According to "Google AI wins second Go game against top player," when he lost his second game in a series of five against Google's AlphaGo last week, world champion Go player Lee Se-dol said that he had gone from being surprised after losing the first game to being "quite speechless" (2016). This follows up the defeat last year by AlphaGo of the reigning European Go champion, which the same BBC News article describes as "an achievement that was not expected for years." Although he has already lost to Google's artificial intelligence machine, Lee did win the fourth game yesterday as we are told in "Artificial intelligence: Go master Lee Se-dol wins against AlphaGo program," (2016). This intelligence contest between humankind and machine was also reported in Thailand's Bangkok Post's "Game over as computer tops Go master," where one of AlphaGo's creators is quoted as saying that what allows AlphaGo to handle the computationally impossible number of possible moves is something "more akin to imagination" (2016). The Post's report in fact is, as they note, a slightly shortened version of the Agence France-Prese (AFP) article "Human vs machine Go showdown kicks off in Seoul" (2016).
_______________________________________
My Yes/No question is:
Does Google's AlphaGo's defeat of the human Go champion support the activation-synthesis theory of dreams?
My answer is:
Yes, it does. This is another of those times where my published Yes/No question is very different to the first question that came to my mind. Initially I was going to ask: Do you think we can create machines that are truly more intelligent than we are? But then I liked the imagination and intuition mentioned in the Post article, so thought of asking whether a machine could have human-like imagination and intuition. Finally, when I was reading Hartmann's "The Function and Meaning of Dreaming" again last night, I made a connection between the activation-synthesis theory of dreams and how AlphaGo is said to function to achieve its amazing conquests over mere humans.
The connection, I thought is that the activation-synthesis might have deeper implications for how our brains work than merely to explain the generation of our dreams: the major role it gives to purely physical and chemical events that are creators of images and more suggests that those things, the stuff of our thoughts, are all purely mechanical in exactly the same way that a computer is, and that with the rise of such sophisticated machines as AlphaGo, we are finally starting to see that in fact anything a human brain can do or create, so too can a strictly mechanical machine of our creation, except that our machines will only be as intelligent as we are for a very short time. I don't think AlphaGo can dream (yet), but it seems to be producing something like intuition and perhaps even imagination (I would not have suggested this had its creator not said it) by purely electrical and chemical means, just as activation-synthesis says the brain works to create dreams, and perhaps as the brain works to create a lot more.
I'm sorry about using so many different sources. I couldn't find one source that contained exactly the right set of background information that I wanted, so I combined a few, and each needs to be cited.
___________
Reference
Artificial intelligence: Go master Lee Se-dol wins against AlphaGo program. (2016, March 13). BBC News. Retrieved from http://www.bbc.com/news/technology-35797102
Game over as computer tops Go master. (2016, March 12). The Bangkok Post. Retrieved from http://www.bangkokpost.com/news/general/895204/game-over-for-human-go-master
Google AI wins second Go game against top player. (2016, March 10). BBC News. Retrieved from http://www.bbc.com/news/technology-35771705
Hartmann, P. (2007). Quest 2 Reading and Writing (2nd ed.). New York: McGraw-Hill.
Human vs machine Go showdown kicks off in Seoul. (2016, March 9). AFP. Retrieved from http://www.afp.com/en/news/human-vs-machine-go-showdown-kicks-seoul
I do not think AlphaGo's victory is related to activation-synthesis theory of dream. AlphaGo is just a programme created by humans and has nothing special with it. It is made, calculated and planned to outwit the human by choosing the best move. Not only that, I think such programme can predict its opponent move and calculate into thousands possibilities before counter and turn over the game. So, to me, it is not so surprising to see that a programme beats human in terms of board games
ReplyDeleteALso, I do not think there are any intuition in machines or programmes as they are not an organism but are something that work due to functions given by human being. According to Activation-Synthesis Theory, it says, "random electrical energy during REM sleep, possibly due to changes in the production of particular neurotransmitters". Neurotransmitter is surely absent in the programme, thus unable to create the same electrical energy. Thus, I thunk that there is no relationship between AlphaGo and Activation-Synthesis Theory.
I would answer No too. Activation-synthesis theory is associated with random stimulation of chaotic memories by the changes of chemicals in our brain during sleep to produce a logical story line as dreams, or beyond (images or other similar things).
ReplyDeleteIn contrast, it seems that AlphaGo’s artificial intelligence (AI) would not related to random activation like such a theory. This is an ability of machine which systematically functions at its active times like the abovementioned example of Non that the program of this machine can accurately analyze and calculate a great number of possibilities of game.
What I am saying is besides lacking of chemical activities, its AI is programmed for exact calculating and planning for optimal results purposely, not randomly, only during the time when this machine is active, not paused. In turn, activities which might support Activation-synthesis theory should concern with randomness, chemicals, inactive state (sleeping state) and even disordered memory.
Help! Some support!
ReplyDeleteIn the meantime, thank you Non and Feem for presenting clear and well-argued reasons for disagreeing with me.
Go competition between human and Google's AlphaGo is an interesting news for many people in this period.
ReplyDeleteBut, what happens if two machines have to compete Go each other? Does it sound more interesting? What is the result we expect from such completion? Hope to see this soon.
It is still not surprising to see the world champion is defeated by AlphaGo to a score of 1-4. I agree with P'Feem's idea that it will be much more interesting if two machines are to compete Go with one another. It will be a competition between programme developers.
ReplyDelete