IA Logo


IA Information
Communication

Dave Mark's books

IA on AI


Archive for the ‘Interviews’ Category

Reddit interviews Peter Norvig

Tuesday, March 2nd, 2010

Reddit interviews Peter Norvig – co-author of the seminal book Artificial Intelligence: A Modern Approach and currently Director of Research (formerly Director of Search Quality) at Google Inc. While not necessarily related directly to game AI, I thought it was interesting anyway.

Basically, the interview is a series of questions that people had submitted ahead of time.
One of the interesting questions was about the difference between “weak AI” and “strong AI” – which involved a definition of terms. He cited the common definition that “strong AI” is human-level problem solving AI. While people aren’t working directly on “strong AI” as such, many people are working on components that may very well lead to it in the future.

He wandered into a description of parallel computing and how it relates to layers of abstractions. I believe this part actually has some relevance to game AI in that we are really only beginning to deal well with splitting out computations into multiple streams. Blackboard architectures and similar ideas really allow for this. (See my commentary on Damián Isla’s interview re blackboards for more.)

2019 per Microsoft

Tuesday, March 3rd, 2009

Ok… it was one thing to listen to Ray Kurzweil’s vision of the next 10 years during his Keynote last year at GDC. It is quite another to see a video from Microsoft of what they think 2019 will look like. Ray suggested that the game developers (and mostly we AI folks) would have a lot to do with the future. I believe that MS would agree with that suggestion.

Interviews with Soren Johnson about Spore

Sunday, July 27th, 2008


If you are a devotee of the Civilization franchise or if you are interested in where the upcoming game Spore is headed, there were recently two interviews with Soren Johnson.

The first was at 1up.com. The second was at Gamasutra. In both, he talks about how his experiences with (and distinct interests in) strategy and simulation AI translated over to the world of Spore.

A-Life, Emergent AI and S.T.A.L.K.E.R

Wednesday, March 5th, 2008

AIGameDev.com has a great, in-depth interview with Dmitriy Iassenev, the AI mastermind for S.T.A.L.K.E.R. The game has a very extensive A-life system that lends a lot of depth to the game. In Dmitriy’s words:

The gist of the A-life is that the characters in the game live their own lives and exist all the time, not only when they are in the player’s field of view. It eventually runs counter to the customary optimization processes used in games development (why perform operations invisible to the player?). Thus, such a scheme is reasonable to be used only when you know exactly what you want to have in the end. We had the game designers’ requirements to have the characters that could not only live inside a certain level, but move between the levels, memorizing the information they obtained during their existence. Consequently, we have decided that each character should come with only one logical essence regardless of the level he is at; whereas we could try to implement that with various tricks involved.

Read more of this very detailed interview over at AIGameDev.com – the place for the killer AI stuff on killer games!

Edit on Bioshock audio interview

Monday, March 3rd, 2008

Some people were complaining that the audio interview with the AI lead on Bioshock was either playing too fast or the applet itself was not loading. I have edited the original page to show a link to the actual audio file. Let me know if that is not working for you.

John Abercrombie (2K Boston) on the AI in Bioshock

Wednesday, February 27th, 2008

At the GDC in San Francisco this past week, I had the opportunity to spend quite a bit of time with John Abercrombie of 2K Boston and the AI lead on Bioshock. It had been a heck of a week for the entire Bioshock team since their title had not only won many awards and accolades prior to GDC but also had a great showing at the Game Developers Choice Awards (they lost the Game of the Year award to Portal… in which there is no shame). Despite having been pestered relentlessly all week to speak about his AI work with the title (he probably talked about it as much as if he had given a lecture on in), he was gracious enough to sit down with me on Sunday morning and chat with me about it on tape.

I want to mention a few things here first…

  • I’m not a professional interviewer. I’m an AI guy just wanting to pick someone’s brain.
  • We were sitting in a hotel lobby so there is a bit of background noise.
  • We didn’t have a white board or anything else so we didn’t really get “down to the metal” on any of these subjects – no code, no diagrams, etc.
  • We could have gone on forever that day, but Laurie and I had to catch a shuttle to the airport.
  • The one requirement his boss put on this interview was that I mention (loudly) that 2K Boston is hiring AI Programmers!

In case the above applet doesn’t load or it is playing at the wrong speed, you can download the actual .mp3 file.

Here’s some questions (paraphrased somewhat) that I asked and topics we discussed throughout the interview (listen to the audio for his answers and descriptions):

(0:41) How did you manage some of the issues with the potentially different factions? That is, splicers fighting splicers, the Big Daddies fighting the splicers, etc.

(5:35) Are the Splicers trying to get at the Little Sisters by default and thereby annoying the Big Daddies or is that just accidental?

(8:00) We discuss the game design mechanic of being able to choose the time and place of your “boss battle” and then actually wander around and watch the boss (Big Daddy) interact – even in combat with the Splicers.

(11:30) John talks about how the Splicers react to being cornered by the Daddies.

(12:50) What other things were you doing with the perception systems?

(19:35) We discuss how the enemies (e.g. the grenadiers) will adapt their tactics to you… both as an intelligent reaction but also to adapt the gameplay to specific difficulties in dealing with the interface and some plasmids.

(21:28) John talks about a small horror story with regard to ballistic physics and the “rifleman’s rule“.

(25:19) John talks about how the Daddies will use things like proximity grenades to box you in – but that how the Splicers don’t even notice them.

