The latest news! It was in my newspaper this morning and on dozens of news sites. I found this on the The Telegraph (UK) site:
“Thousands of years of human knowledge has been learned and surpassed by the world’s smartest computer in just 40 days,
a breakthrough hailed as one of the greatest advances ever in artificial intelligence.
Google DeepMind amazed the world last year when its AI programme AlphaGo beat world champion Lee Sedol at Go,
an ancient and complex game of strategy and intuition which many believed could never be cracked by a machine.
AlphaGo was so effective because it had been programmed with millions of moves of past masters, and could predict its own chances of winning, adjusting its game-plan accordingly.
But now the same team has created a machine that learns from scratch. AlphaGo Zero was taught how to play Go,
but given no additional instructions. Instead it learned the best moves over time, simply by playing millions of games against itself. “ -end quote-
For me, as a Go player myself, this is exciting news, but more exciting is that we have here a machine that learns from scratch.
The development of these self learning algorithms in Artificial Intelligence looks like a huge breakthrough.
Is this the beginning of the end? I mean this: in my newspaper of last Saturday they published an award winning essay.
A translation of the headline could be: The human mind unique? Forget it….
Pretty disappointing message, if you take into account, that we are spending our time on discovering how to understand the unique Self we have.
The argumentation of the author of the essay, Gijsbert Werner, is like this:
The question is: which scientific developments forces us to redefine ourselves and our position in the universe?
After every redefinition we have to conclude that we have become less unique than we thought we were in the universe.
1. The invention of the telescope forced us to accept that our universe is heliocentric and that earth is just a tiny spot in an endless universe.
2. Geology as science forced us to accept, that the earth is not as the bible says five or six thousand years old.
The earth is more than 4 billion years old and only a few million years ago man appeared on that earth.
So, we are not the center of the universe nor are we the crown on creation. We are just a rather unobtrusive species on a hardly noteworthy planet.
We have a complex brain, because our thoughts and experiences have a useful evolutionary function.
From an evolutionary point of view our mind is not a goal as such, but our mind is nothing more than a useful set of tools to stay alive in our (social) environment.
Research already has demonstrated that self-awareness is not a unique feature of homo sapiens anymore.
Neuroscience still may be in its infancy, but with every discovery of the actual patterns and mechanisms
that encode specific experiences and thoughts in the brain, it becomes clearer that our mind has a physical basis.
At a very substantial level our thoughts are formed by interacting molecules and neurons.
The more we can manipulate and intervene in these processes, the more undeniable it will be.
Fortunately, there remains a last line of defence for the defenders of the uniqueness of the human mind.
After all, even if we fully understand how thoughts are formed in our brain, notions like self-consciousness may eventually be specific tothe human brain.
With a bit of luck, we may become more and more clear and transparent, but at least still unique in that respect.
With a computer, which accumulated real knowledge in 40 days which took Go players some 3000 years
and with a Self, which moved from the center of creation and universe to a result of physical processes on a hardly noteworthy planet,
I guess the moment is near, that we have to redefine ourselves again and our place in the universe.
Thank you again for your attention….. ^_^
The Discussion
[13:20] herman Bergson: btw..the glasses in the picture behind me are my own RL glasses :-))
[13:20] CB Axel: LOL
[13:20] CB Axel: Nice touch
[13:20] Gemma (gemma.cleanslate): ha
[13:21] herman Bergson: Take your time to redefine yourself today :-)
[13:21] Particle Physicist Bejiita (bejiita.imako): even if a computer can self learn it can still not feel what it is doing, behind this is a toolset that basically says, if true, store in your code, if false throw away and then it repeats this over and over and thus programs itself by only accepting the correct parts
[13:21] Particle Physicist Bejiita (bejiita.imako): here are the tools they use and I myself just started to explore( note a bit of python knowledge is good here)
[13:22] Particle Physicist Bejiita (bejiita.imako): https://www.tensorflow.org/ this is the machine learning toolkot
[13:22] herman Bergson: Yes Bejiita...remember the Chinese room argument of John Searle to begin with
[13:22] Particle Physicist Bejiita (bejiita.imako): https://www.youtube.com/watch?v=Se9ByBnKb0o&list=PLXO45tsB95cJHXaDKpbwr5fC_CCYylw1f
[13:22] Particle Physicist Bejiita (bejiita.imako): and here a tutorial series how to use it
[13:22] Particle Physicist Bejiita (bejiita.imako): this is i think wat go zero uses
[13:22] herman Bergson: a second matter is that computers are good at simulating cognitive process
[13:22] Particle Physicist Bejiita (bejiita.imako): to learn by itself
[13:23] herman Bergson: AI is mainly about cognitive processes....
