Round 1, Fight! Recently I’ve been trying to show an AI how to create a checkered pattern on a tiled floor for a game. It seems to be learning, but still has quite a bit of trouble around the edges.
The AI uses a generic problem-solving method called a genetic algorithm. Genetic algorithms are really cool because they can produce all kinds of awesome things without any previous knowledge about those things. In this case, it hasn’t the foggiest what it’s doing, but it’s just smashing various maps together and tweaking them trying to get closer and closer to my definition of ‘checkered’ that I gave it.
The problem with genetic algorithms is that they have absolutely no idea what they’re doing. It doesn’t remember what it learned the last time it created a checkered floor (it just starts over again from scratch, tweaking and guessing). It also can’t handle much complexity. “Checkered” is a challenging enough concept, but something like “Ominous” would go right over its head for now.
I came across this article on Polygon describing some of the awesome tech behind the new and improved Storybricks Engine, which purportedly can generate stories from scratch. I was intrigued. The Storybricks team (currently working with the Everquest Next folks), seem to be taking sort of a simulation approach. The rationale is that if every NPC in the world had a complex model of emotions, personality, motivations, goals, desires, and capabilities, then they’d naturally start creating conflict with one another, so quests and stories would (hopefully) write themselves. A player who takes a stroll through the world will have a different experience every time, because the NPCs are constantly fighting and changing relationships.
Apologies for the non-game topic: I stumbled upon an interesting perspective about AI, with a focus on effects to the human economy and lifestyle. Just a disclaimer, I’m a graduate student working on AI, and specifically AI that can replicate human creativity in game design. I’m not an expert, but knowing a bit more about how the sausage is made, I find the video to be overly dramatic and sweeping in its statements. But it’s the comments and discussion about the video that I find the most interesting:
From a non-technical and non-academic vantage point, this type of discussion is frightening, and perhaps even depressing. It evokes strong imagery of either utopian or dystopian societies. I find that odd, because in person it’s extremely difficult to convince someone that computers could ever be more than glorified calculators. Honestly, as I struggle to get my AI to do the most basic of tasks I have a hard time convincing myself! But framed in an economic context, suddenly no one seems to raise a finger of doubt that computers could replace most of what we do.
So, a word of comfort from the perspective of someone who is actually building AI:
This stuff is only just barely getting off the ground. Yes, computers will inevitably become capable of taking over much of what we do, and will probably do it far better than we ever could. But that’s still a really long way off. Right now, “Machine Learning” is essentially just the art of classifying things into neat little piles. It’s really not that smart and requires tons and tons of human intervention and tweaking. From the outside, today’s technology feels like magic. If you’re worried about this stuff taking over your lives, you can take comfort in knowing that we (the magicians) are having a really hard time making progress, and have to claw for every inch. If there’s one thing you can depend on, it’s the glacial pace of academic bureaucracy to stop our robot overlords.
Happy thought for the day,
The Curse of Amnesia is a serious problem in interactive narratives such as pen-and-paper campaigns. If you’re playing an older character who has lived her whole life in a particular city, it doesn’t make sense that you as a player don’t know your way around the city. We want to be able to create any type of character we wish, with any background.
- Amnesia: How do we ensure that the player knows what the character is supposed to know?
- Fourth Wall: How do we prevent the player from knowing what the character isn’t supposed to know, for the purpose of dramatic tension or surprise?
- The character-player relationship in a game is different from the character-reader relationship in a book. In a book we can reveal who the bad guy is to the reader, because the character will still be in the dark, and the reader can only watch helplessly as the character stumbles into a trap. In Interactive Stories, we don’t have that luxury as storytellers.
- Identity Crisis: How do we reconcile the personality differences between how the character would act, and how the player would (or perhaps can) act? Sometimes our acting ability does not do our characters justice, even if we have good intentions.
If you have any thoughts, or even solutions, post them here.
Spectrum Stories are probably the single most popular method in modern games to solve the “Sandbox Paradox”. The basic rundown goes like this: during the entire course of the game, the player will be presented with many (sometimes hundreds) of opportunities to make choices. By far, the most common flavor is “spare bad guy” vs. “kill bad guy”. Sometimes a variant is “just kill everyone” vs. “be stealthy and try not to kill anybody”. Every time you make a choice, it nudges your standing on some greater spectrum.
Today we’re talking about a tried-and-true, predefined solution to the “Sandbox Paradox” in games — Branching Narratives. You may have heard of them in book form as choose-your-own-adventures. The plot of a traditional choose-your-own-adventure story takes the form of a branching tree with many nodes. At each node you can choose one of (typically) two or three options actions that you’d like your character to take next, and you’re taken to the next branch in the story based on what you did.
Writing stories for video games is hard. I mean, could you imagine being an Author and trying to write a novel if your main character was being controlled by someone else? What would you do if your hero just decided not to fight the villain? Even worse, what would you do if they decided to join forces with the villain?
Most game stories are not interactive, and by that, I mean that the player cannot influence the direction of the story. This makes sense, because with the exception of choose-your-own-adventures, it’s impossible to write a complete interactive story beforehand. Games simply don’t have the time or money to write an epic game story for every possible choice that a player could ever make during the course of the game.