifestyles. AI-enabled startups will work on cost-effective solutions to solve difficult problems. Recently, many research companies are interested in providing solutions and investing a lot in AI-based startups. AI-driven products will revolutionize the "smart world. " AI computing tech companies will help to model human speech recognition systems. Also, AI-based startups will focus on perception and reasoning of autonomous robotic systems. AI/ML-based tech startups will introduce smart online education systems for future pandemics. More interestingly, people are also moving for online job opportunities and trying to work from home. Future innovation needs closer relations between academia and industry. Therefore, online platforms need to be introduced that will only focus on academia and industry linkage. Future AI tech-based startups will focus more on research and development to introduce novel products to the market. Accordingly, engineers and many other people should be trained on AI tools and techniques to introduce innovative solutions for the smart world. In addition, integration of many new technologies with AI will be made possible. AI with IoT, smart cities, unmanned aerial vehicles (UAVs), wireless sensor networks, software-defined networks, network management, vehicular ad hoc networks , flying ad hoc networks, wireless communication technologies, ML, reinforcement learning, federated learning and other mechanisms will introduce new technological products.
View
Bibliometric research in the field of artificial intelligence
Article
Full-text available
  • Apr 2021
This paper analyzes the development trend of global artificial intelligence from the perspective of bibliometrics. From the publishing trend, the number of papers published in the field of artificial intelligence is growing rapidly. China and the United States are among the top countries in the output of scientific research papers, and India has the fastest cumulative growth rate. In the research direction, the research on computer vision, image, deep learning and neural network is the research hotspot in recent years.
View
The h-index is no longer an effective correlate of scientific reputation
Article
Full-text available
The impact of individual scientists is commonly quantified using citation-based measures. The most common such measure is the h-index. A scientist’s h-index affects hiring, promotion, and funding decisions, and thus shapes the progress of science. Here we report a large-scale study of scientometric measures, analyzing millions of articles and hundreds of millions of citations across four scientific fields and two data platforms. We find that the correlation of the h-index with awards that indicate recognition by the scientific community has substantially declined. These trends are associated with changing authorship patterns. We show that these declines can be mitigated by fractional allocation of citations among authors, which has been discussed in the literature but not implemented at scale. We find that a fractional analogue of the h-index outperforms other measures as a correlate and predictor of scientific awards. Our results suggest that the use of the h-index in ranking scientists should be reconsidered, and that fractional allocation measures such as h-frac provide more robust alternatives.
View
Games, AI, and Systems
Article
Full-text available
  • Apr 2020
In recent years, we have observed impressive advancements at the intersection of games and artificial intelligence. Often these developments are described in terms of technological progress, while public discourses on their cultural, social and political impact are largely decoupled. I present an alternative rhetoric by speculating about the emergence of AI within social systems. In a radical departure from the dominant discourse, I describe seven roles - Mechanic, Alter/Ego, Observer, Protector, Player, Creator and God - that an AI may assume in the environment of videogames. I reflect on the ramifications of these roles for the idea of an artificial general intelligence (AGI), mainly hoping to irritate the prevailing discussion.
View
A bibliometric study of the research area of videogames using Dimensions.ai database
Article
Full-text available
  • Jan 2019
Videogames are a very interesting area of research for fields as diverse as computer science, health, psychology or even social sciences. Every year a growing number of articles are published in different topics inside this field, so it is very convenient to study the different bibliometric data in order to consolidate the research efforts. Thus, the aim of this work is to conduct a study on the distribution of articles related to videogames in the different fields of research, as well as to measure their interest over time, to identify the sources, countries and authors with the highest scientific production. In order to carry out this analysis, the information system Dimensions.ai has been considered, since it covers a large number of documents and allows for easy downloading and analysis of datasets. According to the study, three countries are the most prolific in this area: USA, Canada and UK. The obtained results also indicate that the fields with the highest number of publications are Information and Computer Sciences, Medical and Health Sciences, and Psychology and Cognitive Sciences, in this order. With regard to the impact of the publications, differences between the number of citations, and the number of Altmetric Attention Score, have been found.
