Thursday, June 16, 2016

On level generation in general, and a new competition

Procedural content generation has been a thing in games at least since Rogue and Elite in the early 80s. Plenty of games feature some kind of procedural generation, for example of levels, maps or dungeons. There's also lots of generation of more auxiliary things, such as skyboxes, trees and other kinds of vegetation, and patterns in general. While Spelunky relit the interest in PCG among indie game developers, AAA developers are increasingly looking at generating various parts of their games. There is also a very active research community on procedural content generation, with lots of papers published every year on new ways of generating things in games. We even wrote a book on this.

Anyway. You probably already knew this, given that you somehow found your way to this blog. What I'm going to say next is also rather obvious:

Essentially all PCG systems are limited to a single game. The level generator for Spelunky only generates Spelunky levels, and will not work with Civilization, or Diablo, or even Super Mario Bros. The Civilization map generator will not work with The Binding of Isaac, or Phoenix, or FTL, or... well, you get my point. In many cases the general algorithm (be it L-systems, binary space partition, diamond-square or something else) could be made to work on other games, but various amounts of game-specific engineering (hacking?) would be necessary; in other cases, the PCG method itself is mostly game-specific engineering and it's hard to discern a general algorithm.

Now, this is actually a problem. It is a problem because we want reusability. We don't want every game designer/developer to have to develop their own PCG method for each new game. Wouldn't it be great if there was software we could just grab, off the shelf to do level generation (or generation of some other kind) in your game? Even for those designers who see PCG as a canvas for creative expression, wouldn't it be great to have something to start with? Most game developers now use some kind of game engine, where physics, collision detection, rendering etc are available out of the box. Even some kinds of PCG is available this way, in particular vegetation through SpeedTree and similar packages. Why couldn't this be the case for level generation?

Let's make another analogy. In research on game-playing AI, there is a growing realization that working only on a single game has its limits. Trying to create champion-level players of Go, Poker, StarCraft or Unreal Tournament is in each case a worthy endeavor and the results are valuable and interesting, but at the same time the resulting solution tends be pretty domain-specific. The world's best Go AI is worthless at any other game than Go, and the same goes for all the other games in the list. There's simply a lot of game-specific engineering.

This is the problem that some recent competitions and frameworks are aiming to overcome. The General Game Playing Competition, the Arcade Learning Learning Environment and the General Video Game AI Competition (GVGAI) each focus on testing game-playing agents on multiple different games. There are many differences between their respective approaches, but also strong similarities.

Tying these threads together, what would it mean to create level generators (or other kinds of game content generators) that without modifications would create good content for a large number of different games? In other words, what would general level generation look like? This is not only a question of making game designers' lives easier (I am of course always interested in that; making game designers' lives easier, or replacing them), but also a very interesting AI and computational creativity problem in its own right.

In a new paper, General Video Game Level Generation, we explore this question. We design a framework based on the GVGAI framework, that allows level generators to connect to any game that is defined in the Video Game Description Language (the basis of GVGAI). The interface gives the level generator information about how the game works, and the level generator then returns a game level in a specified format. In the paper, we describe three different generators based on different approaches, and we test them using both computational means (which agents can play these levels) and through user studies with human players.

Better yet, we are not content with doing this ourselves. We want you to get involved too. That is why we are setting up a level generation track of the General Video Game AI Competition. The competition will run at the International Joint Conference on Artificial Intelligence this year, and we have extensive documentation on how to participate. It is easy to get started, and given how important the question of general level generation is, participating in the first ever competition on general level generation could be a very good investment of your efforts.

Looking forward to seeing what you will do with our framework and competition!

Monday, April 18, 2016

The differences between tinkering and research

Some of us have academic degrees and fancy university jobs, and publish peer-reviewed papers in prestigious journals. Let's call these people researchers. Some (many) others publish bots, hacks, experimental games or apps on blogs, web pages or Twitter while having day jobs that have little to do with their digital creative endeavors. Let's call those people tinkerers.

So what's the difference between researchers and tinkerers?

This is a valid question to ask, given that there are quite a few things that can be - and are - done by both researchers and tinkerers. Like creating deep neural nets for visual style transfer, creating funny Twitter bots, inventing languages for game description and generation, writing interactive fiction or developing Mario-playing AIs. These things have been done by people with PhDs and university affiliations, and they have been done by people who do this for a hobby. Anyone can download the latest deep learning toolkit, game engine or interactive fiction library and get cracking. So why is this called research when the academic does it, and just a curious thing on the Internet when a non-academic does it?

Let me start by outlining some factors that are not defining the difference between tinkering and research.

First of all, it's not about whether you work at a university and have a PhD. People can do excellent research without a PhD, and certainly not everything that a PhD-holder does deserves to called research.

Second, it's not because research always is more principled or has a body of theory supporting it. Nor is there typically a mathematical proof that the software will work or something like that. It's true that there are areas of computer science (and some other disciplines) where research progresses trough painstakingly proving theorems building on other theorems, and I have a lot of respect for such undertakings, but this has little to do with the more applied AI and computer science research I and most of my colleagues do. On a technical level, much of what we do is really not very different from tinkering. Some of it is good code, some of it bad, but typically there is a good (or at least interesting) idea in there.

Third, it's not really the publication venue either. It's true that most of us would trust a peer-reviewed paper in a good conference or journal more than something we find on a blog, and peer review has an important role to fulfill here. But what was once a sharp boundary is now a diffuse continuum, with a myriad of publication venues with different focus and different degrees of stringency. Quite a few papers make it through peer review even though they really shouldn't. Also, the traditional publication process is agonizingly slow, and many of us might just put something online right away instead of waiting until next year when the paper comes out. (I personally think it's prudent to always publish a peer-reviewed paper on substantial projects/artifacts I contribute to, but I sometimes put the thing itself online first.) It is also becoming more common to post preprints of papers on places such as arXiv as soon as they are done, and update them when/if the paper gets accepted into a journal or conference.

So we are back to square one. What, then, is the actual difference between tinkering and research? Let me list four differences, in order of decreasing importance: scholarship, testing, goals and persistence.


Probably the most importance difference between tinkering and research is scholarship. Researchers go out and find out about what other people have done, and then they build on that so they don't have to reinvent the wheel. Or if they do reinvent the wheel, they explain why they have to reinvent the wheel and how and why their wheel is different from all the other wheels out there. In other words, researchers put the thing they have done into context.

For example, almost seven years ago I made some experiments with evolving neural networks to play Super Mario Bros, and published a paper on this. The work (and paper) became fairly well-known in a smallish community, and a bunch of people built on this work in their own research (many of them managed to get better results than I did). Last year, some guy made an experiment with evolving neural networks for Super Mario Bros and made a YouTube video out of it. The video certainly reached more people on the internet than my work did; it makes no mention of any previous work. Seen as tinkering, that work and video is good work; seen as research, it is atrocious because of the complete lack of scholarship. The guy didn't even know he was reinventing the wheel, and didn't care to look it up. Which is fine, as it was probably not meant as research in the first place, and not sold as such.

