Ova nova verzija Go-a je upravo takav sustav gdje je algoritam samo neki početni dio - prvo analizira bazu odigranih poteza, a onda sam eksperimentira i vrti hipotetske situacije i smišlja načine za odvagati koji je potez bolji. U teoriji bi se moglo preskočiti učitavanje baze postojećih poteza i krenuti random potezima, ali bi to vjerojatno bilo puno sporije (učenje, a ne odlučivanje nakon naučenog).
Ne u teoriji, nego je to upravo i ucinjeno u praksi i to i jest po mnogima i najzanimljiviji aspekt:
https://deepmind.com/blog/alphago-zero-learning-scratch/
Previous versions of AlphaGo initially trained on thousands of human amateur and professional games to learn how to play Go. AlphaGo Zero skips this step and learns to play simply by playing games against itself, starting from completely random play. In doing so, it quickly surpassed human level of play and defeated the previously published champion-defeating version of AlphaGo by 100 games to 0.
Sto se tice ihushovog vidjenja AI, podsjeca me na AI effect:
"The AI effect occurs when onlookers discount the behavior of an artificial intelligence program by arguing that it is not real intelligence."
Ili citat D. Hofstadtera: "AI is whatever hasn't been done yet".
Ili M. Kearns: "As soon as someone gets a computer to do it, people say: 'That's not what we meant by intelligence.' People subconsciously are trying to preserve for themselves some special role in the universe."