[13:23] herman Bergson: Why waisting money on simulating feelings by a computer?
[13:24] Gemma (gemma.cleanslate): not a good idea i think too
[13:24] Particle Physicist Bejiita (bejiita.imako): i think it is all about how to make safe self driving cars in the end
[13:24] CB Axel: Maybe computers will take over some day because they don't have feelings to get in the way of thinking.
[13:24] herman Bergson: Take the train Bejiita and you got one :-)
[13:24] Gemma (gemma.cleanslate): bad thought cb
[13:24] Particle Physicist Bejiita (bejiita.imako): hehe
[13:25] Particle Physicist Bejiita (bejiita.imako): but the train never works here in Sweden, you don’t get any value for your money
[13:25] herman Bergson: That indeed is a beloved SF theme CB...
[13:25] Particle Physicist Bejiita (bejiita.imako): maintenance is not what it should be and they always say IT COST TOO MUCH TO FIX, LETS WAIT ¤= YEARS MORE
[13:25] herman Bergson: Terminator 1,2 and 3 are based on it
[13:25] Particle Physicist Bejiita (bejiita.imako): 40
[13:25] Particle Physicist Bejiita (bejiita.imako): thats sad cause trains are smooth in general to ride
[13:25] CB Axel: Vulcans made it work, but their world was destroyed in the end anyway.
[13:26] herman Bergson: Yeah..so was happy Spock survived :-))
[13:26] John Howard Cassio (sticaatsi): ;-)
[13:26] Particle Physicist Bejiita (bejiita.imako): ㋡
[13:26] Gemma (gemma.cleanslate): I think that the future will be very interesting with the changes in genetic knowledge
[13:26] CB Axel: But he was half human. He did have emotions that got in the way occasionally. °͜°
[13:26] Gemma (gemma.cleanslate): to the brain and the mid
[13:27] herman Bergson: But this new achievement of Google DeepMind is significant...
[13:27] Gemma (gemma.cleanslate): and perhaps self
[13:27] herman Bergson: The rules of Go are very simple...
[13:27] herman Bergson: Yes Gemma...
[13:28] herman Bergson: give such a computer the basic rules of whatever...genes....DNA...
[13:28] herman Bergson: and let it reckon
[13:28] Particle Physicist Bejiita (bejiita.imako): it is interesting, i just need to get a grasp of this tensor flow thing, its not entirely easy to understand how it works, basically it operates with 2 or 3 dimensional matrixes of data (tensors)
[13:28] Particle Physicist Bejiita (bejiita.imako): i want to find out how it works, its exciting stuff
[13:28] Gemma (gemma.cleanslate): it has been discovered that ocd is caused by two genes that are lacking something
[13:28] CB Axel: Really?
[13:28] Gemma (gemma.cleanslate): so may be controlled in the future
[13:28] CB Axel: Interesting.
[13:28] Gemma (gemma.cleanslate): yes
[13:28] Particle Physicist Bejiita (bejiita.imako): ok
[13:28] Gemma (gemma.cleanslate): just heard about that the other day
[13:29] Gemma (gemma.cleanslate): so perhaps other mental problems will be changed also
[13:29] herman Bergson: that is what the developers of AlphaGo Zero hope for too Gemma
[13:30] herman Bergson: Yet I think that again we see that the cognitive achievements of machines are somehow equated with the human MIND
[13:30] Particle Physicist Bejiita (bejiita.imako): one very early example of ai (if this is ai now) is a computer who foresaw a totally unknown physics phenomenon by combining all known physics phenomenon, it sorted out a deadly flamethrower like fire disatster in londins subway,
[13:30] Particle Physicist Bejiita (bejiita.imako): the phenomenon is called the trench effect
[13:31] Particle Physicist Bejiita (bejiita.imako): a computer figured it out by itself after being fed data from the disaster
[13:31] Particle Physicist Bejiita (bejiita.imako): they tested it and ut turned out to work just like that
[13:31] herman Bergson: yes Bejiita...that may be the future
[13:31] Particle Physicist Bejiita (bejiita.imako): https://en.wikipedia.org/wiki/Trench_effect
[13:32] herman Bergson: But I wonder...what does it mean to design a machine that has feelings...