Good scholarship is hard. It is easy to miss that someone tackled the same problem as you (or had the same idea as you) last year, or 5 years ago, or 50. People use different words to describe the same things, and publish in out-of-the-way places. Once you found the literature you must read it and actually understand it in order to see how it is similar to or differs from the idea you had. Because good scholarship is not just listing a number of previous works that are vaguely related to what you, but rather telling a believable, coherent and true story where all those previous works fits in, and your own work makes a logical conclusion. Therefore good scholarship takes a lot of searching and reading, a lot of time and effort. It's no wonder that lots of people don't want to spend the time and effort, and would rather get on with the tinkering.

It's common for incoming PhD students and other students to question the need for the scholarship when they could spend the time writing code. So let me go through some reasons for doing good scholarship, in order increasing importance. (Beyond wanting to get a PhD, of course.)

The perhaps most immediately important reason is common civility and courtesy. If you do a thing and tell the world about it, but "forget" to tell the world that person X did something very similar before you did it, then you are being rude to person X. You are insulting person X by not acknowledging the work she or he did. Academics are very sensitive to this, as proper attribution is their bread and butter. In fact, they will generally get offended even if the one you fail to cite other people than themselves. Therefore, the easiest way to get your papers rejected is to not do your related works section.

What about someone who doesn't care what academics think, or about getting published in peer-reviewed journals and conferences? Any point in spending all that time in front of Google Scholar and reading all that technical text written by academics with widely varying writing skills? Yes, obviously. Knowing what others have done, you can build on their work. Stand on the shoulders of giants, or at least atop a gang of midgets who have painstakingly crawled up on the shoulders of taller-than-average people. The more you know, the better your tinkering.

But to see the primary reason to do our scholarship before (or during, or after) tinkering we must lift our eyes beyond our own little fragile egos (yours and mine). It is about the accumulation of knowledge and progress on the scale of the human species. If we learn from each other, we can ultimately push the boundaries of what we collectively know forward and outward; if we don't learn from each other, we are bound to do the same things over and over. And it's way more likely that others will learn from you if you make it clear how what you are doing is different from (and maybe better than) what was done before.

So if you want your little hack or bot or whatnot to contribute to science, in other words to the evolution of humanity, you should do your scholarship.


Here's another big thing. A tinkerer makes a thing and puts it out there. A researcher also tests the thing in some way, and writes up what happens. Tests can take many shapes, as there are many things that can be tested - it depends on what you want to test. Generally the test is about characterizing the thing you made in some way. It could be performance on some sort of benchmark. Or a quantitative characterization with statistics from running your thing multiple times. Or maybe a user study. Or why not a qualitative study, where you really take your time to interact with your software and describe it in detail. The point is that if something is worth making, it's also worth studying and describing. If you don't study it when you're done, you're not learning as much as you could. And if you don't describe it well, nobody else will learn from it.

Interestingly, the tinkering and testing can sometimes be done by different people. There are quite a few academic papers out there that systematically study software that other people built but did not care to study in detail. This ranges from performance analysis of someone else's sorting algorithm, to large parts of the academic field of game studies.


Why do you tinker? Because of the thrill of trying something new? To hone your skills with some tool or programming language? To build useful tools for yourself or others? To get attention? To annoy people? Because you had an idea one night when you couldn't sleep? All of these are perfectly valid reasons, and I must confess to having had all those motivations at point or another.

However, if you read a scientific paper those are usually not the stated reasons for embarking on the research work presented in the paper. Usually, the work is said to be motivated by some scientific problem (e.g. optimizing real-value vectors in high-dimensional spaces, identifying faces in a crowd, generating fun game levels for Super Mario Bros). And that is often the truth, or at least part of the truth, from a certain angle.

While tinkering can be (and often is) done for the hell of it, research is meant to have some kind of goal. Now, it is not always the case that the goal was to get the result that was eventually reported. A key characteristic of research is that we don't really know what the results will be (which is why most grant applications are lies). Sometimes the result comes first, and the goal afterwards. Fleming did not set out to discover Penicillin, but once he did it was very easy to describe his research as solving an urgent problem. Also, he had been working on antibacterial compounds for a long time following different leads, so he recognized the importance of his discovery quickly.

Usually, goals in research are not just goals, but ambitious goals. The reason we don't know what the results of a research project will be is that the project is ambitious; no-one (as far we know) has attempted what we do before so our best guesses at what will happen are just that: guesses. If we understand the system so well that we can predict the results with high accuracy, chances are we are tinkering. Or maybe doing engineering.

Of the papers I've written, I think most of them started with some kind of problem I wanted to solve, so in other words a goal. But many others have been more opportunistic; we had a technology and an idea, and wanted to see what happened because... well, it sounded like a cool thing to do. Interestingly, I have never found it a problem to describe the research as if we had a particular goal in mind when we did it. This is probably because I always keep a number of high-level goals in mind, which implicitly or explicitly help me shape my research ides. This brings us to the next difference between research and tinkering.


You know Einstein's paper that established the special theory of relativity? A guy in his twenties, never having published anything before, publishing a single paper that revolutionized physics? Most papers are not like that.

Most papers report tiny steps towards grand goals. Results that are not in themselves very exciting, but hopefully will help us sometime in the future solve some problem which would be very exciting to solve. Like generating good video games from scratch, curing cancer or algorithms that understand natural language. The vast majority of such breakthroughs don't just happen - they are the results of sustained efforts over years or decades.  Recent progress we have seen in Go-playing builds on decades of research, even though it is sometimes reported as a sudden move by DeepMind.

Tinkerers are content to release something and then forget about it. Researchers carry out sustained efforts over a long time, where individual experiments and papers are part of the puzzle.

Doing research therefore requires having goals on different time scales in mind at any given time, and being able to extract high-level goals from level-goals, and seeing where new results fit into the bigger scheme of things. That is why I consider my attempts (together with colleagues) to chart out the research field and establish grand challenges as some of my most important. See, for example, our paper on challenges for procedural content generation, or on challenges for neuroevolution in games, or our attempt to structure all of research on AI in games.

Interestingly, when I started doing research I did not think I had much persistence at all. I also did not understand how much it was needed. Both of these realizations came later.


This post was inspired by my recent reading of Matti Tedre's "The Science of Computing", a history of the debates about what computer science actually is. He argues that computer science has variously been seen as mathematics, engineering and science, and that this debate has gone back and forth for as long as there has been computing researchers, with no clear end in sight. Reading the book, I felt that most of my research is not science, barely engineering and absolutely not mathematics. But I still think I do valuable and interesting research, so I set out to explain what I am doing.

