Flojomojo #1 Posted June 14, 2017 http://sploid.gizmodo.com/microsofts-ai-just-shattered-the-ms-pac-man-high-score-1796091352 The researchers chose Ms. Pac-Man as it was written to be far less predictable than the original version of the game, which will help the AI they developed better deal with real-world situations that are equally unpredictable. Because as impressive as this feat is, few companies are willing to pay for something that just sits around and plays video games all day. Update, 2:14 p.m. EST/EDT: Microsoft has reached out to us to clarify that the high score for the Atari 2600 version of Ms. Pac-Man is 266,360 points. 5 Quote Share this post Link to post Share on other sites
hepcat #2 Posted June 14, 2017 (edited) Kind of cool, I guess, but I doubt the all time human high score on 4 ghost 2600 Ms. Pac-man is a mere 266K. The game is not that hard; I'm pretty sure I broke 300K on it back in high school. Rolling it would be tough with only 4 lives, but I bet someone has done it before Microsoft's AI. Edit: Yup, I just checked the 2600 high score club results and the high score on Ms. Pac-man is a million... Edited June 14, 2017 by hepcat Quote Share this post Link to post Share on other sites
Flojomojo #3 Posted June 14, 2017 Gizmodo botched the story, was fundamentally confused about the differences between arcade Pac-Man, Atari 2600 versions, and Ms. Pac-Man. Here's a better story https://www.wired.com/story/mircosoft-ai-ms-pac-man/ and here's the source. https://blogs.microsoft.com/next/2017/06/14/divide-conquer-microsoft-researchers-used-ai-master-ms-pac-man/#sm.0001y11attlemfpzrfk1b5haiqep0 If the maximum score is a million, all they can do is hit it. Shattering won't happen. 1 Quote Share this post Link to post Share on other sites
NE146 #4 Posted June 14, 2017 (edited) Lol I don't know if this is the case but it'd be pretty sad if you don't conceptually already understand that arcade ms pac and our beloved VCS port are pretty much entirely different games. It's almost like they grew up in a different era or something!! Regardless though both games get stingy with the scoring once you hit the 300k mark or so.. you can't get any points from ghosts, but also unlike Pac-man you don't get 5000 pt bonuses consistently. Although I don't think the vcs ever gets to a point where the ghosts don't turn blue... do they? Edited June 14, 2017 by NE146 2 Quote Share this post Link to post Share on other sites
+Nathan Strum #5 Posted June 14, 2017 Shouldn't Microsoft be spending their time fixing their own software instead? 12 Quote Share this post Link to post Share on other sites
hepcat #6 Posted June 14, 2017 Lol I don't know if this is the case but it'd be pretty sad if you don't conceptually already understand that arcade ms pac and our beloved VCS port are pretty much entirely different games. It's almost like they grew up in a different era or something!! Regardless though both games get stingy with the scoring once you hit the 300k mark or so.. you can't get any points from ghosts, but also unlike Pac-man you don't get 5000 pt bonuses consistently. Although I don't think the vcs ever gets to a point where the ghosts don't turn blue... do they? If I remember correctly, on the 2600 version you get a one second blue time on every board 8 and higher. It never drops to zero. Also, the 2600 Ms. Pac seems less stingy with bananas on the random fruit levels. Quote Share this post Link to post Share on other sites
negative1 #7 Posted June 15, 2017 i'm impressed. after playing the 2600 version, it definitely has a difficulty curve (not as hard as the arcade though). would like to see more video of the gameplay though. later -1 Quote Share this post Link to post Share on other sites
fiddlepaddle #8 Posted June 15, 2017 Hey! Maybe driverless Ubers will spend their down time playing video games and re-popularize the arcade! 2 Quote Share this post Link to post Share on other sites
+xucaen #9 Posted June 15, 2017 This means it now knows how to evade capture and hunt humans I mean ghosts... o_O Quote Share this post Link to post Share on other sites
Flojomojo #10 Posted June 15, 2017 Shouldn't Microsoft be spending their time fixing their own software instead? Heh. That would imply they're one big happy family. 1 Quote Share this post Link to post Share on other sites
+rdemming #11 Posted June 16, 2017 The complete article on the developer's website (Maluuba) website (follow the link on the bottom of the page to continue reading). Also available is the scientific article and finally the video explaining the algorithm: 2 Quote Share this post Link to post Share on other sites
Andromeda Stardust #12 Posted June 16, 2017 This is really cool, thanks for staring. Perhaps someone invents a "Johny 5" type approach where the bot interfaces the game capturing light and sound from a CRT telivision and servo actuated CX-40 joystick performs the in-game inputs, using the same feedback loops a human would. That would be awesome! Quote Share this post Link to post Share on other sites
negative1 #13 Posted June 19, 2017 if there was more video, it might be possible to see if there are patterns for the 2600 version. given that the ghosts don't go in random directions for the levels (dont remember if they do or dont).. there is a 266k replay here: later -1 Quote Share this post Link to post Share on other sites
tschak909 #14 Posted June 20, 2017 Am going to guess before reading, and say they used a fluid diffusion model (model the maze as a three dimensional model of fluid going through trenches, with certain parts of the maze weighted. Similar to what Norbert Kehrer did when creating an algorithm to automatically place the direction arrows for his automatic mazes...am guessing that a markov model is then built from this fluid model and traversed, as a base-line) -Thom Quote Share this post Link to post Share on other sites
tschak909 #15 Posted June 20, 2017 Ah close, Definitely a parallel Markov traversal like I thought, but with a significant amount of high level domain specific knowledge in each Markov thread, with extreme care to make each thread take care of a specific aspect of a given reward... The algorithm seems to have a high vector factor, and therefore scales quite nicely. The domain specific knowledge describing the reward is the key here, the rest is the grind-work to approach the reward. -Thom (p.s. forget about the fucking articles, they're written by morons, AND ACTUALLY READ THE PAPER!, where the actual data is...it's far more insightful) https://arxiv.org/pdf/1706.04208.pdf 2 Quote Share this post Link to post Share on other sites
+Nathan Strum #16 Posted June 20, 2017 Now let's see them do Jr. Pac-Man. 3 Quote Share this post Link to post Share on other sites
Aquadump Entertainment #17 Posted June 20, 2017 Yeah, but can it solve the riddle of the Sphynx? Quote Share this post Link to post Share on other sites
negative1 #18 Posted June 20, 2017 jr pac man would also be better test, but the thing is, since they've got a solution for ms pac man, it should theoretically also work for all pac-man type games. and most maze games in general. later -1 1 Quote Share this post Link to post Share on other sites
Andromeda Stardust #19 Posted June 21, 2017 jr pac man would also be better test, but the thing is, since they've got a solution for ms pac man, it should theoretically also work for all pac-man type games. and most maze games in general. later -1 I'd like to see any human or AI crack one million on Jr Pacman 2600. The ghosts intelligence and speed are merciless on that game. Even on teddy bear mode (one ghost) I still can't unlock all the mazes... Quote Share this post Link to post Share on other sites
aofl #20 Posted June 25, 2017 I'd like to see them apply similar research to arcade Donkey Kong which IMHO would be much harder to make AI for since the goal isn't grinding through board after board but maximizing points until the kill screen. I wonder if an AI can beat the 1.2 million point human record. A 1 Quote Share this post Link to post Share on other sites