[13:32] herman Bergson: ~Like our brain produces
[13:32] Particle Physicist Bejiita (bejiita.imako): it says nothing about a computer figuring it out but that was how it was done
[13:33] Particle Physicist Bejiita (bejiita.imako): hmm indeed, can it be possible even?
[13:33] John Howard Cassio (sticaatsi): Sooner or later human consciousness will be uploaded into computer memories, and we'll be immortal
[13:33] Particle Physicist Bejiita (bejiita.imako): a computer just see on and off, 1 and 0 no matter what it does
[13:33] John Howard Cassio (sticaatsi): I mean, they will
[13:33] Particle Physicist Bejiita (bejiita.imako): we on the other hand are analog
[13:33] Particle Physicist Bejiita (bejiita.imako): or are we, how does the brain process our senses?
[13:34] herman Bergson: Theoretically you may be right John....
[13:34] CB Axel: Our brains are digital, but in a very complex way, aren't they?
[13:34] herman Bergson: But is it possible to create machines for all individual biochemical/neural processes of all individuals?
[13:34] Particle Physicist Bejiita (bejiita.imako): i doubt anything consisting of billions of electrical on and off switches can feel however, 1 and 0 is so far from the analog continuous world
[13:35] John Howard Cassio (sticaatsi): Neurons are like bits I think
[13:35] Gemma (gemma.cleanslate): and who is paying for it all
[13:35] Gemma (gemma.cleanslate) GIGGLES!!
[13:35] Gemma (gemma.cleanslate): ...LOL...
[13:35] CB Axel: A neuron is getting a signal or not.
[13:35] John Howard Cassio (sticaatsi): they are interconnected
[13:35] herman Bergson: GOOD QUESTION Gemma ^_^
[13:35] John Howard Cassio (sticaatsi): and all works on electrical signals, as in a computer
[13:35] Particle Physicist Bejiita (bejiita.imako): you need DACs and ADCs to translate between the analog world and the computer
[13:35] herman Bergson: and what would be the goal of it...immortality??
[13:35] Gemma (gemma.cleanslate): frightning thought
[13:35] herman Bergson: and what when we run out of electicity John???
[13:36] John Howard Cassio (sticaatsi): Even now computers are like prothesis to our brains
[13:36] herman Bergson: instant death to all?
[13:36] John Howard Cassio (sticaatsi): they enhance our memory
[13:36] Gemma (gemma.cleanslate): almost
[13:36] John Howard Cassio (sticaatsi): our devices are memory aids and boosters
[13:36] Particle Physicist Bejiita (bejiita.imako): indeed, computers are awesome machines and they are much faster then we on doing stuff and what they do they repeat over and over without mistake (as long there is no bugs lurking somewhere)
[13:37] herman Bergson: yet not persons....not a Slef...
[13:37] Particle Physicist Bejiita (bejiita.imako): but they are still dumb,
[13:37] herman Bergson: at least not in my computer :-)
[13:37] John Howard Cassio (sticaatsi): when a strong electrical current interferes with the electrical impulses that make our heats go we die
[13:37] Particle Physicist Bejiita (bejiita.imako): but tensorflow and similar tools might be a bit on the way to smarter computers, a must for a traffic safe auto driving car or something similar
[13:38] John Howard Cassio (sticaatsi): people can be programmed like computers and often are
[13:38] Particle Physicist Bejiita (bejiita.imako): you dint want to travel at 100 km/h and suddenly crash into a wall or lorry because the computer missed something
[13:38] John Howard Cassio (sticaatsi): pavlov and pedagogy
[13:38] herman Bergson: Let's conclude that computers are better at cognitive tasks than man....