The post was also inspired by discussions and arguments with a number of my colleagues, some of which have rather exclusionary ideas of what Science is and how Truth should be attained; and others who don't seem to think there's much difference between a blog post and a journal paper and who question the need to do all of the boring parts of research. I hope I have been able to make a good case for doing good research.

Friday, March 25, 2016

A way to deal with enormous branching factors

When you play Chess, there are on average something like 30 different moves you can make each turn. In other words, the branching factor is around 30. Go, on the other hand, has a branching of 300 or so. The standard Minimax algorithm, traditionally used to play Chess and similar board games, can't handle this branching factor very well. This is one of the two reasons why Go is so much harder for computers to play than Chess is, and why it took 19 years between the victory of AI over humans in Chess and the same victory in Go. (The other reason is the difficulty of state evaluation in Go.)

Let's think a bit about that number: 300. Apparently that's enough to stop Minimax (and Xerxes). Ultimately, Go was conquered with the help of Monte Carlo Tree Search (MCTS), which handles higher branching factors better than Minimax. But still, 300 is not a terribly big number.

Now consider this. Chess and Go are both games where you only move a single piece each turn. Many, perhaps most, turn based games are not like this. In board games such as Risk or Diplomacy you can move a number of units each turn. The same goes for computer strategy games such as Civilization, Advance Wars or XCOM. In trading card games such as Magic and Hearthstone you can play out a large number of cards each turn. Let's call these games multi-action games. (And let's not even get started on a game such as StarCraft, where you move lots of units in tandem but have the added complication of continuous time and space.) One might point out that such games model a number of real-world scenarios with varying degrees of accuracy, so an AI method for playing such games could be useful for non-game applications.

Hero Academy, a multi-action turn-based game with a characteristically huge branching factor.
What's the branching factor of a multi-action game? If you have 6 units to move, and you can move each unit to one of ten possible positions, you have a branching factor of 10 to the power of 6, or one million! Suddenly 300 does not feel like a very big number at all. And this is even simplifying things a bit, assuming that the order in which we move the units doesn't matter.

Obviously, Minimax is incapable of handling a branching factor of a million. A two-ply search means that a trillion positions will need to be examined. Even MCTS runs into very serious problems with this kind of branching factor. The basic problem is that you need to explore all the actions from the root node before doing anything else, and that might already take more time than you have that turn. (There are some clever ways of using MCTS in this space by considering individual unit actions, but that will have to be a topic for another time.)

At EvoStar 2016, we will present a paper that takes a fresh look at this problem and comes with a somewhat surprising solution. The paper is written by Niels Justesen, Tobias Mahlmann and myself; Niels was one of our star masters students at ITU Copenhagen, and Tobias and me supervised his masters thesis. We used a re-implementation of the strategy game Hero Academy as our testbed; Hero Academy is fairly representative of computer strategy games: each turn, a number of action points are available, and any action point can be used to move any character.

The basic idea of the paper is that the branching factor is so huge that we cannot do any tree search. Instead, we will have to concentrate on searching the space of possible action sequences that make up a single turn of the game. A simple state evaluation function will have to suffice when judging the quality of a turn–even though we cannot simulate the actions of the opponent, a good execution of each single turn might be all we need to play well. But even then, the branching factor is too high to search the possibilities of a single turn exhaustively. So we need to search smartly.

Therefore, we turned to evolution. We use an evolutionary algorithm to evolve sequences of actions that could make up a single turn. The fitness function is the quality of the board state of the end of the turn, as measured by some straightforward metrics. Both mutation and crossover are used to provide variation, as is common in evolutionary computation. We call the use of evolution to search actions within a turn Online Evolution.

The crossover scheme for Online Evolution in Hero Academy.
What about the results? Simply put, Online Evolution kicks ass. Evolution outperforms both a standard MCTS and two different greedy search methods by a wide margins. While it would certainly be possible to find MCTS versions that would handle this particular problem better, Online Evolution is at the very least highly competitive on this problem.

This is significant because the use of evolution for planning actions is very underexplored. While evolutionary algorithms have been used extensively to play games, it is almost always through evolving a program of some kind to play the game (for example through neuroevolution). The use of evolution in the actual playing phase seems to be almost virgin territory. We took some inspiration from the only other example of this general idea we know, Rolling Horizon Evolution for the PTSP game. But ultimately, there is no good reason that evolution could not be used to solve all the same problems that MCTS can. In fact, in his recent PhD thesis Cameron McGuinness showed that MCTS can be used to solve many of the same problems that evolution can. There seems to be some deeper connection here, which is well worth exploring.

Are you still with me? Good! Then you can go read our paper to get all the details!

Monday, March 21, 2016

Switching brains and putting the cart before the horse: EvoCommander, an experimental AI game

One of the best ways to make AI that is relevant to game development is to (1) make the AI and (2) make a game around the AI. That is, design a game that needs this particular flavor of AI to work.

To most people in the game industry (and far too many people in academia), this is the equivalent of putting the cart before the horse. The "proper" way to do game AI is to start with a requirement, a problem posed by the game's design, and then come up with some AI to solve this.

But many great innovations would never have been made if we insisted on putting the horse before the cart all the time. And making AI that solves the problems of existing game designs often leads to boring AI, because most games are designed to not need very interesting AI. It's more fun to turn things around - start with the AI and design a game that will need that AI, to prove that such a game design can be interesting.

I am definitely not the first person to think that thought. In particular, Ken Stanley has been involved in a couple of really interesting academic experiments in designing games around evolutionary neural networks, or neuroevolution. NERO is a game where you train the brains of a small army of troops, and then let your army fight other people's armies. Galactic Arms Race (with Erin Hastings as main developer) revolves around picking up and learning to use bizarre weaponry, which is evolved based on players' choices. Petalz (the offspring of my good friend Sebastian Risi and others) is a social network game about flowers, powered by actual evolution. I've been involved in a couple of such attempts myself, such as Infinite Tower Defense which uses several different AI mechanisms to adapt to the player's strategy and preferences, creating an ever-shifting challenge.

Of course, there are several examples of commercial games that seem to have been built partly to showcase interesting AI as well, including classics such as Creatures and Black and White. And there are many games that have chosen non-standard, AI-heavy solutions to design problems. But it would take us too far to dig deeper into those games, as I think it's about time that I get to the reason I wrote this blog post.

The reason I wrote this blog post is EvoCommander (game, paper). EvoCommander is a game designed and implemented by Daniel Jallov, while he was a student at ITU Copenhagen under the supervision of Sebastian Risi and myself; we also contributed to the game design.

A mission being played in EvoCommander.

