Discussion Topic |
|
This thread has been locked |
BASE104
Social climber
An Oil Field
|
|
Dec 25, 2018 - 09:56am PT
|
I really disagree with the approach that many of you use.
Religion and speculative philosophy was important once....in the 15th century.
And to Werner’s position that all life has intelligence, common among Buddhists and Hindu’s, where does the mind reside in a single called organism?
Single cellular critters completely dominated life for 3 billion years. Complex life is a fairly recent occurrence.
I have been studying botany in my spare time. Many plant genomes are much larger than the human genome. Plants and their adaptations are fascinating.
|
|
Ed Hartouni
Trad climber
Livermore, CA
|
|
Dec 25, 2018 - 09:58am PT
|
...not to mention lateral gene transfer...
happy solstice to you all!
|
|
MikeL
Social climber
Southern Arizona
|
|
Dec 27, 2018 - 07:40am PT
|
HFCS: I'm sensing competing minds here...
I’d suggest the sensation is not another’s mind, but your own.
|
|
Don Paul
Social climber
Washington DC
|
|
Dec 27, 2018 - 09:05am PT
|
Religion and speculative philosophy was important once....in the 15th century.
Agreed, and even for thousands of years before, the laws of God became the laws of men over time. Religions weren't a primitive version of science, they were primitive legal and moral systems. I've cited to the Canon Law, mainly for emotional effects, in legal briefs. Islamic Law is even more of a synthesis of law and religion. Islamic scholars prove things with citations, consider precedent and public policy, etc. just like lawyers. The civil and criminal codes of Afghanistan are based on a codification of the Hanafi school of Islam. As recently as about 250 years ago, the dominant legal theory in common law systems (US, UK, etc) was called Natural Law. Instead of making rules and boundaries, judges only discovered them. A similar idea to the Pythagorians who believed mathematics was real, and the world only an approximation. Although I guess you can still say that some things are inherently logical or fair, and when they become settled law, are unlikely to ever change.
|
|
Trump
climber
|
|
Dec 27, 2018 - 09:05am PT
|
Interesting article in the Times about recent developments in chess computing in the latest and greatest AlphaZero chess agent:
https://www.nytimes.com/2018/12/26/science/chess-artificial-intelligence.html
“It clearly displays a breed of intellect that humans have not seen before, and that we will be mulling over for a long time to come.”
“Most unnerving was that AlphaZero seemed to express insight.”
“By discovering the principles of chess on its own, AlphaZero developed a style of play that “reflects the truth” about the game rather than “the priorities and prejudices of programmers,” Mr. Kasparov wrote in a commentary accompanying the Science article.”
Doesn’t sound like it has any opinions on the awesomeness of its own ability to appreciate pretty flowers, but if its evolution follows our human pattern, maybe just give it time. My sense though is that the developers will need to add a little survival bias into the agent, and a little survival of the fittest to its environment, in order to get that result.
|
|
WBraun
climber
|
|
Dec 27, 2018 - 09:27am PT
|
The greatest achievement by any materialist will and has always been defeated by the absolute.
All living entities and everything in the entire material cosmic manifestation is part parcel and subordinate to the Absolute .....
|
|
Trump
climber
|
|
Dec 27, 2018 - 10:43am PT
|
Ok. I don’t doubt that God is a better chess player than this particular machine. I’m just commenting on the smaller question of consciousness, as reflected by this example of a materialistic insight into playing chess and the appreciation that human chess masters have for that materialistic consciousness.
|
|
WBraun
climber
|
|
Dec 27, 2018 - 11:00am PT
|
I know ^^^^ and you are correct with your analysis as far as that goes ....
|
|
MH2
Boulder climber
Andy Cairns
|
|
Dec 28, 2018 - 08:39am PT
|
I think it has long been realized that machines that learn how to play chess might best improve by playing against themselves. I had that thought, too, during my time of interest in how human-made systems might learn.
It has also long been possible to hook together networks of nodes that work in ways similar to how neurons interact with one another. However, there had been a problem with finding an efficient way to correct errors made by the network in a direction that improved its performance. The "deep learning" algorithms are a significant advance in getting a network to use its output to adjust its connections in a way to move toward a target.
Faster processors have also made a big difference.
Chess is a well-studied problem. It is interesting to see what (mostly) unsupervised learning, or self-teaching, can be done by a machine. However, it may tell us more about the nature of chess than about the nature of the human mind.
But chess could also prove to be a good choice for humans learning how to make machines that are more general problem-solvers.
|
|
Ed Hartouni
Trad climber
Livermore, CA
|
|
Dec 28, 2018 - 09:28am PT
|
[Click to View YouTube Video]
Dan King is a British grandmaster and chess "explainer" and coach. He comments on a game AlphaZero plays, in which there are three rather deep moves made.
I found the commentary rather interesting because a human who is an expert on chess (who plays at a high level, and is a coach for other high level players) was taken by surprise with AlphaZero's understanding.
That was evident in Kasparov's comments in Science also, that AlphaZero learned chess and has attained a very high level of understanding in the game.
From a mechanistic point of view, AlphaZero accomplishes this not so much by raw computer power, as Stockfish does, but by architecture, and one that is specifically geared to a particular task, chess. While the architecture was "built by humans" the chess that was learned did not have any human guidance. The insight that AlphaZero gained playing chess is genuinely new, and surprising to human chess players, and could be said to have "emerged" from both the architectures employed and from its experience playing the game.