[13:39] Gemma (gemma.cleanslate): ok
[13:39] Particle Physicist Bejiita (bejiita.imako): ah
[13:39] herman Bergson: but that they still are symbol manipulating machines only
[13:39] Particle Physicist Bejiita (bejiita.imako): how do you make a computer "see" what is a road and what is not for example
[13:39] herman Bergson: The big difference is that WE know what we are doing because we have a Self....computers don't
[13:40] Particle Physicist Bejiita (bejiita.imako): it have no concept of either roads driving or safety
[13:40] herman Bergson: they just do what they do
[13:40] John Howard Cassio (sticaatsi): roads can be installed sensors to guide cars
[13:40] John Howard Cassio (sticaatsi): Computers are evolving fast
[13:40] herman Bergson: Just imagine the computer that says...oh my..what am I doing...should I be doing this?
[13:40] Particle Physicist Bejiita (bejiita.imako): but what it does it repeats with great accuracy, far better and faster then we, it is the interpretation of our analog world , images sounds ect that is difficult for a computer to do
[13:40] Particle Physicist Bejiita (bejiita.imako): as said it sees only 1 and 0
[13:41] John Howard Cassio (sticaatsi): I bet we are computers too in a way
[13:41] John Howard Cassio (sticaatsi): most of what we do responds to an inner program we have
[13:41] herman Bergson: it is the other way around John....
[13:41] Gemma (gemma.cleanslate): self
[13:41] John Howard Cassio (sticaatsi): main objective of the program is reproduction and survival of the species
[13:41] John Howard Cassio (sticaatsi): it is both ways
[13:41] herman Bergson: computers do thing s WE also do but not as good as the machine
[13:41] John Howard Cassio (sticaatsi): we create machines that imitate ourselves
[13:42] John Howard Cassio (sticaatsi): look at a car attentively, you'll see a car has eyes, a nose and an anus
[13:42] Particle Physicist Bejiita (bejiita.imako): in the computers infancy (ENIAC time) they were referred to as gigant electric brains)
[13:42] John Howard Cassio (sticaatsi): and the anus is behind while the eyes are in the front
[13:42] Gemma (gemma.cleanslate) GIGGLES!!
[13:42] Gemma (gemma.cleanslate): ...LOL...
[13:42] herman Bergson: we are not like computers but computers are like us :-)
[13:42] Particle Physicist Bejiita (bejiita.imako): (falls over anime style) THUD!
[13:42] John Howard Cassio (sticaatsi): we make what we know.
[13:42] John Howard Cassio (sticaatsi): computers are going to be integrated into the human body
[13:43] herman Bergson: But computer can make what we do not know....
[13:43] John Howard Cassio (sticaatsi): right now there are workers that are inserted microchips underskin
[13:43] John Howard Cassio (sticaatsi): they are already out there
[13:43] CB Axel: There are also people with computer aided prosthetics.
[13:43] herman Bergson: The AlphaGo Zero computer played Go in a way even the professional never had thought of
[13:43] Gemma (gemma.cleanslate): yes
[13:44] herman Bergson: In opening and end games it came up with totally new strategies...and WON always..
[13:44] Particle Physicist Bejiita (bejiita.imako): just like that other computer program foresaw the trench effect, even way before the time of tensorflow although i think tensorflow is a bit mor advanced, that other computer had all known physics data in it and knew how to combine it
[13:44] herman Bergson: really amazing...
[13:44] John Howard Cassio (sticaatsi): in information processing tasks they must be imbattable
[13:44] Particle Physicist Bejiita (bejiita.imako): it is indeed
[13:45] herman Bergson: Well...to conclude this lecture...my advise..
[13:45] herman Bergson: Buy yourself all a computer,,,,
[13:45] herman Bergson: Quite handy :-)
[13:45] Particle Physicist Bejiita (bejiita.imako): try tensorflow!
[13:46] Particle Physicist Bejiita (bejiita.imako): ㋡
[13:46] herman Bergson: Thank you all again for your participation :-)
[13:46] Gemma (gemma.cleanslate): ♥ Thank Youuuuuuuuuu!! ♥
[13:46] Particle Physicist Bejiita (bejiita.imako): i’ll at last see if i can learn something of it
[13:46] Ciska Riverstone: thank you
[13:46] Particle Physicist Bejiita (bejiita.imako): as said it seem a bit tricky
[13:46] herman Bergson: Class dismissed....^_^
[13:46] Particle Physicist Bejiita (bejiita.imako): nice again ㋡
No comments:
Post a Comment