View
Grandmaster level in StarCraft II using multi-agent reinforcement learning
Article
Full-text available
  • Oct 2019
  • NATURE
Many real-world applications require artificial agents to compete and coordinate with other agents in complex environments. As a stepping stone to this goal, the domain of StarCraft has emerged as an important challenge for artificial intelligence research, owing to its iconic and enduring status among the most difficult professional esports and its relevance to the real world in terms of its raw complexity and multi-agent challenges. Over the course of a decade and numerous competitions1–3, the strongest agents have simplified important aspects of the game, utilized superhuman capabilities, or employed hand-crafted sub-systems⁴. Despite these advantages, no previous agent has come close to matching the overall skill of top StarCraft players. We chose to address the challenge of StarCraft using general-purpose learning methods that are in principle applicable to other complex domains: a multi-agent reinforcement learning algorithm that uses data from both human and agent games within a diverse league of continually adapting strategies and counter-strategies, each represented by deep neural networks5,6. We evaluated our agent, AlphaStar, in the full game of StarCraft II, through a series of online games against human players. AlphaStar was rated at Grandmaster level for all three StarCraft races and above 99.8% of officially ranked human players.
View
A Bibliometric Analysis of the Explainable Artificial Intelligence Research Field
Chapter
Full-text available
  • May 2018
This paper presents the results of a bibliometric study of the recent research on eXplainable Artificial Intelligence (XAI) systems. We took a global look at the contributions of scholars in XAI as well as in the subfields of AI that are mostly involved in the development of XAI systems. It is worthy to remark that we found out that about one third of contributions in XAI come from the fuzzy logic community. Accordingly, we went in depth with the actual connections of fuzzy logic contributions with AI to promote and improve XAI systems in the broad sense. Finally, we outlined new research directions aimed at strengthening the integration of different fields of AI, including fuzzy logic, toward the common objective of making AI accessible to people.
View
Building Machines That Learn and Think Like People
Article
Full-text available
  • Apr 2016
  • BEHAV BRAIN SCI
Recent progress in artificial intelligence (AI) has renewed interest in building systems that learn and think like people. Many advances have come from using deep neural networks trained end-to-end in tasks such as object recognition, video games, and board games, achieving performance that equals or even beats humans in some respects. Despite their biological inspiration and performance achievements, these systems differ from human intelligence in crucial ways. We review progress in cognitive science suggesting that truly human-like learning and thinking machines will have to reach beyond current engineering trends in both what they learn, and how they learn it. Specifically, we argue that these machines should (a) build causal models of the world that support explanation and understanding, rather than merely solving pattern recognition problems; (b) ground learning in intuitive theories of physics and psychology, to support and enrich the knowledge that is learned; and (c) harness compositionality and learning-to-learn to rapidly acquire and generalize knowledge to new tasks and situations. We suggest concrete challenges and promising routes towards these goals that can combine the strengths of recent neural network advances with more structured cognitive models.
View
Mastering the game of Go with deep neural networks and tree search
Article
Full-text available
  • Jan 2016
The game of Go has long been viewed as the most challenging of classic games for artificial intelligence owing to its enormous search space and the difficulty of evaluating board positions and moves. Here we introduce a new approach to computer Go that uses ‘value networks’ to evaluate board positions and ‘policy networks’ to select moves. These deep neural networks are trained by a novel combination of supervised learning from human expert games, and reinforcement learning from games of self-play. Without any lookahead search, the neural networks play Go at the level of state-of-the-art Monte Carlo tree search programs that simulate thousands of random games of self-play. We also introduce a new search algorithm that combines Monte Carlo simulation with value and policy networks. Using this search algorithm, our program AlphaGo achieved a 99.8% winning rate against other Go programs, and defeated the human European Go champion by 5 games to 0. This is the first time that a computer program has defeated a human professional player in the full-sized game of Go, a feat previously thought to be at least a decade away.
View
View
DeepStack: Expert-Level Artificial Intelligence in No-Limit Poker
Article
  • Jan 2017
Artificial intelligence has seen a number of breakthroughs in recent years, with games often serving as significant milestones. A common feature of games with these successes is that they involve information symmetry among the players, where all players have identical information. This property of perfect information, though, is far more common in games than in real-world problems. Poker is the quintessential game of imperfect information, and it has been a longstanding challenge problem in artificial intelligence. In this paper we introduce DeepStack, a new algorithm for imperfect information settings such as poker. It combines recursive reasoning to handle information asymmetry, decomposition to focus computation on the relevant decision, and a form of intuition about arbitrary poker situations that is automatically learned from self-play games using deep learning. In a study involving dozens of participants and 44,000 hands of poker, DeepStack becomes the first computer program to beat professional poker players in heads-up no-limit Texas hold'em. Furthermore, we show this approach dramatically reduces worst-case exploitability compared to the abstraction paradigm that has been favored for over a decade.
View