trial assigning Nibbana and dhamma into AI AlphaGo's architecture.
beginning
Eight months have been passed since I posted here the writing titled 'What are Buddha awakened and how?'. I have been fine everyday in my life.
The reason I refrained from writing and discussing any more is because I wanted to continue them after confirming the 'immediacy of applying dhamma' mentioned in A3:54 through solving the most important problems that remained unfinished in my life. Here is the corresponding passage of 'Paribbājakasutta'. 'It is in this way, too, that the Dhamma is directly visible, immediate, inviting one to come and see, applicable, to be personally experienced by the wise.'
Now, I have experienced the 'immediacy of applying dhamma' through solving serious problems remaining in my life incomplete, so I as ex programmer am writing this article with 'good' intention (sankhara of paticcasammuppada) to introduce 'programmer's pesrpective on dhamma'.
AI(Artificial Intelligence)
Ever since the technical footing of computers in the 1930s and 1940s became solid, attempts have continuously been made to give computers the ability to think like human beings.
Since year 2000, several projects to create virtual brains by using supercomputer have been ongoing at IBM(Dr. Dhamendra Modha's team), Swiss Federal Institute of Technology in Lausanne's laboratory(under professor Henry Markram) and RIKEN(Rikagaku Kenkyusho) research laboratory in Japan.
At the same time, another projects named as IBM's "SyNAPSE Project" and "SpiNNaker Project" of professor Steve Furber of University of Manchester in UK have been executed to develop 'chip set' to replace the expensive supercomputer required for the implementation of the virtual brain and achieved already meaningful results.
However, as far as computing power is concerned, computers with the ability to exceed human beings are still lagging behind humans in the fields of understanding or translating human language or identifying objects(roses, apples, dogs, cats etc.) until recently.
Then, after the publication of a paper of 'deep learning' by Professor Geoffrey Hinton of the University of Toronto, Canada in 2006, it came to have a breakthrough turning point which can get out of the state of stagnation.
Most of the efforts for computers to be simulated human beings before this paper are that programmers develope more sophisticated algorithms(logic) and mounted on the computer. But results were always disappointed.
However, after the paper of professor Geoffrey Hinton, AI researchers let the computer creates its own algorithm for identifying things(roses, apples, dogs, cats etc.) from big data, ie. a large amount of data provided. Actually, this trial opened a new era in which computers can surpass human beings even in areas other than computing.
AlphaGo
The game 'Go' among similar games created by humans is known as the most complex game in recorded history.
The latest version of AlphaGo developed by DeepMind Technologies Limited(Ceo. Demis Hassabis) is a Go appl(application) that learns the Go algorithm by computer itself according to the 'deep learning' theory of professor Geoffrey Hinton, unlike the already existing Go appls in which programmer finds and mounts Go algorithm of the appl.
From early year 2016 to mid year 2017, top ranking Go pros(professionals) from the three countries in the Far East Asia (Korea, China, Japan) challenged AlphaGo, but eventually it ended with AlphaGo's perfect victory. Even though it is a limited field of Go(game), while most people in the world do not know, the era that computers learned by itself exceed professionals(human beings) have already been arrived!
tasks of AlphaGo and dhamma of human
If the expressions ‘tasks from S/W engineering’ and ‘dhamma mirroring from S/W engineering’ appearing in Figure 1 and Figure 2 are unfamiliar at this moment, the article 'Buddha's teaching from programmer's view' might help to alleviate that unfamiliarity.
FIG. 1 is a diagram for visually exposing the difference between classic version AlphaGo computer and AI version AlphaGo computer. In other words, for the classic version, the programmer must perform all three categories of tasks, but in the case of the AI version, the computer performs the tasks in the second category on its own, so the programmer only needs to perform the tasks in the other two categories. In Figure 1, the DeepMind Technologies Limited is expressed as the 'manufacturer' of the tasks of 2nd category.
FIG. 2 is a diagram for visually exposing a difference between AI version AlphaGo computer and a human. In other words, while the AI version AlphaGo computer carries out only the tasks in the second category on its own, humans perform all the tasks in three categories themselves. In Figure 2, the Nibbana is expressed as the 'origin' of the dhamma.
I have already described the relationship between Nibbana and the three categories of dhamma in Chapter 3 of the article 'Buddha's teaching from programmer's view'. Also, 8 months ago, I have already stated as follows (based on SN56:11 Dhammacakkappavattana sutta) that the dhamma eight fold path is self-operated task (to create new appls and to maintain already created appls).
In other words, Gottama did not reach to enlightenment by practiciing the eight fold path, but fully realized that the eight fold path is a sort of task that makes appls!
ending
In the sense that humans have the same three categories of dhamma, all living humans are equal. But every living human beings are not equal in terms of the qualitative difference between my appl and other's appl that are changing at this moment even and changed through countless samsara(rebirth) of life.
Anyway, my current appl is still with a lot of incomplete unfinished conditions, so I will continue to improve my appl until the day of death, and I wish to confirm the improved result in my life remained. fin.
.(2017.9.3)
[이전] [위로] [다음]