The tutorials that are being held at the IEEE CoG 2020 are the following:
Details about each of the tutorials can be found below:
As we develop our systems for generation and to play games, we want to have a clear understanding of the power of the method and the human cooperativeness in order to demonstratively prove an AI is meeting with its requirements. Further, games are artistic objects with aesthetic concerns which are best evaluated by human subjects testing. These human trials are is not usually performed by computer scientists.
In this respect, while focusing on correcting the technical dimensions of the system, it is easy to overlook player dimensions that are the major stakeholder of the product. Player dimension broadly constitutes their ability to feel that they are a part of the system and can easily immerse in the play. Though this does not mean, removing complexities and unpredictability that are sometimes crucial for an adventurous play journey. This means to remove confusions that are unnecessary and that design is taking into account all; visceral, behavioural and reflective levels of processing of player’s mind. This means considering all effects such as calmness, anxiety, hope, fear and expectations from player’s side.
The tutorial will examine the process of ethical approvals, participant selection, setup of the space and equipment, surveys of participants, data collection methods, data storage, and analysis. The tutorial is broadly on user experience investigation and a focus will be made as well on tools for the understanding of aesthetics and other similarly subjectively defined generation criteria used in many PCG and games objects.
The remainder of the tutorial will engage participants in a demo of the data collection process with several board games in order to demonstrate processes and answer questions. The audience will take away the appreciation of user experience testing and will comprehend its effectiveness upon decreasing the cognitive load of game designers and players. Moreover, the participants will visualize the technicalities of the user experience investigation.
Joseph Alexander Brown (S'09--M'14; jbrown@innopolis.ru) was born in Niagara-on-the-Lake, ON, Canada, on July 6, 1985. He received the B.Sc. (Hons.) with first-class standing in computer science with a concentration in software engineering, and M.Sc. in computer science from Brock University, St. Catharines, ON, Canada in 2007 and 2009, respectively. He received the Ph.D. in computer science from the University of Guelph in 2014. He previously worked for Magna International Inc. as a Manufacturing Systems Analyst and as a visiting researcher at ITU Copenhagen. He is currently an Assistant Professor and head of the Artificial Intelligence in Games Development Lab at Innopolis University in Innopolis, Republic of Tatarstan, Russia and an Adjunct Professor of Computer Science at Brock University, St. Catharines, ON, Canada.
Hamna Aslam (h.aslam@innopolis.ru) received the B.Sc. (Hons. with first-class standing) in Computer Engineering and the M.Sc. in Computer Engineering from UET, Lahore, Pakistan. Previously she was a visiting lecturer at the Institute of Business Management, University of Engineering and Technology Lahore (UET), Pakistan. Besides, Hamna Aslam worked in the energy sector of the government of Pakistan in the capacity of Software Developer. Presently she is a Ph.D. student as well as an instructor at Innopolis University since December 2015. Her research interests include Human Factors in Gaming, Intuitive Game Design and User testing.
In the tutorial, the whole figure of current game AI in game industry is explained by referring the actual cases of game industry in Japan and the world. In these 20 years, game AI technologies drastically evolved, and the fields of game AI have expanded. Game AI technologies have the two main fields: one is AI in game and the other is AI outside game.
AI in game means technologies used in a game dynamically, and AI outside game means AI technologies used for game development processes. AI in game has developed from 1994 to now mainly by using symbolic AI, but in contrast AI outside game has developed in these five years by using connectionism techniques of neural net and deep learning. In the lecture, both of them are explained.
(1) AIs in game consist from Meta-AI, Character AI, and Spatial AI. Meta-AI is AI to control a whole game by observing and changing a game world dynamically from a top of view. Character AI is a brain of a game character to make a decision by itself. Spatial AI is to recognize a world and abstract some terrain features. These three AIs cooperate each other in real-time and produce new user experience. Each AI and the whole system is explained in the former part of the lecture.