That AlphaZero cannot explain why it did what it did highlights another issue that pervades this thread, that we cannot explain how we do what we do in many instances, and this is frequently taken as a proof that the mechanistic description of "mind" is a failure. But we have a human chess community composed of people with incredible talent telling us that AlphaZero expresses insight, it does not play like a machine.
Perhaps the "hard problem" is not so hard after all, and is not a "problem" at all, but merely a feature of the particular architecture of the brain from which "mind" emerges.
Interestingly, having figured this out, humans now face a future in which these machines will have profound insights which we may not be able to understand, but will be "true" in a testable manner, and applied to science and technology.
While the report that Einstein said that there were no more than 12 people in the world that would understand his General Theory of Relativity (as proclaimed in a NYTimes article) was not true (Einstein never said this), it seems true, and the insight so profound that most people today don't understand it (Don Paul will surely interject his take on this). Yet most people today in developed nations use the insight everyday for the most mundane tasks, in the form of their location on the planet. The accuracy of that location depends on corrections to the GPS satellite time which your GPS uses to calculate your position, and would be off without correcting for gravitational effects, and not Newton's gravity, but Einstein's.
Perhaps no one will care if we build machines to deliver these insights to us, insights that were once delivered by humans.
|
|
WBraun
climber
|
|
Dec 28, 2018 - 09:50am PT
|
The machine ultimately accomplishes nothing as it has no soul.
The living entity is not and never a soulless machine.
The living entity has actually nothing to do with the material plane itself.
The living entity in the material world is nothing but a fish out of the water and not in its true constitutional natural position.
Thus mankind attempt to AI will never achieve the actual desired result to free it from material bondage for life itself has nothing to do ultimately with the material energies .....
(In the material world, one can play god like children make believing in the sandpit)
|
|
WBraun
climber
|
|
Dec 28, 2018 - 09:58am PT
|
It has no soul.
Soul = non material
All life itself is non-material.
Life always comes from life .....
|
|
Ed Hartouni
Trad climber
Livermore, CA
|
|
Dec 28, 2018 - 10:08am PT
|
Life always comes from life .....
a contemporary perspective
|
|
MikeL
Social climber
Southern Arizona
|
|
Dec 28, 2018 - 10:13am PT
|
Ed: . . . the insight [re: the General Theory of Relativity] [is] so profound that most people today don't understand it . . . most people today in developed nations use the insight everyday for the most mundane tasks, in the form of their location on the planet.
(It’s possible that I may be about to say the same thing as you did there.)
Relying upon a process and its effects (ala, instrumentalism) may not constitute understanding IF “understanding” means complete, accurate, and final knowledge. When I feel around the corner at an entryway for a light switch in the dark, and I find one, I expect a light to go on somewhere in the room. I don’t really have a grasp of electricity, the room’s circuitry, or where the lights are in a room. I just know what to do to see. Things just happen, and I don’t care too much how they happen.
Going from that kind of instrumental understanding to real knowledge about what’s really what has no understanding other than loose associations. I know what I need to know.
Ditto for understanding mind. Your argument about the “hard problem” doesn’t bring light to “what’s really what” anymore than an archaic tribal religion does about the world. What we all follow is, “If I do X, then usually Y happens.”
Most so-called understanding (even scientific) always appears to be some kind of kludge—a highly refined kludge, but a kludge nonetheless. They don’t have any relevance one way or the other as to the extent to which we *really* understand reality.
Our paradigms and visions are generally consistent with one another—so much that we might notice that they invariably go together. That, imo, makes them ideologies, all of them.
|
|
Ed Hartouni
Trad climber
Livermore, CA
|
|
Dec 28, 2018 - 10:35am PT
|
'...may not constitute understanding IF “understanding” means complete, accurate, and final knowledge.'
you've consistently made this point, my criticism is complicated because I agree in the goal of "complete, accurate and final knowledge," I am willing to declare that provisional knowledge constitutes understanding.
Because general relativity is a provisional theory does not mean that the corrections in your GPS are incorrect, it means that they may get even better when we understand the "complete, accurate and final" theory that governs the calculation.
We also understand where our understanding is not "complete, accurate and final" and what the implications of that are, often this provides the place where we revise our provisional understanding, that is the "place" where it breaks, where it does not "work" anymore (by "work" I mean fails to properly predict outcomes, within our finite measurement accuracy and precision).
One can argue that no theory should be believed until it can be demonstrated to be "complete, accurate and final," the instrumentalist approach is that provisional theories match our ability to measure (instrumentally) and that a theory provides useful understanding at the level we can sense (a "sensible" theory).
Your trust in technologies is at least a tacit acknowledgment of the usefulness of provisional theory, in many cases you are willing to trust that limited understanding with your life.
your avatar image shows you wearing glasses. You never responded to my question regarding the reason why... in particular, we say the glasses "correct" your vision, what does that mean to you?
|
|
healyje
Trad climber
Portland, Oregon
|
|
Dec 28, 2018 - 11:28am PT
|
I think the alphZero story is a bit over-hyped. The notion of "insights" has more to do with a machine learning program uncovered lines of reasoning and application humans hadn't; I have no doubt the surprise of novel approaches does feel like "insights", but I'd suggest that's more romantic anthropomorphizing than anything else. And that AlphaZero cannot explain why it did what it did is more a human failing which will likely be addressed quite a bit sooner than later.
|
|
|
SuperTopo on the Web
|