Playing with games

Let's talk about digital games (computer games), their relation to play, their relation to you and to artificial intelligence. One way they could be divided up is the following: there are games you play, games you play with, games that play you and games that play with you. The first category is very common, the second rather common, the third a current research topic and the fourth an unknown territory awaiting (or perhaps provoking) new advances in artificial intelligence.

Games you play
These are the games that most people come to think of when you mention computer games, and that usually dominate sales charts. Games like Doom, Asteroids, Tetris, StarCraft, Angry Birds and Call of Duty. Games of progression (Juul) where there are goals and there is a clear sense of what it means to perform the game better or worse. Games where "instrumental play" (Taylor), i.e. playing the game like you were performing a sport, is possible (but not necessary). In these games, the rules are set and enforced by the game and it's up to you as a player to do your best according to what the game thinks you should do and within the possibility space the game affords.

It is very common that these games feature some sort of artificial intelligence, typically in the role of controlling the opponents and allies in order to provide a challenge. But there are also examples of AI-based dynamic difficulty adjustment and other forms of experience management, such as the AI Director in Left 4 Dead. Additionally, AI techniques can be used during the design and development phases of such games, especially for procedural content generation.

Games you play with
Some games have less explicit goals, less clear sense of progression and perhaps less structure overall. These are games which are sometimes describes as more of toys than games, and which when replayed from scratch may play out completely differently. Games of emergence, like Sim City, Minecraft and The Sims. As these games try to hard to make you feel like you can do whatever the hell you want, you are encouraged to express your own desires, dreams and whims using the mechanics of the game. Of course, no games have so far been made where you can actually do whatever you want; therefore, you are likely to run up against the limits of the simulation and game mechanics sooner or later. (Sooner, if you are of the "non serviam" player type.) Of course, most of the games that are meant to be played can be played with with varying degrees of success. For example, you can decide to build sculptures instead of clearing lines in Tetris, and you can enact little dances in Doom. The very fuzzy boundary beetwen games you play and games you play with can be further illustrated with how hard it is to classify Civilization or Skyrim as one or the other of these.

There are more potential roles for artificial intelligence in games you play with than there are in games you play. Because of the impossibility of scripting a free-form game, they are often crucially relying on a complicated simulation of the game world, many parts of which could be called AI. A good example is the use of "smart objects" that match characters' needs with provided services in The Sims. Given the risk of such free-form simulations getting out of hand due to emergent unbalance, there is also a role for experience managers here (e.g. causing an earthquake if your city is doing too well).

Games that play you
You can become a challenge for the game. That is, if the game sees as its objective to make you happy, to keep you challenged, to make you sad, to keep you playing or perhaps to make you stop playing. You might not be inclined to do as the game wants you to do, and then the game needs to manipulate you. Luckily (for the game) the game knows quite a lot about you, as you keep feeding it with information all the time, via the keyboard, mouse or gamepad interface you use to interact with the game. (The game might also have other nefarious ways to glean information about what you feel and think, such as watching what you do via a Kinect device or webcam, or measuring your stress level via galvanic skin response.) Thus, the game can take actions by introducing new elements, new levels, changing music and art, distributing rewards, changing difficulty level and behaviour of in-game characters etc, and measure the success of these actions by your behaviour. This is isomorphic to how you take actions such as jumping, shooting, looting and choosing dialogue options when you play a game, and measure the success of these actions through score, progress meters, achievements etc.

Games that play you is at the center of our AI research program here at IT University of Copenhagen. Georgios and me recently wrote a survey/position paper that outlines our view of how to achieve such games. We call our approach "Experience-driven Procedural Content Generation". This is a very AI-heavy approach, where various AI techniques are used both for modelling what you think and feel while playing the game, predicting what you will do under various circumstances, and creating new game content that will bring the game closer to its goal according to the derived model. An example of this approach is our work with a version Super Mario Bros with adaptive levels. This game wants to keep you entertained. Therefore it has let hundreds of players play pairs of different levels, and asked the players which of the levels were most entertaining. When you play it, it will match your play style with its model of player experience, and deliver a freshly created level which it thinks will keep you optimally entertained.

Games that play with you
Remember the games you used to play as a kid? Not the digital games or board games, but the ones where you never quite knew in advance what the game was about or what the rules were, because the basic premise was "to play". The ones where you went "let's pretend I'm an astronaut, and this is our spaceship" and your friend went "ok, but I'm a space knight and I ride next to your spaceship on my space horse!". Now that you're a grown-up you don't do this very often, except if you engage in improvisational theatre. In particular, you don't do this while playing games on your computer. The basic idea of a digital game is that you play against or with a stable set of rules that is more or less predictable and therefore "fair"; if you fail (or things play out differently than expected) it was your own fault or simply bad luck. The game enforces these rules with such strictness that anything that works is correct. Any change in the rules during the game session (such as removing your jump ability or making your friends your enemies) is carefully scripted.