(2) AIs outside game are various fields such as auto-balancing of setting parameters, QA-AI (Quality Assurance AI), or procedural contents generation. They support game development and game developers. Especially, number of cases of QA-AI in game industry much increased in these five years. Some examples of QA-AI are there by using deep learning to find bugs or hacking cases to analyze a screen image. Some examples are to analyze user logs to abstract a user's features. Reinforcement learning is also used in game development to make a character clever in a game by training a neural network of character. Some cases in game industry are explained in the latter part.
The turning point of game AI in game industry has come now. Symbolic AI and connectionism AI will become fused. Old style of game AI will change and a new style of game AI will change game industry in the next 5 years. Now learning techniques is too unstable to implement in a game, but it should be researched to become useful in a game. The possibility of combination of symbolic techniques and learning techniques for games are explained at the last part.
Youichiro Miyake ( Lead AI Researcher , SQUARE ENIX CO., LTD.) (miyakey@square-enix.com / y.m.4160@gmail.com) has been in development of video game titles while researching game AI technologies as the lead AI researcher at SQUARE ENIX. He developed the technical design of AI for the following game titles: CHROME HOUNDS, Demon's Souls, Armored Core V developed by FROM SOFTWARE, and FINAL FANTASY XIV, FINAL FANTASY XV, and Kingdom Hearts 3 developed by SQUARE ENIX. He is the chair of the SIG-AI in IGDA Japan, and also the board member of DiGRA Japan and the Society of Art and Science. He is the author of the books "AI in FINAL FANTASY XV" "Game AI methods for Digital Games" “Western Philosophy for Artificial Intelligence” (2016) "Eastern Philosophy for Artificial Intelligence" and "The theory of Game AI"(2018).
Ludii (https://ludii.games) is a general game system for creating, playing, evaluating and improving a variety of different game types, including board, card, dice, tile, hand, graph and mathematical games. Ludii game descriptions are more concise and human readable than games described in other formats such as GDL, making it easier to design, test and improve new games. The Ludii repository currently includes more than 200 game descriptions, many with additional variants, providing a rich source that users can extend and adapt into their own designs using its custom ludeme editor and visual interface.
Ludii provides a platform for evaluating AI agents over a range of game types that vary in complexity from Tic-Tac-Toe to Taikyoku Shogi, which is many orders of magnitude more complex than Go (this tutorial is being run in conjunction with the Ludii AI Competition at this year’s CoG). But Ludii is primarily a game design tool aimed at making it as easy as possible for users to model and test new game ideas and improve on old ones. It is being developed as part of the ERC-funded Digital Ludeme Project for the purpose of analysing historical games and improving reconstructions where possible, but has equal application to modern games as well.
This tutorial will be very hands-on. It will begin with a short demonstration of Ludii, including its basic usage, main features and functionality, and an introduction to its game description language. Several examples of using the system for game analysis and improvement will be demonstrated. Participants will be invited to experiment with the system to design and test their own games on-the-spot. We will conclude with a game design challenge — what is the best rule set that participants can devise with the help of Ludii given a certain set of equipment and constraints? — aimed to emulate the real-world reconstruction tasks that the system was designed to tackle.
