Are the SQ and AlphaGo based on similar algorithm?
2 replies
eastpeace
8 years ago #114863
Are the SQ and AlphaGo based on similar algorithm?
Would the SQ4 become the “AlphaGo” in trading markets? 😀
I have no idea about GA or GP, even if I have bought the software for a long time.
Would you please tell me more about GA/GP of SQ ?
geektrader
8 years ago #135748
Would you help out in defining what “AlphaGo” is?
Â
P.S.: Ah I see:Â http://www.cnbc.com/2016/03/08/google-deepminds-alphago-takes-on-go-champion-lee-sedol-in-ai-milestone-in-seoul.html
Mark Fric
8 years ago #135751
they both belong to the category “machine learning”, which means that machine is using some kind of algorithms to “learn” or evaluate  some rules automaticaly, without humans teling it exactly what to do. There are multiple ways how machine learning can be implemented.
Â
AlphaGo uses neural networks and SQ uses genetic programing. It’s not that one is necesarily better than the other, they are simply diferent approaches, each has its own strengths and weaknesses, and each is more suitable for different aplications.
Â
I’ve seen attempts of implementing neural networks in trading, but I don’t know any sucess case, moreover neural networks are really a black box, where you don’t see any actual rules behind its decisions.
Â
Â
There are some other possibilities of machine learning, and what i’m trying to achieve is to make SQ4 flexible enough so that anybody would be able to implement his own machine learning engine that can be added to the program, so also new different approaches can be used.
Mark
StrategyQuant architect
Viewing 2 replies - 1 through 2 (of 2 total)