There are some attempts to undermine the status of rules in computer games, for example the game B.U.T.T.O.N (see Wilson's article) where the rules are underspecified. Players are required to invent new rules, which are not observable or enforceable by the game, as they play. However, what this game does is simply to reduce the role of the computer in the game, and mixing in more of classic non-digital play. The game does not express any desire on its own, and is even more of a mindless automaton than most of the classic games you play. This is of course unappealing from the perspective of an AI aficionado interested in technological progress, as is the fact that the game needs to be played in physical space with other humans.

Imagine that we could replace these people with AI, and integrate this system into the game itself. Imagine a game that is actually playful. A game that wants to play with you, not because it wants to fulfil some sort of goal (keeping you playing, or entertained), but because it wants to keep itself entertained, and explore the possibility space of the play session together with you. A game that treats you as a means, not as an end. That could refuse to play with you if you are too boring or hard to play with. That gives you the feeling of really playing with someone, not just against a ruleset.

Unfortunately, we don't really know how yet how to create a game that plays with you. How can we implement a desire in the game for having fun and a lust for exploring the possibilities of interaction with a player? How can we implement the concrete mechanisms where the game co-creates the rules with you and plays along, both constructively and obstructively? Actually, this is not unfortunate at all. It means that we have a new research problem for AI and games, and yet another angle of attack on the elusive problem of artificial intelligence. For we could surely not have any real artificial intelligence that was not capable of playing and enjoying itself.

One idea for how to approach this problem is Schmidhuber's theory of artificial curiosity, where agents choose what to learn so that they optimise their predicted learning rate. There is also related work in evolutionary and developmental robotics. Another strand of work which could potentially inform the development of games that play with you is work in mixed-initiative procedural content generation, such as Tanagra and Sketchaworld. In these systems, the designer and the software take turns to work on the game content, typically so that the software acts as an resourceful and inventive assistant to the designer. Last year, I published a paper outlining a simple system that tries to do mixed-initiative rule generation while you're playing the game. Or perhaps you're playing the system, and the game emerges from the play session. The system tries to understand what rules you're playing according to, and then enforcing these rules, which might force you to play differently (especially if it has misunderstood you) and see the flaws in your implicit design concept. What is missing from this system is intentions and desires on part of the system; the game is still trying to please you, rather than itself. I should get to work right away on changing this.

This blog post was inspired in part by Sicart's article Against Procedurality, and the opposition it sets out between rule-centric play/interpretation and player-centric play/interpretation. I have nothing against any of these, and think that both are interesting. However, I think that it's important to point that just because something is procedural, it's not necessarily rigid, predictable and intentionally designed; it could be adaptive, emergent and have a will of its own.

Help us test new strategy game ideas

We have two new playtests/surveys going on, this time about strategy games. Those of us working on AI in the Center for Computer Games Research are interested in creating adaptive mechanisms for games, and automating the generation of various forms of game content. For strategy games, we are working on automatic map generation, automatic rule generation and AI techniques for competently playing strategy games with little domain information.

What we want from you is some help with evaluating the results of our recent endeavours.

The first survey is about automatically generated rulesets and general AI. If you take this survey, you will play a tutorial scenario and then two different (very short) scenarios in a strategy game you have never seen before. You will then be asked which one you prefer, which opponent AI was best, and a couple of other relevant questions. It will take you about 10 minutes. Click here to participate in the playtest survey.

You can also read the paper describing some of the previous research leading up to the system used in the current playtest. However, you are only allowed to read the paper after you've taken the survey. No cheating!

The second survey is about automatically generated maps for the classic real-time strategy game StarCraft. If you take this survey, you will be asked to look at ten different StarCraft maps and judge their relative qualities. It helps if you have played StarCraft, but it's not necessary. This survey will take you about 5 minutes to complete. Click here to take part in the StarCraft survey.

When you're finished, you could reward yourself with reading a paper on how the maps were created (though the current version of the system has evolved a bit). But only after you've taken the survey, or you'll bias the results...

Seriously though: thanks a lot for helping us with this!

On the Game AI versus traditional AI debate

Luke Dicken wrote a very nice blog post the other day on the differences between how AI is viewed in academia (and among students) and in the game industry. He focuses on the real-time requirements of game AI, and how little processing power most games make available for AI. He also talks about how NPC behaviour in games needs to be entertaining, not just high-performing, and reminds the reader that in striking contrast to e.g. AI for robots, game AI is allowed to cheat if it makes the game better.

Kevin dill wrote a response post where he points out that "traditional AI" is simply trying to solve very different problems than those faced by AI in games. Far from being simplistic and primitive, the techniques devised specifically within game AI are well suited to their specific purpose: reliably providing interesting NPC behaviour, while being understandable and moddable by designers. Perhaps academic AI research should take a hint or two from industrial game AI, rather than the opposite.

I am not fundamentally disagreeing with anything Luke or Kevin say. I think they both make several good points. However, I'd like to point out that I have a rather different and, I think, broader perspective on what game AI is. Both posts make implicit assumptions on what games are and what AI is, which I think are limiting.