Cameron Browne (Maastricht University) is an Associate Professor at Maastricht University's Department of Data Science and Knowledge Engineering (DKE), where he is running the €2m ERC-funded Digital Ludeme Project. Cameron received his PhD from the Queensland University of Technology (QUT) in 2009, winning a Dean’s Award for Outstanding Thesis and producing the world's first published computer-generated games. He is the author of the books Hex Strategy, Connection Games and Evolutionary Game Design, which won the 2012 GECCO “Humies” award for human-competitive results in evolutionary computation. He is a Section Editor of the IEEE Transactions on Games and the International Computer Games Association (ICGA) journal, and was the founder and Editor-in-Chief of Game & Puzzle Design journal. Email: cameron.browne@maastrichtuniversity.nl
Matthew Stephenson (Maastricht University) is a Postdoctoral Research Assistant at Maastricht University's Department of Data Science and Knowledge Engineering (DKE). Matthew received a PhD from the Research School of Computer Science of the Australian National University (ANU). His Thesis focused on procedural content generation techniques for physics-based game environments, particularly for the popular video game Angry Birds and the general video game AI (GVGAI) framework. His primary research interests and publications centre around various topics in AI for games, including general video game AI, procedural content generation, agent development, computational complexity analysis and deceptive game design. Email: matthew.stephenson@maastrichtuniversity.nl
In this tutorial, we introduce the communication game "Are You a Werewolf" (also known as "Werewolf" or "Mafia"), and present a hands-on tutorial for creating AI agents that can play this game. Participants of this tutorial will program their own Werewolf Agents, and compete against each other in the end.
Games, including board games, card games, and video games, have been used as test-beds for AI for a long time. Game-playing AI agents have been used to explore several technologies in Artificial Intelligence and Machine Learning. However, when humans play board or card games, they usually communicate with each other a lot. For example, in Settlers of Catan players usually engage in an informal session of promises, cajoling and negotiation before trading resources. In some games ("communication games") this communication is a central part of the rules. Werewolf is such a communication game, where the players have to communicate, convince, negotiate and even lie to each other.
Few studies in the Game-playing AI literature have analyzed these communication games. Creating an AI agent to play a game such as Werewolf require many skills that are not commonly used in other AI challenge games, such as communication skills, negotiation and communication, multi-agent coordination, reading intentions, and modeling the motivation of other players.
The AIWolf project proposes the Werewolf game as a new standard AI game challenge, to explore communication skills in a GAME AI context. Similar to RoboCup, the goal of this project is to create a communicative AI player that can play Werewolf with humans. As an initial step, we have designed a platform to allow competitions between Werewolf playing AI agents using either Natural Language or a specific protocol. We have used this platform to hold AI Werewolf competitions in Japan for the past 6 years, and in 2019 we held our first international competition with over 30 participants.
The goal of this tutorial is to increase the awareness of the AIWolf project among the CoG community, and welcome more people to this interesting and fun challenge. We will introduce the rules of the game as well as how to use our server platform and protocol. Then, using a sample agent as basis, we will help the tutorial participants to generate their own Werewolf AI agents. Finally, the participants will pit their agents against each other in a contest.
ATSUSHI TAKEDA, Graduate Student, University of Tokyo, takeda@torilab.net, 1st winner of last year's AIWolf competition. He will speak about how to program AIWolf agent in Java.
CLAUS ARANHA, Assistant Professor, University of Tsukuba, caranha@cs.tsukuba.ac.jp, Last year's AIWolf competition organizer (at ANAC on IJCAI 2019). He will speak about how to manage international competition of Werewolf game.
TAKASHI OTSUKI, Professor, Yamagata University, otsuki@yz.yamagata-u.ac.jp, C# platform programmer, and co-organizer of Werewolf protocol. He will speak about how to make standard AIWolf agent.
HIROTAKA OSAWA, Assistant Professor, University of Tsukuba, osawa@iit.tsukuba.ac.jp, Last year's competition organizer, and co-organizer of Werewolf protocol. He will speak about the history of Werewolf competition.
Link to the presenter/organizer web page or the tutorial page: https://aiwolf.org/
Game jams provides multiple levels of research opportunities, namely game development, team-based learning, community building, and as a method of research inquiry. Since the inception of game jams, especially the Global Game Jam, both participation rate and research interests have grown significantly. In this tutorial, we invite game jam enthusiasts to join us to brainstorm, share, and collaboratively write academic papers, and present to each other. Keywords: game jam, hackathon, collaborative research event
Game development activities and events (game jams) bring game enthusiasts worldwide together to make games. The Global Game Jam recorded tens of thousands of people creating unique games under forty-eight hours since 2009 [1].