EvoCommander's game play revolves around training your little combat robot, and then unleashing it against human or computer-controlled enemies. The robot's brain is a neural network, and training happens through neuroevolution. You train the robot by giving it a task and deciding what it gets rewarded for; for example, you can reward it for approaching the enemy, using one of its weapons, or simply keeping distance; you can also punish if for any of these things. Like a good dog, your little robot will learn to do things so as to maximize reward and minimize punishment, but these things are not always what you had in mind when you decided what to reward. Unlike a game like Pokemon, where "training" is simply progression along a pretermined skill trajectory, in EvoCommander training really is an art, with in principle limitless and open-ended outcomes. In this regard, the premise of the game resembles that of NERO (mentioned above).

Fierce PvP battle in the EvoCommander arena.
A key difference to that game, and also a key design innovation in EvoCommander, is the brain switching mechanic. You can train multiple different "brains" (neural networks) for different behaviors, some of which may be attacking tactics, others tactics for hiding behind walls etc. When battling an opponent, you can then decide which brain to use at each point in time. This gives you constant but indirect control over the robot. It also gives you considerable leeway in selecting your strategy, both in the training phase and the playing phase. You may decide to train complicated generic behaviors (remember that you can start training new brains from any brain you trained so far) and only switch brains rarely. Or you may train brains that only do simple things, and use brain switching as a kind of macro-action, a bit like a combo move in Street Fighter.

The robot bootcamp, where you train your brains.
As an experimental research game, EvoCommander is not as polished as your typical commercial game. However, that is not the point. The point is to take an interesting AI method and show how it can be the basis for a game design, and in the process invent a game design that would not be possible without this AI method.

You are welcome play the game and/or read the paper yourself to find out more!

Further reading: I've written in the past about why academia and game industry don't always get along, and strategies for overcoming this. Building AI-based games to show how interesting AI can be useful in games is one of my favorite strategies. An analysis (and idea repository) for how AI can be used in games can be found in our recent paper on AI-based game design patterns.

Monday, January 25, 2016

Some rules for the modern scientist

If you don't make your papers publicly available on your webpage, you don't want people to read them. Making them available means providing a link to a downloadable version in pdf format, preferably hosted on your own web site. You cannot blame anyone for not reading and citing your work if it cannot very easily be found online without a paywall. All respectable academic publishers allow self-archiving of publications these days; if they don't allow it, they are not respectable.

If you don't have a Google Scholar profile, you don't want to be hired. Like it or not, citation counts and h-index are among the least bad metrics for evaluating researchers. And like it or not, metrics are necessary because people on hiring committees do not really have the time to read your papers on detail, nor do they typically have the necessary context to understand their significance.

If you never try to describe your research in more popular terms than an academic paper, you don't want anyone outside your field to know about what you are doing. All academic fields are full of jargon which acts as an effective deterrent to non-specialists understanding your work. Every so often, try to give a public lecture, write a blog post or record a video that explains your work so that ordinary people would understand it. Skip the jargon. Your target audience could be your high school friends who became hair dressers or car mechanics, or why not your parents. Don't forget to ask someone from your target audience if they understood what you wrote or talked about. You will learn a lot.

If you don't network with your research colleagues on Facebook and/or Twitter, you don't really care about keeping up to date with what happens in your research community. Conferences happen only a few times per year, and research happens faster than that. Mailing lists are mainly used for calls for papers. Your peers will talk about events, papers, ideas, results online. To stay relevant, you need to keep up to date on what's happening. If you desperately want to keep your private life online separate from your professional life, create alternate social network accounts for professional networking. But really, if you are passionate about your research, why would you want to keep it separate from your private life?

Monday, January 11, 2016

Why video games are essential for inventing artificial intelligence

The most important thing for humanity to do right now is to invent true artificial intelligence (AI): machines or software that can think and act independently in a wide variety of situations. Once we have artificial intelligence, it can help us solve all manner of other problems.

Luckily, thousands of researchers around work on inventing artificial intelligence. While most of them work on ways of using known AI algorithms to solve new problems, some work on the overarching problem of artificial general intelligence. I do both. As I see it, addressing applied problems spur the invention of new algorithms, and the availability of new algorithms make it possible to address new problems. Having concrete problems to try to solve with AI is necessary in order to make progress; if you try to invent AI without having something to use it for, you will not know where to start. My chosen domain is games, and I will explain why this is the most relevant domain to work on if you are serious about AI.

I talk about this a lot. All the time, some would say.

But first, let us acknowledge that AI has gotten a lot of attention recently, in particular work on "deep learning" is being discussed in mainstream press as well as turned into startups that get bought by giant companies for bizarre amounts of money. There have been some very impressive advances during the past few years in identifying objects in images, understanding speech, matching names to faces, translating text and other such tasks. By some measures, the winner of the recent ImageNet contest is better than humans at correctly naming things in images; sometimes I think Facebook's algorithms are better than me at recognizing the faces of my acquaintances.
Example image from the ImageNet competition. "Deep" neural networks have been made to learn to label images extremely well lately. Image courtesy of

With few exceptions, the tasks that deep neural networks have excelled at are what are called pattern recognition problems. Basically, take some large amount of data (an image, a song, a text) and output some other (typically smaller) data, such as a name, a category, another image or a text in another language. To learn to do this, they look at tons of data to find patterns. In other words, the neural networks are learning to do the same work as our brain's sensory systems: sight, hearing, touch and so on. To a lesser extent they can also do some of the job of our brain's language centra.

However, this is not all that intelligence is. We humans don't just sit around and watch things all day. We do things: solve problems by taking decisions and carrying them out. We move about and we manipulate our surroundings. (Sure, some days we stay in bed almost all day, but most of the rest of the time we are active in one way or another.) Our intelligence evolved to help us survive in a hostile environment, and doing that meant both reacting to the world and planning complicated sequences of actions, as well as adapting to changing circumstances. Pattern recognition - identifying objects and faces, understanding speech and so on - is an important component of intelligence, but should really be thought of as one part of a complete system which is constantly working on figuring out what to do next. Trying to invent artificial intelligence while only focusing on pattern recognition is like trying to invent the car while only focusing on the wheels.

In order to build a complete artificial intelligence we therefore need to build a system that takes actions in some kind of environment. How can we do this? Perhaps the most obvious idea is to embody artificial intelligence in robots. And indeed, robotics has shown us how even the most mundane tasks, such as walking in terrain or grabbing strangely shaped objects, are really rather hard to accomplish for robots. In the eighties, robotics research largely refocused on these kind of "simple" problems, which led to progress in applications as well as a better understanding of what intelligence is all about. The last few decades of progress in robotics has fed into the development of self-driving cars, which is likely to become one of the areas where AI technology will revolutionize society in the near future.

Once upon a time, when I thought I was a roboticist of sorts. The carlike robot in the foreground was supposed to learn to independently drive around the University of Essex campus (in the background). I think it's fair to say we underestimated the problem. (From left to right: me (now at NYU), Renzo de Nardi (Google), Simon Lucas (Essex), Richard Newcombe (Facebook/Oculus), Hugo Marques (ETH Zurich).)

