Finally finished 360 days later ... very tired but happy I got through it!
In “Build Deeper - Deep Learning Beginner's Guide by Thimira Amaratunga
This book confirms other predictive system results that I have seen, where it has often been found that we human as a species who fancy ourselves as psychics or using other la-di-da methodologies can at best achieve around an 80% accuracy rate, even with good regular practice and tuning. The more accustomed you are toward reaching ever higher accuracy & precision percentile targets the more the distance to the next little increase in goal horizon. Still it does bring into question the abilities of Science and machine systems designing new machine systems, often through excluding what are regarded as unrepeatable subjective methods in favour of repeatable objectiveness. Outliers and other non-obvious patterns & so on are pushing back the boundaries at the edge of our cultural belief systems.
I don't think that any computer scientist would dispute the point that modern AI or machine learning is nowhere near the threshold of 'consciousness' or even 'general intelligence'. But it's not uncommon for words to have a different meaning within a technical field compared to how they are used in everyday communication. In regular English 'chaos' means unpredictable, whereas in mathematics it refers to the tendency of sensitive nonlinear systems to exhibit emergent attraction basins that can potentially be extremely predictable. Those are arguably even antonyms. Another example would be terms 'deterministic/nondeterministic' in Computer Science, which also differ strongly from their meanings in regular English. The point is that if you feel the need to grandstand on these trivialities, you clearly don't understand the fundamentals of the subject matter under discussion.
If you're into Computer Science and Machine Learning in particular, read on.
If I were clever enough, I would write this review as a fugue. This is the formal structure that Hofstadter uses throughoutGödel, Escher, Bach. Whether the whole book is a fugue, I'm not smart enough to tell. But the fugue is used as a metaphor for layers of brain activity, thoughts, superimposed over the “hardware” of the brain, the neurons.
In fact, though I would recommend starting at the beginning of the book, I suppose one might begin anywhere and read through and back again, a'la Finnegan's Wake. No, the book isn't designed this way, but considering that I couldn't discern a solid central idea until page 302 of the book, and that this was only one of several theses in the book, I wouldn't be surprised if it proved possible to begin anywhere.
The idea presented there is “To suggest ways of reconciling the software of mind with the hardware of brain is a main goal of this book.”
The question is, does it succeed?
I would argue that it does not.
And it does not matter.
There are some works, such as Giorgio De Santilliana's Hamlet's Mill or Daniel Schacter's Searching for Memory that are so vast and all-encompassing that it is difficult to pin down one central thesis. These are the kind of works that you might not understand in your lifetime, the thoughts of a genius transposed directly to paper that, unless you are an equally-gifted person or a savant, you cannot hope to fully comprehend. Still, the threads and nuggets of gold that are spread throughout make it worth the time spent in the dark mines of incomprehension, if only to find that one fist-sized chunk of precious metal and appreciate its beauty set against the background of your own ignorance.
As far as I can tell, the book is really about intelligence, both human and artificial. Hofstadter does a lot of preliminary work priming the reader's brain with assumptions taken from theoretical mathematics and computer programming. But don't let that scare you off! I'm no math whiz, but I found most of the logical puzzles at least comprehensible after a few careful reads. Hofstadter also gives the occasional exercise, leaving the reader without an answer to his question. Like all good teachers, Hofstadter understands that the students who work things out on their own are the best prepared students. That doesn't meant that you won't understand many of the book's salient points if you can't successfully answer his questions. You can. But in order to understand the finer points, I suppose one would have to have a pretty good grasp on the answers to those questions.
And it didn't matter.
What did matter, for me, was having a little bit of a background in the idea of nested hierarchies and a smidgen of knowledge in non-linear dynamics (aka “chaos theory”). For the former, I'd recommend Valerie Ahl's seminal Hierarchy Theory: A Vision, Vocabulary, and Epistemology . For the latter, just do what you were going to do anyway and look it up on Wikipedia. I won't tell anyone.
The idea of nested hierarchies is central to the understanding of what makes human intelligence different from machine intelligence. The short story is this: human thought is structured from the ground up according to the basic laws of physics, in particular, electricity, because it is through electricity that neural networks . . . well, network. The issue is that the layers interceding between neural electrical firings and human thought are tangled. They are explainable, or ought to be explainable, by a series of “tangled” layers that lead up to the higher functioning of thought. Again, this is one of the central points of the book.
And this is the point where Hofstadter utterly fails.
And it doesn't matter.
You see, Hofstadter never convincingly shows those transitional layers between neural activity and thought, though he claims they must be there. He claims that it should be possible to create an Artificial Intelligence (AI) that is every bit as human as human intelligence. The problem is, how do you define human intelligence?
Hofstadter presents the problem like this:
Historically, people have been naïve about what qualities, if mechanized, would undeniably constitute intelligence. Sometimes it seems as though each new step towards AI, rather than producing something which everyone agrees is real intelligence, merely reveals what real intelligence is not. If intelligence involves learning, creativity, emotional responses, a sense of beauty, a sense of self, then there is a long road ahead, and it may be that these will only be realized when we have totally duplicated a living brain.
One of the big issues in identifying whether an AI is actually intelligent is the notion of “slipperiness”. The concept here is that human thoughts can deal in a larger possibility space (my words) than machine “intelligence”. Hofstadter quotes from an article in The New Yorker, in which two statements are made that, while possible, would constitute lunacy on the part of anyone who actually believed them. They are:
If Leonardo da Vinci had been born a female the ceiling of the Sistine Chapel might never have been painted.
And if Michelangelo had been Siamese twins, the work would have been completed in half the time.
Then he points out another sentence that was “printed without blushing”:
I think he [Professor Philipp Frank] would have enjoyed both of these books enormously.
Hofstadter comments: “Now poor Professor Frank is dead; and clearly it is nonsense to suggest that someone could read books written after his death. So why wasn't this serious sentence scoffed at? Somehow, in some difficult-to-pin-down sense, the parameters slipped in this sentence do not violate our sense of 'possibility' as much as in the earlier examples.”
This allowable playfulness is something so complex and multi-layered, that an AI would be hard-pressed to correctly parse an “appropriate” reaction.
This is just one case portraying the difficulty inherent in trying to define and understand intelligence and the connection between brain hardware and mind-thought. The book is rife with them. I'm not convinced that Hofstadter was fully convinced that there will ever be a machine so “intelligent” as to completely mirror human thought.
And, one last time, it doesn't matter.
This book has set me to thinking, thinking hard, about what it means to be human. Not merely as an intellectual exercise, but deep in my emotional breadbasket, if you will, I feel human in a way that I can't explain when I think about the difficulty of trying to translate my hopes, fears, love, creativity, wordplay, happiness, sadness, and ambitions into machine language. There has been a lot of talk lately about “singularity,” that moment when machines become self-aware. I'm beginning to think that it will never happen. And I'm fine with that.
Besides, Hofstadter gives an implicit warning when quoting Marvin Minsky, who said:
When intelligent machines are constructed, we should not be surprised to find them as confused and as stubborn as men in their convictions about mind-matter, consciousness, free will, and the like.
In other words, if we do somehow construct true Artificial Intelligence, with the same capacity for thought and feeling as human beings, whose to say the “person” we create isn't going to turn out to be a real douchebag?