Game jams provide for multidisciplinary teams, and inject creativity and individual learning opportunities without the risky shortcomings of typical game and software development projects [2]. Game jams have the potential to facilitate industry partnerships, the onboarding of students, and can open doors to new career pathways [3].
Early studies in game jams have touched on team-based collaboration work, creativity in action, and software production [4]. Since the introduction to game jams research, the types of experiments and inquiry have grown significantly. For instance, recent research experimented running game jams for sharing indigenous culture [5], team-based learning approaches in classrooms [6], and establishing a global curriculum for teaching game development skills to the next generation [7].
In this tutorial, we are conducting a ‘paper jam’ in the spirit of game jams. Specifically, we aim to facilitate a collaborative environment for game jam researchers to share research ideas and propose experimental methodologies and formats.
The format of this tutorial follows the paper jam held at International Conference on Game Jams, Hackathons and Game Creation Events (ICGJ) in 2019 [8], and the previous workshops held at Foundations of Digital Games (FDG) [9]. The structure of the tutorial will be as follows:
The scope of paper jam encompasses many academic and industry areas related to game development, including but not limited to: collaborative work, brainstorming, education and learning, art direction, sound design, community building, tools and processes, game jam reports, and research methodologies. While this is intended to be a paper jam, and not a game jam, participants may build experimental artefacts to explore their research ideas as a team. Near the end of the tutorial, each team will present to the room on what they have written.
Participants of the academic paper jam are expected to bring their writing instrument of choice, whether it is digital or analogue. The tutorial will require power and internet access, and a few large tables for drawing and sharing ideas. Whiteboards and standing poster walls will be essential for facilitating note taking in a group discussion. Finally, we will strive to host the tutorial in sun-lit environment early in the morning to get the most out of our collective brain powers.
Dr Xavier Ho is a Senior Software Engineer at CSIRO Data61, where he leads a team working on an interactive graph data visualisation tool. He is Lightning Talks Chair at ICGJ 2020, served as the FDG 2019 Poster Chair, and the GGJ 2017 Survey Chair. His research intersects game jams both as research and participation. Apart from his day job, he runs a freelancing videography business for tech meetups and latenight events, and he enjoys making generative coding art. He is a member of Archery Australia and have been practicing target archery casually for 2 years with recurve bows.
Dr Allan Fowler is the Chair of the International Conference on Games Jams, Hackathons and Game Creation Events. He has organized several workshops at the Foundations of Digital Games Conference (2019, 2015) and DiGRA (2013). Fowler is a published author, game designer, and dedicated educator. Apart from looking after his two gorgeous hobbits, in his spare time he makes games and practices martial arts and is a keen amateur photographer. Dr Fowler holds fifth degree black belts in both Gojuryu and Shorin-ryu karate and has competed in State and International Tournaments.
Links to the presenter/organizer web page or the tutorial page: https://aiwolf.org/
[1] “Global Game Jam History.” [Online]. Available: https://globalgamejam.org/history. [2] M. Cook, G. Smith, T. Thompson, J. Togelius, and A. Zook, “Hackademics: A Case for Game Jams At Academic Conferences,” in Proceedings of the 10th Foundation of Digital Games, 2015.
Automated game design is a rapidly growing field within game AI research that will have a huge impact on the future of game development. The budgets of the biggest games are continuing to swell, the numbers of small independent studios are on the rise, and more people than ever are experimenting with making games themselves. The future of all of these creative endeavours will be deeply impacted by the availability of intelligent tools that can teach, collaborate with and lead people as they create games.
Getting started in automated game design can be intimidating. Lots of the example systems are huge monoliths that are made up of lots of bespoke technology, which can be daunting. On top of this it can involve learning new game engines, finding new ways to distribute your work, thinking of new approaches to evaluation. Many of the field's core challenges and problems lie outside of the topics normally covered by game AI.