Now, working with robots clearly has its downsides. Robots are expensive, complex and slow. When I started my PhD, my plan was to build robot software that would learn evolutionarily from its mistakes in order to develop increasingly complex and general intelligence. But I soon realized that in order for my robots to learn from their experiences, they would have to attempt each task thousands of times, with each attempt maybe taking a few minutes. This meant that even a simple experiment would take several days - even if the robot would not break down (it usually would) or start behaving differently as the batteries depleted or motors warmed up. In order to learn any more complex intelligence I would have to build an excessively complex (and expensive) robot with advanced sensors and actuators, further increasing the risk of breakdown. I also would have to develop some very complex environments where complex skills could be learned. This all adds up, and quickly becomes unmanageable. Problems such as these is why the field of evolutionary robotics has not scaled up to evolve more complex intelligence.
The basic idea of evolutionary robotics: use Darwinian evolution implemented as a computer algorithm to teach robot AIs (for example neural networks) to solve tasks. Image taken from a recent survey paper.

I was too ambitious and impatient for that. I wanted to create complex intelligence that could learn from experience. So I turned to video games.

Games and artificial intelligence have a long history together. Even since before artificial intelligence was recognized as a field, early pioneers of computer science wrote game-playing programs because they wanted to test whether computers could solve tasks that seemed to require "intelligence". Alan Turing, arguably the principal inventor of computer science, (re)invented the Minimax algorithm and used it to play Chess. (As no computer had been built yet, he performed the calculations himself using pen and paper.) Arthur Samuel was the first to invent the form of machine learning that is now called reinforcement learning; he used it in a program that learned to play Checkers by playing against itself. Much later, IBM's Deep Blue computer famously won against the reigning grandmaster of Chess, Gary Kasparov, in a much-publicized 1997 event. Currently, many researchers around the world work on developing better software for playing the board game Go, where the best software is still no match for the best humans.

Arthur Samuel in 1957, playing checkers against a mainframe computer the size of a room and thousands of times less computer power than your phone. The computer won. 

Classic board game such as Chess, Checkers and Go are nice and easy to work with as they are very simple to model in code and can be simulated extremely fast - you could easily make millions of moves per second on a modern computer - which is indispensable for many AI techniques. Also, they seem to require thinking to play well, and have the property that they take "a minute to learn, but a lifetime to master". It is indeed the case that games have a lot to do with learning, and good games are able to constantly teach us more about how to play them. Indeed, to some extent the fun in playing a game consists in learning it and when there is nothing more to learn we largely stop enjoying them. This suggests that better-designed games are also better benchmarks for artificial intelligence. However, judging from the fact that now have (relatively simple) computer programs that can play Chess better than any human, it is clear that you don't need to be truly, generally intelligent to play such games well. When you think about it, they exercise only a very narrow range of human thinking skills; it's all about turn-based movements on a discrete grid of a few pieces with very well-defined, deterministic behavior.

Deep Blue beats Kasparov in Chess.

But, despite what your grandfather might want you to believe, there's more to games than classical board games. In addition to all kinds of modern boundary-pushing board games, card games and role-playing games, there's also video games. Video games owe their massive popularity at least partly to that they engage multiple senses and multiple cognitive skills. Take a game such as Super Mario Bros. It requires you not only to have quick reactions, visual understanding and motoric coordination, but also to plan a path through the level, decide about tradeoffs between various path options (which include different risks and rewards), predict the future position and state of enemies and other characters of the level, predict the physics of your own running and jumping, and balance the demands of limited time to finish the level with multiple goals. Other games introduce demands of handling incomplete information (e.g. StarCraft), understanding narrative (e.g. Skyrim), or very long-term planning (e.g. Civilization).

A very simple car racing game I hacked up for a bunch of experiments that went into my PhD thesis. We also used it for a competition.

On top of this, video games are run inside controllable environments in a computer and many (though not all) video games can be sped up to many times the original speed. It is simple and cheap to get started, and experiments can be run many thousand times in quick succession, allowing the use of learning algorithms.

After the first year, the Simulated Car Racing Competition switched to the TORCS racing game, which is more advanced. It also undeniably looks better.

So it is not surprising that AI researchers have increasingly turned to video games as benchmarks recently. Researchers such as myself have adapted a number of video games to function as AI benchmarks. In many cases we have organized competitions where researchers can submit their best game-playing AIs and test them against the best that other researchers can produce; having recurring competitions based on the same game allows competitors to refine their approaches and methods, hoping to win next year. Games for which we have run such competitions include Super Mario Bros (paper), StarCraft (paper), the TORCS racing game (paper), Ms Pac-Man (paper), a generic Street Fighter-style figthing game (paper), Angry Birds (paper), Unreal Tournament (paper) and several others. In most of these competitions, we have seen performance of the winning AI player improve every time the competition is run. These competitions play an important role in catalyzing research in the community, and every year many papers are published where the competition software is used for benchmarking some new AI method. Thus, we advance AI through game-based competitions.

There's a problem with the picture I just painted. Can you spot it?

That's right. Game specificity. The problem is that improving how well an artificial intelligence plays a particular game is not necessarily helping us improve artificial intelligence in general. It's true that in most of the game-based competitions mentioned above we have seen the submitted AIs get better every time the competition ran. But in most cases, the improvements were not because of better AI algorithms, but because of even more ingenious ways of using these algorithms for the particular problems. Sometimes this meant relegating the AI to a more peripheral role. For example, in the car racing competition the first years were dominated by AIs that used evolutionary algorithms to train a neural network to keep the car on the track. In later years, most of the best submissions used hand-crafted "dumb" methods to keep the car on the track, but used learning algorithms to learn the shape of the track to adapt the driving. This is a clever solution to a very specific engineering problem but says very little about intelligence in general.
The game Ms. Pac-Man was used in a successful recent AI competition, which saw academic researchers, students and enthusiasts from around the world submit their AI Pac-Man players. The challenge of playing Pac-Man spurred researchers to invent new versions of state-of-the-art algorithms to try to get ahead. Unfortunately, the best computer players so far are no better than an intermediate human.
In order to make sure that what such a competition measures is anything approaching actual intelligence, we need to recast the problem. To do this, it's a great idea to define what it is we want to measure: general intelligence. Shane Legg and Marcus Hutter have proposed a very useful definition of intelligence, which is roughly the average performance of an agent on all possible problems. (In their original formulation, each problem's contribution to the average is weighed by its simplicity, but let's disregard that for now.) Obviously, testing an AI on all possible problems is not an option, as there are infinitely many problems. But maybe we could test our AI on just a sizable number of diverse problems? For example on a number of different video games?

