Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
argonaut
on March 9, 2016
|
parent
|
context
|
favorite
| on:
AlphaGo beats the world champion Lee Sedol in firs...
Even without the neural net system, AI is able to beat most amateurs, and predict moves experts would make.
Houshalter
on March 10, 2016
[–]
I'm not sure that's correct. MCTS has well known weaknesses, and isn't even a predictive algorithm. MCTS on it's own couldn't get anywhere near beating the top Go champion, that requires deepminds neural networks.
argonaut
on March 11, 2016
|
parent
[–]
http://www.milesbrundage.com/blog-posts/alphago-and-ai-progr...
The best Go program before AlphaGo was CrazyStone, ranked at 5-dan ("high amateur" range).
Houshalter
on March 11, 2016
|
root
|
parent
[–]
There's a massive skill difference between ameatures and professionals. It couldn't even beat the top ameatures.
argonaut
on March 12, 2016
|
root
|
parent
[–]
Which is why I said the intuition was amateur-pro level. I did not say it could beat every amateur-pro in the world.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search:
