A 'Brief' History of Game AI Up to AlphaGo | Hacker News Reader

A 'Brief' History of Game AI Up to AlphaGo

andreykurenkov.com

71 pointsandreyk18 comments

It's a nice historical overview of using search and ML for formal games like chess.

But - this is not the game developer usage of "game AI". Where are wargames, where are artificial characters, where is interactive storytelling, where are realtime control systems, and so on, and so forth? The doc is missing a lot of the actual history of game AI, which is much wider and more varied than presented here (and ML approaches are still in minority).

Anyone interested in game AI history should definitely check out the past proceedings of the AIIDE conference: http://www.aiide.org

What are you talking about. Storytelling is not part of game AI. Not for Chess. At best, the storytelling can be considered to be purposely making the AI consider fewer ply of moves, to create a more fun experience for lower leveled players.

I can't tell if you're trolling or are just not up to date on game AI, so assuming the latter: you have to think beyond chess. There's much more info on interactive storytelling in the AIIDE proceedings I linked to, it's a very active area of research. Academically it goes back at least to the mid 90s, Oz project at CMU for example, or MIT Media Lab projects. And in commercial production, interactive stories are as old as the industry, probably the best early example is Infocom but back then there wasn't much AI driving it; it's not until the 2000s that people started actually building "drama managers" for shipping games to introduce dramatic conflict dynamically.

You can measure research on lots of directions. The purpose of AlphaGo was to win the game, so discussion about AI which are not built to win is largely irrelevant. The difficulty of Go had nothing to do with creating tension (where in many games humans play, AI players can trivially win if the designers wanted them to). In Go this was not the case, as AI lagged behind grandmaster play significantly. Thus, the discussion of research that interests you that has nothing to do with the goals of Go AI, is tangential at best. It's not fair to say the article is faulty just because you wanted them to discuss your favorite research conference.

You are being pedantic - it is pretty clear from the intro I am summarizing the history of AI research into playing games and mean that by 'Game AI'; there is not a single reference to video games, and indeed it would be ridicolous to try and include any amount of summary of that since most video game AI is a lot more heuristic/hacked together/ad hoc that the Game AI stuff I summarized.

I'm glad that you've mentioned Kaissa, the famous chess program from Soviet Union.

Kaissa was developed in 1972. Before that, in 1963, there was ITEP (ИТЭФ) chess program and in 1965 this program won in the international computer chess match. See details here: https://chessprogramming.wikispaces.com/ITEP+Chess+Program#S...

By the way, the first computer game in USSR was made in 1957 on computer Strela-1, the author is Yuri Pervin. It was a variant of Dominoes board game in the player vs computer mode. The game was described in "Problemi Kibernetiki", issue 3, pages 171-180.

Very interesting, thanks for the info!

Term "Game AI" is usually used in context of game development, where it means something completely different: not an agent that is optimized for top perfomance, but an agent that is optimized for maximum player fun. It's a completely different task, solved with completely different methods.

>"since then computers have continued to become exponentially faster, and today humanity’s best Chess players are likely no match for programs you can run on your smartphone."

Is this really true? Would a smartphone chess app beat the worlds top ranked player today?

>Chess engines continue to improve. In 2009, chess engines running on slower hardware have reached the grandmaster level. A mobile phone won a category 6 tournament with a performance rating 2898: chess engine Hiarcs 13 running inside Pocket Fritz 4 on the mobile phone HTC Touch HD won the Copa Mercosur tournament in Buenos Aires, Argentina with 9 wins and 1 draw on August 4–14, 2009.

https://en.wikipedia.org/wiki/Computer_chess#Computers_versu...

oh wow! thanks for the link

So, if someone out there can explain, is a program like AlphaGo deterministic? Meaning if the opponent plays the exact same moves in 2 consecutive games will the 2nd game be a replica of the first?

My naive guess is that the MTC search part of the algorithm will necessarily introduce some variation in play, no?

I would be surprised if AG didn't start with a randomly generated seed to prevent this from happening. Of course that just moves the question to, if you start with the same seed and do the same moves...

yep, I got that tidbit from Kasparov himself I think

Reply on news.ycombinator.com