The first thing that comes to mind here is to just to take a bunch of existing games for some game console, preferably one that could be easily emulated and sped up to many times real time speed, and build an AI benchmark on them. This is what the Arcade Learning Environment (ALE) does. ALE lets you test your AI on more than a hundred games released for 70s vintage Atari 2600 console. The AI agents get feeds of the screen at pixel level, and have to respond with a joystick command. ALE has been used in a number of experiments, including those by the original developers of the framework. Perhaps most famously, Google Deep Mind published a paper in Nature showing how they could learn to play several of the games with superhuman skill using deep learning (Q-learning on a deep convolutional network).

The Atari 2600 video game console. Note the 1977-vintage graphics of the Combat game on the TV in the background. Note, also, the 1977-vintage wooden panel. Your current game console probably suffers from a lack of wooden panels. Not in picture: 128 bytes of memory. That's a few million times less than your cell phone.

ALE is an excellent AI benchmark, but has a key limitation. The problem with using Atari 2600 games is that there is only a finite number of them, and developing new games is a trick process. The Atari 2600 is notoriously hard to program, and the hardware limitations of the console tightly constrain what sort of games can be implemented. More importantly, all of the existing games are known and available to everyone. This makes it possible to tune your AI to each particular game. Not only to train your AI for each game (DeepMind's results depend on playing each game many thousands of times to train on it) but to tune your whole system to work better on the games you know you will train on.

Can we do better than this? Yes we can! If we want to approximate testing our AI on all possible problems, the best we can do is to test it on a number of unseen problems. That is, the designer of the AI should not know which problems it is being tested on before the test. At least, this was our reasoning when we designed the General Video Game Playing Competition.

"Frogger" for the Atari 2600.

The General Video Game Playing Competition (GVGAI) allows anyone to submit their best AI players to a special server, which will then use them to play ten games that no-one (except the competition organizers) have seen before. These games are of the type that you could find on home computers or in arcades in the early eighties; some of them are based on existing games such as Boulder Dash, Pac-Man, Space Invaders, Sokoban and Missile Command. The winner of the competition is the AI that plays these unseen games best. Therefore, it is impossible for the creator of the AI to tune their software to any particular game. Around 50 games are available for training your AI on, and every iteration of the competition increases this number as the testing games from the previous iteration become available to train on.

"Frogger" implemented in VGDL in the General Video Game Playing framework.

Now, 50 games is not such a large number; where do we get new games from? To start with, all the games are programmed in something called the Video Game Description Language (VGDL). This is a simple language we designed to to be able to write games in a compact and human-readable way, a bit like how HTML is used to write web pages. The language is designed explicitly to be able to encode classical arcade games; this means that the games are all based on the movement of and interaction sprites in two dimensions. But that goes for essentially all video games before Wolfenstein 3D. In any case, the simplicity of this language makes it very easy to write new games, either from scratch or as variations on existing games. (Incidentally, as an offshoot of this project we are exploring the use of VGDL as a prototyping tool for game developers.)

A simple version of the classic puzzle game Sokoban implemented in VGDL.

Even if it's simple to write new games, that doesn't solve the fundamental problem that someone has to write them, and design them first. For the GVG-AI competition to reach its full potential as a test of general AI, we need an endless supply of new games. For this, we need to generate them. We need software that can produce new games at the press of a button, and these need to be good games that are not only playable but also require genuine skill to win. (As a side effect, such games are likely to be enjoyable for humans.)

Boulder Dash implemented in VGDL in the General Video Game Playing framework.

I know, designing software that can design complete new games (that are also good in some sense) sounds quite hard. And it is. However, I and a couple of others have been working on this problem on and off for a couple of years, and I'm firmly convinced it is doable. Cameron Browne has already managed to build a complete generator for playable (and enjoyable) board games, and some of our recent work has focused on generating simple VGDL games, though there is much left to do. Also, it is clearly possible to generate parts of games, such as game levels; there has been plenty of research within the last five years on procedural content generation - the automatic generation of game content. Researchers have demonstrated that methods such as evolutionary algorithms, planning and answer set programming can automatically create levels, maps, stories, items and geometry, and basically any other content type for games. Now, the research challenges are to make these methods general (so that they work for all games, not just for a particular game) and more comprehensive, so that they can generate all aspects of a game including the rules. Most of the generative methods include some form of simulation of the games that are being generated, suggesting that the problems of game playing and game generation are intricately connected and should be considered together whenever possible.
Yavalath, a board game designed by computer program designed by Cameron Browne. Proof that computers can design complete games.
Once we have extended the General Video Game Playing Competition with automated game generation, we have a much better way of testing generic game-playing ability than we have ever had before. The software can of course also be used outside of the competition, providing a way to easily test the general intelligence of game-playing AI.

So far we have only talked about how to best test or evaluate the general intelligence of a computer program, not how to best create one. Well, this post is about why video games are essential for inventing AI, and I think that I have explained that pretty well: they can be used to fairly and accurately benchmark AIs. But for completeness, let us consider which are the most promising methods for creating AIs of this kind. As mentioned above, (deep) neural networks have recently attracted lots of attention because of some spectacular results in pattern recognition. I believe neural networks and similar pattern recognition methods will have an important role to play for evaluating game states and suggesting actions in various game states. In many cases, evolutionary algorithms are more suitable than gradient-based methods when training neural networks for games.

But intelligence can not only be pattern recognition. (This is for the same reason that behaviorism is not a complete account of human behavior: people don't just map stimuli to responses§, sometimes they also think.) Intelligence must also incorporate some aspect of planning, where future sequences of actions can be played out in simulation before deciding what to do. Recently an algorithm called Monte Carlo Tree Search, which simulates the consequences of long sequences of actions by doing statistics of random actions, has worked wonders on the board game Go. It has also done very well on GVGAI. Another family of algorithms that has recently shown great promise on game planning tasks is rolling horizon evolution. Here, evolutionary algorithms are used not for long-term learning, but for short-term action planning.

I think the next wave of advances in general video game-playing AIs will come from ingenious combinations of neural networks, evolution and tree search. And from algorithms inspired by these methods. The important thing is that both pattern recognition and planning will be necessary in various different capacities. As usual in research, we cannot predict what will work well in the future (otherwise it wouldn't be called research), but I bet that exploring various combinations of these method will inspire the invention of the next generation of AI algorithms.

A video game such as The Elder Scrolls V: Skyrim requires a wide variety of cognitive skills to play well.

Now, you might object that this is a very limited view of intelligence and AI. What about text recognition, listening comprehension, storytelling, bodily coordination, irony and romance? Our game-playing AIs can't do any of this, no matter if it can play all the arcade games in the world perfectly. To this I say: patience! None of these things are required for playing early arcade games, that is true. But as we master these games and move on to include other genres of games in our benchmark, such as role-playing games, adventure games, simulation games and social network games, many of these skills will be required to play well. As we gradually increase the diversity of games we include in our benchmark, we will also gradually increase the breadth of cognitive skills necessary to play well. Of course, our game-playing AIs will have to get more advanced to cope. Understanding language, images, stories, facial expression and humor will be necessary. And don't forget that closely coupled with the challenge of general video game playing is the challenge of general video game generation, where plenty of other types of intelligence will be necessary. I am convinced that video games (in general) challenges all forms of intelligence except perhaps those closely related to bodily movement, and therefore that video game (in general) is the best testbed for artificial intelligence. An AI that can play almost any video game and create a wide variety of video games is, by any reasonable standard, intelligent.

