Game AI Roundup Week #8 2008: 9 Stories, 3 Quotes, 1 Video, 2 Books

Novack on February 23, 2008

Saturdays at are dedicated to rounding up Smart Links from around the web. There’s been lots of news this week due to GDC, including some interesting announcements in San Francisco!

This roundup is brought to you by Novack and Alex Champandard. If you have any news or tips for next week, be sure to email them in to: editors at

Remember there’s a mini-blog over at (RSS) with game AI news from the web as it happens.

Intel Purchases Project Offset

Sam McGrath announced that Project Offset has been acquired by Intel. The company has some interesting game engine technology generally, and artificial intelligence in particular.

“The AI system is designed to reduce work needed by the developer. Rather than forcing the programmer to create a collection of state machines, agents create plans at runtime through goal-oriented action planning. An agent-centric representation of the world is automatically generated rather than having a designer build it from scratch. Visual output for agent states and decisions can be viewed at runtime rather than requiring a programmer to enter the debugger.

  • Agents automatically determine how to sequence their actions, allowing the developer to only concern himself with what each action should do.

  • An agent s assigned goals and actions can be modified at runtime in the editor allowing for fast iteration and experimentation.

  • A navigation graph is generated by the editor to help the agents navigate the world.

  • Agents can track and prioritize multiple threats and suspicions to interact with.

  • At any point in the editor, an agents can be told to draw the states of different internal systems to assist in the debugging process.

  • Sensory system such as vision will display what other entities are being noticed and why.

  • Memory systems such as threat tracking will display which entities an agent has classified as threats and how dangerous he perceives them.

  • Planning systems such as motion path finding will display the current path being followed along with the current state of any paths that are being computed.

Find out more about the technology here:

GDC08: Thwarting the XNA AI challenge

Located in the back of Moscone West’s show floor, was the XNA Game Studio AI Challenge, where contestants (divided into two teams of four each hour) had 50 minutes to program an artificial intelligence script (in C#) that would defeat the other three members on their team in a shoot-em-up multiplayer match (think: four Geometry Wars ships shooting each other for scores). The winner of each round received ‘Invite Only passes to the XNA party.’

Natural Language Conversations in Flash

Long time member of the community Pete King sent in a link to his most recent work going beyond dialog trees:

“Conversation trees are good but don’t really allow you much freedom. In discussing this with a work colleague a while ago, he alerted me to an old Ultima game that would allow you to free type keywords in to the characters. […] So here is the results of about four days work in experimenting with this approach.”

xaitment Launches New AI Modules

xaitment GmbH, one of the leading developers and service providers of artificial intelligence for the games and simulation industries, announced that it has launched five new artificial intelligence (AI) modules ranging from standard to high-level AI functionality. The modules, which position the xaitEngine as one of the most flexible AI solutions on the market, were shown at GDC 2008 in the German Pavilion.

GDC 2008 Quotes

Ray Kurzweil, in his keynote on the The Future and Gaming, had the following to say about games and AI:

“The term ‘game’ is regrettable because it implies it’s not real. Also artificial intelligence (AI) is an unfortunate term since that implies that it’s fake or ersatz intelligence, whereas I see it as a real intelligence.”

Soren Johnson, in his “Cheating Ways” lecture about Civilisation 4, commented on the design process:

“My only interest is in creating the most fun experience possible. I’m always careful, when coding AI, to be designing for the good of the experience and not for the good of the AI itself.”

Peter Molyneux also made more announcements about a game in production Lionhead that say everything and nothing at the same time:

“In the six years since Black & White was released they have been working improving the great things that underpinned that game, and now they are getting pretty close to something that is very significant. It’s a very big thing. A new franchise and as close to something revolutionary as I have ever seen. And we’re close to announcing it.”.

Raydon Corp to use AI.implant

Virtual learning company injects artificial intelligence into its gaming technology initiative, using product from Presagis. By eliminating the need to develop an internal computer-generated entity capability, AI.implant enables the Raydon development team to save resources and focus on creating more compelling dynamic scenarios that effectively engage a video game-savvy generation of trainees.

3D Travelling Salesman using Genetic Algorithms

The travelling salesman problem (TSP) is a problem in discrete or combinatorial optimization. It is a prominent illustration of a class of problems in computational complexity theory which are classified as NP-hard.

“Given a number of cities and the costs of travelling from any city to any other city, what is the least-cost round-trip route that visits each city exactly once and then returns to the starting city?”

AI for NPCs in an Interactive Drama

Thomas Amundsen recently made available a paper about learning behaviors through demonstration:

“…there have not been many attempts to create complete interactive drama systems. I am working with a group that will be the first to undertake this daunting challenge, and one of the most important components in such a system is artificial intelligence (AI) to control the behaviors of non-player characters. […] This paper will describe my attempt to design a system that will allow an expert user to demonstrate behaviors to the system, which the system will use to learn how to behave on its own using case-based reasoning. The end result of this work will not only benefit interactive dramas but may help in the design of game AI for other genres of video games or simulations.”

Thought-controlled computers a reality

“The headset could be used to improve the realism of emotional responses of AI characters in games […] If you laughed or felt happy after killing a character in a game then your virtual buddy could admonish you for being callous”

Autodesk To Acquire AI Developer Kynogon

The Paris based AI middleware vendor has been bought by Autodesk. Here’s what the company had to say:

“The future of videogames is about sophisticated game play, with more believable characters and environments,’ said Petit. ‘Our goal is to simplify videogame development by providing cutting-edge tools to create, animate and integrate 3D assets into game engines. The acquisition of Kynogon is a key component in reaching this goal.”

Artificial Intelligence A Systems Approach

This recently released book on AI contains ideas that are increasingly applied in games:

“Written by M. Tim Jones, and edited by Infinity Science Press.Covers not only AI theory, but modern applications e.g., game programming, machine learning, swarming, artificial immune systems, genetic algorithms, pattern recognition, numerical optimization, data mining, and more.”

Complexity Theory: A Modern Approach

There’s an online draft of a recently released book available for free.

“It is intended as a text for an advanced undergraduate course or introductory graduate course, or as a reference for researchers and students in computer science and allied fields such as mathematics and physics. The plan is to keep the draft on the web, even after the book is published, as a resource for teachers and students from developing countries.”

Particularly of interest is a chapter about decision trees. Sadly, the PDF doesn’t include figures.

Dynamic Difficulty Adjustment

Destructoid has an interesting discussion about gameplay that adapts to the player’s skills:

“Videogames as we know them were built on the idea of challenge. Arcade games were always about (and still are about) getting as far as you can go, mastering the gameplay (or at least the minutiae of the levels) and going until you either ran out of money or you contracted premature arthritis in your wrists.

As games got more complicated the management of challenge turned into something of an art. Too much challenge, and players got frustrated and quit. Too little challenge, and players got bored and quit. So the idea of dynamic difficulty adjustment (DDA) was born.”

Stay tuned next week for more smart links from around the web!

Discussion 1 Comments

RobinB on February 24th, 2008

Hey cool, I'm in the picture of the XNA AI challenge at the GDC (only my beard is visible though)! I didn't win it, was too ambitious and didn't finish my mastermind algorithm in the 50 minutes. The guy next to me, a lead programmer from Insomniac, won the first round. GDC was awesome, looking forward to the next year! The AI talks were not really revealing (maybe apart from that noone uses any fancy academia algorithms, really), but the discussions at the roundtables (Dave, are the notes and audio meant to be made public for the non-GDC crowd?) and the AI dinner were really insightful.

If you'd like to add a comment or question on this page, simply log-in to the site. You can create an account from the sign-up page if necessary... It takes less than a minute!