Really good video! Thanks for posting it

No One Gonna Talk About Alphazero Elmo

On April 20, 2013, GPS Shogi defeated highly rated professional shogi player Miura Hiroyuki in a 102-move game in over 8 hours. A computer is now ready to beat any professional, even Habu. I hope Hidetchi will review the GPS win over Miura.

Actually Akara was not in its best condition. At 8:59, you will see that last three of the list got very little voting point. They are the cluster version of Gekisashi, Bonanza, and YSS. Akara works with many CPUs(169 of Xeon), so to make it work efficiently, developers had to remake their programs. But they couldn't make their new programs stable, and also they couldn't test them sufficiently enough times. As a result, they ended up using only 24 CPUs out of 169.

This is a great video. It seems many times Geki-sashi saw better moves than the other engines. Who do you think would win if Akara played Geki-sashi? Do the engines really try to convince other engines how to vote? Or was that a joke.

Well, Shogi is a different game. Related, but different. Although Chess players seem to have a big advantage over non Chess players in their improvement speed in Shogi at first. A really big one. At least with all the examples I've seen so far. Out of curiosity, what's the age group you're referring to? (I'll deduce the rating myself, if you don't want to write more).

I mean in shogi, of course.

Nice! I like these AI games. I want to study artificial intelligence in college :)

Hmm interesting. Hidetchi just claimed that Gekisashi's moves differed in that they SEEMED more rational, but it doesn't mean at all they were better (reading complex variations and correct evaluation of resulting positions can make a seemingly irrational move better than a "logical" one. I would be interested in the difference in strength between Akara and Gekisashi too. just to see how much Akara gained from the multi-engine settings. With Gekisashi on same hardware as it was here..

How can the AIs decide on how many votes to do per move? This game was awesome!!!!

is there a game which the STRONGEST player got defeated by an AI?

Heh funny. You think your understanding of structure is better than a female pro's? As hidetchi said in the video, it's common to see that maneuver in pro games. From the few pro games I have gone over, I've already encountered this maneuver several times even with bishops exchanged and this open diagonal. However, as hidetchi said, it DOES look strange, agreed. It just seems to be considered ok. And she reached a good position (with the possible move suggested by other pros mentioned later)

no, only a women professional (see Hidetchi's video response)

@IchibanNo01 lol...:)

I think its funny how I'm in the top 100 of my age group in Western Chess and when I look at this im like WTF

Comp is stronger. And alfazero stronger than comp