(27:23) John talks about how the design decision surrounding whether to have the Splicers and Daddies avoid fire.

(28:13) I ask about the different tactics that the enemies use and how they select between them. Also, how they had to do workarounds to keep players from exploiting things like the pathfinding grid. He goes into detail about how this was beginning to get computationally expensive.

(34:26) What’s your biggest horror story from the whole process? (John responds with an amusing anecdote about some early behavior tests.)

(37:24) What’s the coolest thing that you pulled off?

(38:53) On Soren Johnson’s scale from “good AI” to “fun AI”, where do you think Bioshock fits?

(40:50) Do Splicers continue to spawn on a level until you leave it?

(43:00) We briefly digress into talking about the audio (which won awards).

(44:28) Not counting development time, how many times did you play through the game?

(45:24) The mandatory “2K is hiring” plug!

(49:04) John tells the story about his wonderful moment after the game is shipped. (As game developers, don’t we all want to hear this?)

(Total length 50:56)

If you want to hear more of John’s comments on various subjects, you can listen to the GDC AI Roundtable audio files and read my notes here. He spoke up once or twice in there.

Again, I want to thank John for taking the time with me for the interview… and being a pretty good companion throughout the week of GDC. I enjoyed it quite a bit.

This interview brought to you in association with…

Academic AI vs. Game AI

Saturday, December 15th, 2007

There is an interesting post/rant/exploration by Adam Russell at AI-blog.net entitled Thoughts on Industry / Academic collaboration. He talks about his frustration on how academia seems to not “get it” with regards to game AI.

I made a comment/rant on the post myself, which I will repost here just for the sake of saking.

I think part of the problem that continues to be the fence between the game AI world and academia is the game worlds continued insistence that we have to strip down our AI to “fake AI” in order to wedge it into games.

I’m getting tired of the rubber stamp statements that “our players don’t want realistic behaviors… they want FUN behaviors!” And yet, in review after review of the latest games, people bitch about the AI not being realistic enough. We hear it. We acknowledge it. But when it comes to developing the next cycle, the edict from on high is “we don’t have enough clock cycles to do that nifty XYZ technique.”

As Moore’s Law trips merrily along from year to year, we have more and more processing available to us. In theory, that should give us, as game developers, the overhead we need to close the gap between the need for 60 FPS in our games and the academics who don’t really care if they are rendering their half-ass, low poly bots at 4 FPS.

Another point on this subject… I’m sick of hearing designers – and even AI programmers – make the statement “but it’s not predictable!” about agent-based, emergent AI. Uh… isn’t that the point? Again, look at the reviews and the comments from our customers. “The AI sucks because it is too predictable.” Even the implication via statements such as “you can beat this level by doing XYZ to the AI because…” means that there is a shallowness to our creations. Why? Is it because we are lazy and don’t want to write more complicated code? Is it because we are scared of the unpredictability of non-deterministic models? Is it because our designers would better be served writing static movie screen-plays than game levels? What holds us back?

I’m not saying that academia is the answer. Sometimes it seems that they can get so wrapped up in an esoteric sojourn that they cease to realize that what they are doing is not even remotely relevant. However, some of the concepts and techniques that they take the time to explore (because they don’t have producers and ship-dates) are things that can map over into the game world. And, if we are truly interested in putting realism into our games (which can be fun for the player!), then what academia comes up with should be noted by us. Adapted maybe, but noted nonetheless.

Behavior Trees

Friday, December 14th, 2007

Time for a taste of the Lyon, France Game Developers Conference!

Alex Champandard at AIGameDev.com posted part 1 of a presentation he gave on the use of behavior trees in game AI.

Seriously good stuff!

(note: there may be a problem viewing the videos with IE – they work fine in Firefox.)

Temporal Coherence and Planning

Tuesday, December 11th, 2007

Alex at AIGameDev has a great essay up entitled “Memento, Temporal Coherence and Debugging Planners“. In it, he talks about how planning algorithms have the problem of having their assumptions about the world fall quickly out of scope as the world changes. One solution is to continually replan from scratch – which can become quite expensive to do for numerous agents.

He offers a couple of solutions – and the comments on the post have turned into a rather interesting discussion on the caveats and possibilities. Check it out!

Working towards better MMORPG AI

Friday, December 7th, 2007

According to this blog posting on an MMORPG blog, a development team working on “The Chronicles of Spellborn” has some ideas on how to make mob AI a little more engaging. The post points to an entry in CoS’s development journal. I need to do a bit more reading on the game and their ideas (the article is a little sparse on details) but at least their hearts are in the right place.

One interesting point that fell out of the comments on that journal is that people are actually worried that the behavior will become predictable. That was startling to me since the concept they were proposing actually makes the behavior less predictable. One example cited was that most MMO mobs will attack the player that attacked them ad nauseum (predictable and shallow). The CoS mob will break off that one player that attacked them and go after the weakest player in the raid group. Yes, that’s somewhat predictable, but only because it makes sense.

One of the comments in the journal responded that the idea was “brilliant”. To me, the idea is bloody obvious. And really, it’s not all that hard or computationally expensive to implement. I just don’t see why MMO AI hasn’t advanced beyond where it is right now. *sigh*

Add to Google Reader or Homepage

Latest blog posts:

IA News

IA on AI

Post-Play'em




Content ©2002-2010 by Intrinsic Algorithm L.L.C.

OGDA