"But why, then, are not most AI researchers working on general video game playing and generation?"
To this I say: patience!

A game such as Civilization V requires a different, but just as wide, skillset to play well.

This blog post became pretty long - I had really intended it to be just a fraction of what it is. But there was so much to explain. In case you've read this far, you might very well have forgotten what I said in the beginning by now. So let me recapitulate:

It is crucial for artificial intelligence research to have good testbeds. Games are excellent AI testbeds because they pose a wide variety of challenges, similarly to robotics, and are highly engaging. But they are also simpler, cheaper and faster, permitting a lot of research that is not practically possible with robotics. Board games have been used in AI research since the field started, but in the last decade more and more researchers have moved to video games because they offer more diverse and relevant challenges. (They are also more fun.) Competitions play a big role in this. But putting too much effort into AI for a single game has limited value for AI in general. Therefore we created the General Video Game Playing Competition and its associated software framework. This is meant to be the most complete game-based benchmark for general intelligence. AIs are evaluated on playing not a single video game, but on multiple games which the AI designer has not seen before. It is likely that the next breakthroughs in general video game playing will come from a combination of neural networks, evolutionary algorithms and Monte Carlo tree search. Coupled with the challenge of playing these games, is the challenge of generating new games and new content for these games. The plan is to have an infinite supply of games to test AIs on. While playing and generating simple arcade games tests a large variety of cognitive capacities - more diverse than any other AI benchmark - we are not yet at the stage where we test all of intelligence. But there is no reason to think we would not get there, given the wide variety of intelligence that is needed to play and design modern video games.

If you want to know more about these topics, I've linked various blog posts, articles and books from the text above. Most of the research I'm doing (and that we do in the NYU Game Innovation Lab) is in some way connected to the overall project I've been describing here. I recently put together a little overview of research I've done in the past few years, with links to most of my recent papers. Many more papers can be found on my web page. It is also important to realize that most of this work has a dual purpose: to advance artificial intelligence and to make games more fun. Many of the technologies that we are developing are to a lesser or greater extent focused on improving games. It's important to note that there is still important work to be done in advancing AI for particular games. In another recent blog post, I tried to envision what video games would be like if we had actual artificial intelligence. A somewhat recent paper I co-wrote tries to outline the whole field of AI in games, but it's rather long and complex so I would advise you to read the blog post first. You can also peruse the proceedings of the Computational Intelligence and Games and Artificial Intelligence and Interactive Digital Entertainment conference series.

Finally, it's important to note that there is plenty of room to get involved in this line of research (I think of it all as an overarching meta-project) as there are many, many open research questions. So if you are not already doing research on this I think you should start working in this field. It's exciting, because it's the future. So what are you waiting for?

Tuesday, November 17, 2015

Neuroevolution in games

Neuroevolution - the evolution of weights and/or topology for neural networks - is a common and powerful method in evolutionary robotics and machine learning. In the last decade or so, we have seen a large number of applications of neuroevolution in games. Evolved neural networks have been used to play games, model players, generate content and even enable completely new game genres. To some extent, games seem to be replacing the small mobile robots ubiquitous in evolutionary robotics and simple benchmarks used in reinforcement learning research.

Sebastian Risi and I have written a survey on neuroevolution in games, including a discussion of future research challenges. The main reason is that there was no survey of neuroevolution in games in existence; the other reason was that we wanted a tutorial overview to hand out to the students in our Modern AI for Games course.

A while back we asked the community to send us comments and suggestions for important work we might have overlooked. We received a lot of useful input and incorporated most of the suggested work. Thank you for very much for the help!

Now we are happy to announce that the paper will finally be published in the IEEE Transactions on Computational Intelligence and AI in Games (TCIAIG). We hope you like it!

TCIAIG early access:

A preprint of the manuscript is available here:

Wednesday, October 07, 2015

What if videogames had actual AI?

Is there any artificial intelligence in a typical videogame? Depends on what you mean. The kind of AI that goes into games typically deal with pathfinding and expressing behaviors that were designed by human designers. The sort of AI that we work on in university research labs is often trying to achieve more ambitious goals, and therefore often not yet mature enough to use in an actual game. This article has an excellent discussion of the difference, including a suggestion (from Alex Champandard) that the "next giant leap of game AI is actually artificial intelligence". And there's indeed lots of things we could do in games if we only had the AI techniques to do it.

So let's step into the future, and assume that many of the various AI techniques we are working on at the moment have reached perfection, and we could make games that use them. In other words, let's imagine what games would be like if we had good enough AI for anything we wanted to do with AI in games. Imagine that you are playing a game of the future.

You are playing an "open world" game, something like Grand Theft Auto or Skyrim. Instead of going straight to the next mission objective in the city you are in, you decide to drive (or ride) five hours in some randomly chosen direction. The game makes up the landscape as you go along, and you end up in a new city that no human player has visited before. In this city, you can enter any house (though you might have to pick a few locks), talk to everyone you meet, and involve yourself in a completely new set of intrigues and carry out new missions. If you would have gone in a different direction, you would have reached a different city with different architecture, different people and different missions. Or a huge forest with realistic animals and eremites, or a secret research lab, or whatever the game engine comes up with.

Talking to these people you find in the new city is as easy as just talking to the screen. The characters respond to you in natural language that takes into account what you just said. These lines are not read by an actor but generated in real-time by the game. You could also communicate with the game though waving your hands around, dancing, exhibiting emotions or other exotic modalities. Of course, in many (most?) cases you are still pushing buttons on a keyboard or controller, as that is often the most efficient way of telling the game what you want to do.

Perhaps needless to say, but all the non-player characters (NPCs) navigate and generally behave in a thoroughly believable way. For example, they will not get stuck running into walls or repeat the same sentence over and over (well, not more than an ordinary human would). This also means that you have interesting adversaries and collaborators to play any game with, without having to resort to waiting for your friends to come online or have to resort to being matched with annoying thirteen year olds.

Within the open world game, there are other games to play, for example by accessing virtual game consoles within the game or proposing to play a game with some NPC. These NPCs are capable of playing the various sub-games at whatever level of proficiency that fits with the game fiction, and they play with human-like playing styles. It is also possible to play the core game at different resolutions, for example as a management game or as a game involving the control of individual body parts, by zooming in or out. Whatever rules, mechanics and content are necessary to play these sub-games or derived games are invented by the game engine on the spot. Any of these games can be lifted out of the main game and played on its own.