The main game examples used by Luke and Kevin are Red Dead Redemption, Dragon Age, Battlefield Bad Company 2 and Left 4 Dead. While these are impressive games, they are all representatives of a pretty small subspace of gaming: AAA first-person story-driven games with a real-time component and graphics that require them to run on a home console or computer. They are also all targeted at a classic "hardcore gamer" audience. These are the sort of games that we typically talk about when discussing games, and these are the sort of games that everybody wants to work on. But not the games that most people play. It's like if in the automotive industry, everybody would want to work on the next Porsche, while most people drive a Toyota.

Bejeweled, FarmVille and Diner Dash don't have a first-person perspective, don't have complex graphics and don't have a story in the same sense as the games above. Yet as far as I know, they have more players than those games. Importantly, they don't have NPCs that need to be controlled by AI, but still they present a number of interesting AI problems. Even traditional hardcore strategy games like Civilization, StarCraft or Total War present hard AI problems which are only insufficiently solved by the techniques used in the game industry.

The other limiting assumption is that AI is used for controlling NPC behaviour. In fact, this is only one of many applications for the bountiful toolbox of techniques found in artificial intelligence. AI techniques can also be used to generate game content (levels, maps, rules, puzzles etc), model players, adapt various aspects of the game (such as the difficulty or the reward schedule), match players in online games, control artificial economy, debug game mechanics or implementations, and so on.

Between all the myriad types of games out there and the multitude of interesting AI problems within them, I feel there's more than enough to work on even for an academic like me. Real-time pathfinding and planning for FPS and RTS games is all great, and I look forward to playing the results, but I'm happy to see someone else doing that specific work.

If you're interested in the "other" game AI work I'm involved, you might want to read our recent survey papers on generating game content and on adapting games based on player models.

New Infinite TD playtest

Elvis has now done significant additional development on his Infinite Tower Defense game. Please help us by playing the game and answering

Infinite Tower Defence: please help us playtest the prototype

My student Elvis Alistar has developed a first version of his "Infinite Tower Defence" game, which aims to continuously creating personalized challenges in response to your playing style. The game is an application of experience-driven procedural content generation to tower defence games, a very popular yet relatively complex and cerebral genre of casual strategy games.

We are currently looking for playtesters. Please head over to http://itu.dk/people/elal/ and play a few rounds of the game, and then answer our questionnaire about it. Those who provide best feedback will be included in the credits for the game!

Second CfP for CIG 2011. Yes, the deadline has been extended!

Second call for papers -- deadline extended!
Call for tutorial proposals

2011 IEEE Conference on Computational Intelligence and Games

COEX, Seoul
South Korea, August 31-September 3, 2011

http://sclab.yonsei.ac.kr/~cig

Games have proven to be an ideal domain for the study of computational intelligence as not only are they fun to play and interesting to observe, but they provide competitive and dynamic environments that model many real-world problems. Additionally, methods from computational intelligence promise to have a big impact on game technology and development, assisting designers and developers and enabling new types of computer games.

The 2011 IEEE Conference on Computational Intelligence and Games brings together leading researchers and practitioners from academia and industry to discuss recent advances and explore future directions in this quickly moving field.

Topics of interest include, but are not limited to, the following:

* Learning in games
* Coevolution in games
* Neural-based approaches for games
* Fuzzy-based approaches for games
* Player/Opponent modeling in games
* CI/AI-based game design
* Multi-agent and multi-strategy learning
* Applications of game theory
* CI for Player Affective Modeling
* Intelligent Interactive Narrative
* Imperfect information and non-deterministic games
* Player satisfaction and experience in games
* Theoretical or empirical analysis of CI techniques for games
* Comparative studies and game-based benchmarking
* Computational and artificial intelligence in:
o Video games
o Board and card games
o Economic or mathematical games
o Serious games
o Augmented and mixed-reality games
o Games for mobile platforms

The conference will consist of a single track of oral presentations, tutorial and workshop/special sessions, and live competitions. The proceedings will be placed in IEEE Xplore, and made freely available on the conference website after the conference.

IMPORTANT DATES:
Tutorial and special session proposal deadline: March 15, 2011
Paper submission deadline: March 30, 2011 -- extended!
Decision notification: May 15, 2011
Camera-ready submission: June 15, 2011
Conference dates: August 31-September 3, 2011

General Chair : Sung-Bae Cho
Program Co-Chairs: Simon Lucas and Phllip Hingston
Competitions Chair: Julian Togelius
Publicity Chair: Clare Bates Congdon
Proceedings Chair: Mike Preuss
Tutorials and special sessions chair: Georgios Yannakakis
Local Chairs: Kyung-Joong Kim, Kyu-Baek Hwang, Eun-Youn Kim

CIG 2011

The 2011 IEEE Conference on Computational Intelligence and Games will run in Seoul, Korea, August 31-September 3.

This conference builds on the massive success of last year's conference, and we hope even to surpass the quantity and quality of that conference. So if you're at all interested in CI and/or AI in games, you need to be there.

Submission deadline March 15.