In this tutorial we'll look at the breadth of automated game design - surveying what's been done so far, looking at the current problems faced by research and by the games industry, and discussing uncharted territory that you could be the first to explore. In doing so, we'll not just look at the AI side of things, but also talk a little bit about what game design is, what kinds of problems game designers solve, and how AI can be expected to help in big and small ways.
The tutorial won't just be theoretical, but will also discuss and hopefully demonstrate some practical issues in automated game design, too. Even if you've never made a game before, we'll talk about what game engines you can use, how to build your own automated game design frameworks on top, and the kinds of genres and problems that make good starting points.
Attendees are encouraged to bring along a laptop, in order to take part in some practical aspects of the tutorial, but it is not compulsory. We hope to provide a fun experience with some inspiring examples, and to leave you feeling confident and encouraged to do research into automated game design!
Dr. Michael Cook is an AI researcher and game designer working at Queen Mary University of London. He currently holds a Royal Academy of Engineering research fellowship, investigating the application of automated game design to the games industry. He is best known for his research into automated game design, particularly ANGELINA the game-designing AI system, and as the organiser of PROCJAM, the Procedural Generation Jam. He is also the developer of the action-hacking game Rogue Process.
https://www.possibilityspace.org
This tutorial will cover numerous different representations for creating maps for use in games as instances of procedural content generation. The type of map located is strongly dependent on the representation used. Terrain maps, dungeon maps, mazes, and side-scroller levels will be used as examples. Representations covered will include direct representations, multiple types of generative representation, and will demonstrate how to create collections of fitness function by using techniques from artificial immune system technology. Terrain maps employ L-systems to permit complex multi-scale representation. Mazes representations will include highly scalable cellular automata technology.
Level maps will be demonstrated with required content and tile-decomposition techniques that enable the creation of enormous maps with great speed that contain objects required by the designer. The new \do what's possible" representation that can lay out thousands of rooms using an automaton described in a few hundred bytes will be demonstrated. The creation of multiple instances of a map with a controlled set of tactical properties will be covered in this tutorial, as will dual maps in which different types of barriers are combined to yield maps with distinct tactical properties for different types of agents.
The tutorial assumes novice level familiarity with evolutionary computation and familiarity with standard algorithms such as dynamic programming and A*. A theme within the tutorial is the design of fitness functions for maps, a place where the new artificial immune technology substantially simplifies design.
Professor Daniel Ashlock has 297 peer reviewed scientific publications, many of which deal with the issues of representation for games using evolutionary computation. He has presented tutorials and invited plenary lectures on representation at IEEE'sWCCI, CEC, CIG, and CIBCB conferences as well as at numerous universities. Dr. Ashlock serves as an associate editor for the IEEE Transactions on Evolutionary Computation, the IEEE Transactions on Games, and Biosystems. He is a member if the IEEE CIS Technical Committee on Bioinformatics and Biomedical Engineering and is a past chair of the IEEE Technical Committee on Games. Dr. Ashlock has written books on evolutionary computation, on the representation of game playing agents, and on representation for automatic content generation. Dr. Ashlock has served as the general chair of three IEEE Conferences, including CIG, and served on the program committee of more than twenty IEEE conferences. Daniel Ashlock Department of Mathematics and Statistics University of Guelph dashlock@uoguelph.ca https://eldar.mathstat.uoguelph.ca/dashlock/ (Vita) https://eldar.mathstat.uoguelph.ca/dashlock/vita.html
This is a lecture intended for the designer/artist/therapist of the interactive system for "wellness entertainment". In recent years, the fields of serious games and medical/welfare/rehabilitation applications are gaining attention in the field of entertainment computing. The presenter has been promoting R & D activities in this area for 30 years, and recently targeting "Wellness" or "Well-being" entertainment using bio-sensing technology. This lecture contains four parts of practical techniques/ideas to realize effective interactive system for "wellness entertainment" - environment & interface, concept of biofeedback, special techniques using commercial bio-sensors and cutting-edge topics.