The game senses how you feel while playing the game, and figures out which aspects of it you are good at as well as which parts you like (and conversely, which parts you suck at and despise). Based on this, the game constantly adapts itself to be more to your liking, for example by giving you more story, challenges and experiences that you will like in that new city which you reached by driving five hours in a randomly chosen direction. Or perhaps by changing its own rules. It's not just that the game is giving you more of what you already liked and mastered. Rather more sophisticatedly, the game models what you preferred in the past, and creates new content that answers to your evolving skills and preferences as you keep playing.

Although the game you are playing is endless, of infinite resolution and continuously adapts to your changing tastes and capabilities, you might still want to play something else at some point. So why not design and make your own game? Maybe because it's hard and requires lots of work? Sure, it's true that back in 2015 it required hundreds of people working for years to make a high profile game, and a handful of highly skilled professionals to make any notable game at all. But now that it's the future and we have advanced AI, this can be used not only inside of the game but also in the game design and development and process. So you simply switch the game engine to edit mode and start sketching a game idea. A bit of a storyline here, a character there, some mechanics over here and a set piece on top of it. The game engine immediately fills in the missing parts and provides you with a complete, playable game. Some of it is suggestions: if you have sketched an in-game economy but have no money sink, the game engine will suggest one for you, and if you have designed gaps that the player character can not jump over, the game engine will suggest changes to the gaps or to the jump mechanic. You can continue sketching, and the game engine will convert your sketches into details, or jump right in and start modifying the details of the game; whatever you do, the game engine will work with you to flesh out your ideas into a complete game with art, levels and characters. At any time you can jump in and play the game yourself, and you can also watch a number of artificial players play various parts of the game, including players that play like you would have played the game or like your friends (with different tastes and skills) would have played it.

If you ask me, I'd say that this is a rather enticing vision of the future. I'll certainly play a lot of games if this is what games will look like in a decade or so. But will they? Will we have the AI techniques to make all this possible? Well, me and a bunch of other people in the CI/AI in Games research community are certainly working on it. (Whether that means that progress is more or less likely to happen is another question...) My team and I are in some form working on all of the things discussed above, except the natural interaction parts (talking to the game etc).

If you are interested in knowing more about these topics, I recently wrote a blog post summarizing what I've been working on in the last few years. Last year, I also co-wrote a survey paper trying to give a panoramic overview of AI in games research and another survey paper about computational game creativity. Also, our in-progress book about procedural content generation covers many of these topics. You might also want to look at the general video game playing competition (and its results) and the sentient sketchbook and ropossum AI-assisted level design systems. For work on believable NPC behavior, check out the Mario AI Turing Test competition and procedural personas.

Finally, I've always been in favor of better collaboration between AI researchers and game developers. I wrote a post last year about why this collaboration doesn't always work so well, and what could be done about that.

Thursday, July 30, 2015

Revolutionary algorithms

Edit: Apparently it is not clear to everyone that the following post is satire. Well, it is.

You have surely heard about evolutionary algorithms and how they, inspired by Darwinian evolution in the natural world, are excellent general-purpose search and optimization algorithms. You might also know about neural networks, which are learning algorithms inspired by biological brains, and cellular automata, inspired by biological developmental processes. The success of these types of natural computation has spurred other attempts to base algorithms on natural phenomena, such as particle swarm optimization, ant colony routing, honey bees algorithm, and cat swarm optimization. These algorithms are popular not only for their biological inspiration but also for their proven performance on many hard computational problems.

However, in an era where unfettered market forces force bankruptcy upon liberal-democratic countries as a result of bank bailouts dictated by the global financial elite, the neoliberal ideological basis of such algorithms can be called into question. After all, they are based on a model of individual betterment at the expense of the weaker members of population, an all-consuming "creative" destruction process where disenfranchised individuals are ruthlessly discarded. "Survival of the fittest" describes the elimination process by which the invisible hand strangles the weak; "self-organization" is the capitalist excuse for exploiting non-unionized labor. Common evolutionary algorithm operators like survivor selection represent the violence inherent in the system.

But there is an alternative: algorithms for socially just optimization based on models of the workers' struggle and the liberation of the oppressed. While rarely discussed in major  (corporate-sponsored) conferences, revolutionary algorithms have certain similarities with  the better-known evolutionary algorithms. The basic structure of a revolutionary algorithm is as follows (Marks and Leanin 2005):

  1. Initialize the population with n individuals drawn at random.
  2. Evaluate all individuals to assign a fitness value to them; sort the population according to fitness.
  3. Remove the most fit part of the population (the "elite").
  4. Calculate the average fitness in the population; assign this fitness to all individuals.
  5. Increase fitness of the whole population linearly according to a five-generation plan.
  6. Repeat step 5 until maximum fitness has been reached.

As you surely understand, this simple scheme does away with the need for elimination of lower-performing individuals while assuring orderly fitness growth according to plan. Just like in evolutionary computation, a number of modifications to the basic scheme have been proposed, and proponents of the various "schools" that have grown up around specific types of algorithms do not always see eye to eye. Here are some of the most important new operators:

  • Forced population migration (Sztaln 2006): While in evolutionary computation much effort is is spent on diversity maintenance, in revolutionary computation it is important to counteract the damaging effects of diversity. Forced population migration moves whole parts of the population around in memory space, so as to counteract any dangerous clustering of similar individuals.
  • Continuous anti-elitism (Polpotte 2008): While standard revolutionary algorithms only cull the elite in the initialization phase, the radical scheme suggested by Polpotte eliminates the most-fit part of the population every generation. When no fitness differences can be discerned, which individuals to remove can be determined based on arbitrary factors.
  • Great leap mutation (Maocedong 2007): This modification of the basic scheme is particularly useful when the initial population has a very low average fitness. Here, the population is sorted into small "villages" and each village is told to accomplish its development goals on its own, including creating its own search operators.

More recently a newer generation of researchers have questioned some of the basic assumptions underlying revolutionary computation, such as the stable identity of individuals and the boundaries of the population array. The replacement of some parts of the population with others has been decried as a form colonialism. Revolutionary algorithms of the poststructuralist variety therefore eschew strict divisions between individuals and practice adding random variables to instruction pointers and array indexes. This naturally meets resistance from antiquated, orthonormative models of computation and operating systems. In this context, it is important to remember that "segmentation fault" is just a form of norm transgression.

In the end, those algorithms that are most efficient will win; society cannot afford substandard optimization. And in the same way as the success of evolutionary algorithms is predicated on the success of Darwinian evolution in nature, the success of revolutionary algorithms is predicated on the success of the ideologies and movements that they are modeled on.

(This post was inspired by discussions with Daan Wierstra, Mark Nelson, Spyros Samothrakis and probably others.)