Twitter | Search | |
Demis Hassabis
When I say 'thought' and 'realisation' I just mean the output of value net. It was around 70% at move 79 and then dived on move 87
Reply Retweet Like More
comex 13 Mar 16
Replying to @demishassabis
You tease! What is it now?
Reply Retweet Like
S e O n ⚾ 13 Mar 16
Replying to @demishassabis
πŸ‘Œ Excellent explanation. Needed that. thank you
Reply Retweet Like
μœ€λ™μ‹ 13 Mar 16
Replying to @demishassabis
this 4th game will be first fail ever?
Reply Retweet Like
Demon 13 Mar 16
Replying to @demishassabis
Reply Retweet Like
Jed Brown 13 Mar 16
Replying to @demishassabis
I'm really curious about how Alphago missed something. Would it learn from it's mistake right away for future games?
Reply Retweet Like
IzzyMandelbaum 13 Mar 16
Replying to @dsyun @demishassabis
It lost twice to the Euro Champ. But it didn't lose the official matches they played.
Reply Retweet Like
S e O n ⚾ 13 Mar 16
Replying to @demishassabis
where is it now?
Reply Retweet Like
Vallen Valiant 13 Mar 16
Replying to @demishassabis
I am just surprised that it takes this long, I would have assumed that AG would have found the error immediately.
Reply Retweet Like
CML PoP 13 Mar 16
Replying to @demishassabis
breathtaking
Reply Retweet Like
Sze Kai Chen πŸ˜’ πŸ€– 13 Mar 16
Replying to @demishassabis
Reply Retweet Like
Kevin CantΓΊ πŸ³οΈβ€πŸŒˆ 13 Mar 16
reportedly they've been training it on some other world class human players (or it is training them)
Reply Retweet Like
Arifin Kusuma 13 Mar 16
Replying to @demishassabis
should have sacrifice black p10 group with black p8.
Reply Retweet Like
Jed Brown 13 Mar 16
Maybe both. :) It's also played a lot of games against itself. I wonder if Lee made a move it wasn't prepared for
Reply Retweet Like
Jianhong Wang 13 Mar 16
Replying to @demishassabis
great game!
Reply Retweet Like
Joseph Abrahamson 13 Mar 16
Replying to @demishassabis
How does it report that the mistake was on 79?
Reply Retweet Like
Rhett Robinson 13 Mar 16
Replying to @demishassabis
Sounds like Lee played outside the training data, and AlphaGo made a poor fit to the trained data.
Reply Retweet Like