The first part is short lecture of "Max8"(cycling'74) - famous and powerful platform/environments to develop realtime multimedia system with sensors. This "Sketching"(physical computing) lecture with many effective examples will help future development. The popular interface "Gainer" had retired few years ago, so this lecture also contains three types of "post-Gainer" new sensing/interface techniques using Arduino.
The second part is a lecture about "biofeedback and human wellness" related to "Wellness Entertainment" System Design. The idea of biofeedback is well-known in the rehabilitation and welfare field, and recent bio-sensing technology and multimedia technology enable to realize effective biofeedback system easily. Generally, human mental power for homeostasis works in emotion/feeling channel - which leads us "Wellness" or "Well-being" status. Especially, reaction information from unconscious muscles is gaining attention as "interoceptive sense" leading to emotion/feeling - this is the reason the organizer pays particular attention to the EMG sensing.
The third part is the special techniques using bio-sensing systems which are commercially available. (1) The "Myo" armband - three original techniques will be opened - how to use 2 or 3 Myo at the same time, how to avoid Myo's sleeping, and how to get directly data such as 8ch-EMG and 9-D data via Bluetooth. (2) The "MUSE" headband - how to get directly data such as 4ch-brain waves and 3D-direction via Bluetooth. (3) The combination application using "double Myo" and "MUSE" will be shown with the detail of Max8 programming - as a realtime interactive multimedia performance work in media art.
The fourth part is two cutting-edge bio-sensing topics. (1) The presenter and collaborator developed a novel 4ch EMG sensor system called "VPP-SUAC". The whole development data is open-source in the Web. While Myo is an "exclusive for arm" EMG sensor, our sensor focuses on sensing legs and lower body EMG which are important in welfare and rehabilitation area. (2) The other is a new special tactile/rubbing sensor called "PAW sensor", which is a small PCB with a cylinder of urethane foam on it. The output information of this sensor is four channel voltages which is time-shared conversion, which means the nuances of rubbing/touching the urethane foam with fingers. In this lecture, the attendees will experience the special eight sensors and its experimental application which generates 3-D realtime sounds/graphics for healing entertainment or prevention of dementia.
Yoichi Nagashima, composer/researcher/PE, was born in 1958 in Japan. He studied nuclear physics and music at Kyoto University. As the engineer of Kawai Musical Instruments, he developed some sound generator LSIs, and designed some electronic musical instruments, and produced musical softwares. From 1991, He has been the director of "Art & Science Laboratory" and produces many interactive tools of real-time music performance with sensor/MIDI. From 2000, he has been also the profrssor of SUAC(Shizouka University of Art and Culture), Faculty of Design, and teaches multi-media, computer music and media-art. He organized and was the General Chair of NIME04. He supported over 150 works and projects of interactive/multimedia installations from 2000, composed/performed many works of computer music, and organized/performed many lectures/workshops in many places all over the world. Yoichi Nagashima Professor, PE Department of Design, Shizuoka University of Art and Culture 1-1, chuo-2, Naka-ku, Hamamatsu, Shizuoka, 430-0906 JAPAN Tel : +81-53-457-6215 Fax : +81-53-457-6210 Email : nagasm@suac.ac.jp Lecture/Workshop and papers of Yoichi Nagashima (in English): https://nagasm.org/ASL/profile/
Presenter's web page:https://nagasm.org/ASL/ Related web page (in Japanese): https://nagasm.org/Sketching/ Tutorial page will be opened later. This is ICEC2018 workshop webpage by the presenter, and the CoG2020 tutorial will be arranged/shrinked from this workshop: https://nagasm.org/ICEC2018workshop/
"intermediate" - "advanced" The designer/researcher/specialist who experiences an interactive system that uses Max8, sensors, and Arduino is preferable.
- To know a lot of cases with interactive system design - To construct many relations at the same time in the system - To escape from the necessity to use PC for a part of the system - To unite the pursuit of the entertainment and the educational effect