Weekends at AiGameDev.com are dedicated to rounding up smart links from the web relating to artificial intelligence and game development. This week, the article’s trend topic seems to be open worlds; as always, there are some good articles and blog posts for you to read. Remember, there’s also lots of great content to be found in the forums here! (All you have to do is introduce yourself.) Also don’t forget the Twitter account for random thoughts!
This post is brought to you by Marcos Novacovsky (aka “Novack”) and Alex Champandard. If you have any news or tips for next week, be sure to email them in to editors at AiGameDev.com. Remember there’s a mini-blog over at news.AiGameDev.com (RSS) with game AI news from the web as it happens.
Making Things Worse
On his blog BrainWorks, Ted Vessenes (who’s project is focused on building better bots for Quake 3) published an interesting article about programming different skill levels of AI.
“One of the major challenges of game AI is creating AI that’s plays at an appropriate skill level for a range a players. Most games need AI in easy, medium, and hard settings so that as players become better, they can face more appropriate challenges. There are three typical methods for solving this problem […]”
He then goes discusses giving extra health and damage to the bots, as well as trying to actually make them smarter. Read the comments for our take on the issue.
Scripting AI vs. Scripting Scenes
The blog An Entertaining Grime posted an article analyzing the side effects of scripting scenes in an open world style game like GTA, as opposed to AI driven scenes.
“One interesting side effect of the amount of freedom allotted the player in Grand Theft Auto IV is that whenever the developers try to constrict that freedom it seems absolutely stifling. Take for instance the myriad of chase sequences littered throughout the game; about half of them rely on the game’s artificial intelligence to provide the basics for the chases. These free-form segments are most common when running away from the police but also occur sporadically during the missions-proper. On the other hand are pre-scripted sequences that Rockstar seems to have meticulously designed with the sole purpose of outdoing every car-chase ever placed in a videogame. The problem is that, although these scripted chases are cool, they really interrupt the otherwise phenomenally open game.”
An AI Programmer’s Profile
Computer And VideoGames featured an article about open-world games, in which Ed del Castillo (among others) makes a comment. He believes that improving AI is one of the most important areas for enhancing open-world gaming.
“The industry is talent-starved — we really need great coders and multi-dimensional people, as in order to have great AI, you need a person who can create systems that simulate life. […] That person needs to be someone introspective, philosophical, and a viewer of people — working out how to fake salient features in people.”
Fourth International Conference on Games Research and Development 2008
A call for papers has been issued, and among other topics Game AI and Agents for Games development topics were suggested.
“The Fourth International Conference on Games Research and Development 2008 (CyberGames 2008) will be held during 27-30 October 2008 at the Beijing Normal University, China. CALL FOR PARTICIPATION: To explore the latest developments in the game and interactive entertainment industry.”
Catastrophic Errors in Machine Learning
This isn’t directly related to game AI, but this article by Anand Rajaraman covers the perils of machine learning. It’s the exact same reasons that keep game developers away from this kind of technology still:
“The big surprise is that Google still uses the manually-crafted formula for its search results. They haven’t cut over to the machine learned model yet. Peter suggests two reasons for this. The first is hubris: the human experts who created the algorithm believe they can do better than a machine-learned model. The second reason is more interesting. Google’s search team worries that machine-learned models may be susceptible to catastrophic errors on searches that look very different from the training data. They believe the manually crafted model is less susceptible to such catastrophic errors on unforeseen query types.”
Parkour and AI
In a recent interview with Gamasutra, Senior Producer Tim Bennison commented about the AI in Prototype, in particular concerning the movement through an open-world.
“A whole big part of our game is this locomotion system. It’s this adaptive parkour system. That’s the term I use. What it means is that the AI is involved in predicting the environment, looking ahead, and determines how your character’s going to locomote, including over dynamic objects, not just fixed objects.”
Animation is not traditionally considered part of AI, but for games it’s unavoidable. You’ll find many articles on AiGameDev.com about character animation.
Commander AI for Multiplayer Games
Brian Crecente from Kotaku, wrote an article on the multiplayer mechanics of the upcoming PS3 fps Resistance 2. It seems like the AI will play some sort of commanding role, almost like playing chess with itself.
“Once spawned, the game’s AI looks at the larger battle taking place on the map and issues automatic orders to your and your squad, marking your new objective on the map with a star in a circle. Objectives can include taking a control point, defending an area, even coming to the rescue of another squad under heavy fire or taking out one particular member of an opposing squad.
Resistance 2 automatically matches you up with an enemy squad, making sure to give the two opposing forces the same sorts of orders so they’re always fighting one another. The AI also does it’s best to make sure that when possible your objectives and the objectives of other squads in the larger army aren’t near one another, so team mates will be more inclined to stick together and follow orders.”
While it probably doesn’t take many lines of code to implement this, it’s certainly a very novel application of AI to games.
AI: Beyond Logic
On the blog Anyway Games, Aaron Miller posted an article about personality driven AI characters.
“When depth is a goal, AI should include self-contained variables. A character’s actions should not be determined solely by a “personality” type plus environmental circumstances. Within that personality, there should be a variable range.”
PathEngine Updated to Version 5.16
Through Develop.com, we knew about the PathEngine latest version release.
The PathEngine SDK has been updated to version 5.16. The latest version is largely comprised of optimisations, particularly in the pre-generation of ground meshes from 3D source geometry - with the developers going as far as quoting order of magnitude reductions in processing times for certain environments.
Polyworld: Using Evolution to Design Artificial Intelligence
Another interesting Google Tech Talk on AI. (Which we reported a few months ago, but it’s resurfaced on the social media channels this week.)
“This presentation is about a potential shortcut to artificial intelligence by trading mind-design for world-design using artificial evolution. Evolutionary algorithms are a pump for turning CPU cycles into brain designs. With exponentially increasing CPU cycles while our understanding of intelligence is almost a flat-line, the evolutionary route to AI is a centerpiece of most Kurzweilian singularity scenarios. This talk introduces the Polyworld artificial life simulator as well as results from our ongoing attempt to evolve artificial intelligence and further the Singularity.”
Stay tuned next week for more smart links from around the web!