9 ms·

Neural Annealing: Toward a Neural Theory of Everything

86 points by dangirsh 7y ago

hirundo 7y ago
> being young is like microdosing on LSD all the time If we invent significant life extension, microdosing may be a required part of the protocol just to keep the brain plastic enough to function. It may already be beneficial as a routine geriatric prescription. "Go play with grandma, she's on her weekly trip."
ngcc_hk 7y ago
Another major area of life. Nice to read.
taneq 7y ago
> But we’ve also mostly been doing it wrong, trying to explain the brain using methods that couldn’t possibly generate insight about the things we care about. Wow, it must suck to work in neuroscience and find out that your entire field is pointless because this guy says so.
- loa_in_ 7y ago
  This comment is inflammatory, how is it at the top? Statements saying we're doing it wrong are either revolutionary or misinformed, and I don't see that it's obvious either way
  - taneq 7y ago
    I thought it was a reasonable response to an inflammatory position. If you think you've discovered a new approach that works better than the current ones, you can announce it without dismissing an entire discipline as hopelessly ineffective.
    - xtiansimon 7y ago
      I read the article (once) and didn’t encounter this passage nor much criticism in general. Most of the article outlines in a positive way the annealing idea and it’s related extents. Inflammatory ‘comment’, sure, but hardly a main theme for the article.
  - mannykannot 7y ago
    It will not be obvious for some while whether the author is revolutionary or misinformed, but, historically, revolutions have usually depended on painstaking groundwork of the sort the author disparages in this statement. The discovery of gravity built on painstaking and thorough astrometry, and the discovery of evolution built on painstaking and thorough taxonomy. In neither case could it justifiably be said that these precursors "couldn’t possibly generate insight about the things we care about."
crimsonalucard 7y ago
To hold a concept of certain complexity within the brain the brain itself must have greater complexity than the concept. Thus humans can never understand the brain because the brain has equal complexity to itself and for the brain to understand the brain then it must have greater complexity than itself which is impossible. Like many other things in life, we can only hope to understand a simplified and symbolic representation of brain. The problem is... it's quite possible that human intelligence itself has no higher level abstraction that we are capable of holding in our head. The processes that create consciousness, stripped of all the fluff and irrelevant details, may be of sufficient complexity that our brains will not be able understand what's going on. In short I fear that the ultimate goal of neurology may be a fruitless endeavor.
- schmilk 7y ago
  This sounds eerily similar to some of the early proofs of God. I think there may be a flaw in this line of thinking. |Thus humans can never understand the brain because the brain has equal complexity to itself and for the brain to understand the brain then it must have greater complexity than itself which is impossible. We don't need to know the entire state of the brain at any given moment to understand the mechanics of the brain, just as we don't need to know the entire state of every molecule on planet Earth in order to understand it's physics.
  - crimsonalucard 7y ago
    Sigh. My statement of impossibility was understanding the brain understanding the brain as a whole. This is the only part that is IMPOSSIBLE. Your statement of modularizing portions of understanding into pieces to deal with complexity is isomorphic to my statement that we can only hope to understand A representation of the brain as a simplification. The chemists who understand molecules Understand the atom of the as a symbolic representation. The physicist understands the atom in a more detailed manner and sees the molecule as a symbolic representation. In short each human as individuals CAN ONLY understand the BRAIN as a symbolic simplification. This is the BEST possible outcome. What you are SAYING is the EXACT same thing I am SAYING. The one difference is... I am taking it ONE STEP further with a speculation on the nature of intelligence. There exists concepts in this world that are fundamental and cannot be modularized. It is a very reasonable speculation that the understanding of consciousness itself cannot be further modularized or subdivided. I am saying that in order to understand consciousness it may very well be, that we have to understand consciousness as a complex whole and this complexity may be too big for us to hold in our heads. This IS a realistic possibility. We are already seeing the limits to this with the black box nature of the neural nets we are generating. A possible future simulation of Consciousness may very well be encoded in the same blackbox that our deep fakes are also trapped in. Either way this was a speculation. You (and many others) completely misunderstanding and taking it the wrong way.
ilaksh 7y ago
This is really interesting but I would almost use the word "pseudoscience". I mean a lot of it seems to be on the right track to me in a broad way, but its problematic because it is a mashup of real scientific ideas but the process seems to be more like a philosophical essay. You can't get scientific or engineering progress from philosophy. There are good reasons that most of academia moved on from philosophy.
- Sevrene 7y ago
  You can't get scientific progress from philosophy? Wait until you try justify your belief of knowledge without philosophy. The fact is philosophy underpins much of science. You can't just 'move on' from that when it literally forms the foundations. That being said, you might be right about the article itself. Just as you can have bad science, you can have bad philosophy too.
- normalnorm 7y ago
  > There are good reasons that most of academia moved on from philosophy. There are philosophy departments in almost every reputable University I know... Everything that is non-STEM is underfunded, it doesn't matter if it is philosophy, anthropology, sociology, history, etc. This has nothing to do with academia itself, but with the managerial society we live in, where MBA-types decide on the value of everything with simplistic, dumb and short-sighted economic metrics. Science absolutely needs philosophy. Science is philosophy. There is not a lot of encouragement to do real science/philosophy these days, because this requires deep thinking and following all sorts of unknown paths. Everything must be justified in terms of what gadgets can be built with the discoveries. We are going through a profoundly anti-intellectual stage in western culture.
  - allovernow 7y ago
    >Science is philosophy No. Philosophy does not speak the language of mathematics. Without that tool it's mostly just ego masturbation. I'm not saying it's not interesting, but it's fairly easy for an unusually intelligent person to build cohesive, rational arguments in favor of any arbitrary point. Without some measure of validity philosophers are condemned to continuously argue over the same ideas and assumptions. Philosophy evolved into science when quantifiable validation metrics were introduced. What's more, the hardest and most profound of sciences outside of applied math, physics, simply cannot be advanced without a skilled application of math, without which any amount of essays or books is worthless.
AstralStorm 7y ago
I do not like the lack of actual support in the data for this idea. Not in developmental science, not in neuroscience. Not in how we understand cellular mechanics to work. Neurons and glial cells are born, grow, die and chemically change all the time. Psychedelic therapy is in the infancy. Go and run trials before building sand castles of theory. Most importantly, the testable assumption is what, adding energy would change belief? But of what form and sort? Neurons organize into attractors? Then what do they look like and what's the process? Without that it is woo.
- johnsonmx 7y ago
  Hi AstralStorm, You may want to check out some of the background research to get a sense of what this piece is trying to do-- I can recommend: - Atasoy's YouTube talk on CSHW (as linked and transcribed here: https://qualiacomputing.com/2017/06/18/connectome-specific-harmonic-waves-on-lsd/ https://qualiacomputing.com/2017/06/18/connectome-specific-h... ) - Carhart-Harris and Friston's REBUS model: http://pharmrev.aspetjournals.org/content/71/3/316/tab-article-info http://pharmrev.aspetjournals.org/content/71/3/316/tab-artic... - More recently, Friston's Waves of Prediction: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3000426 https://journals.plos.org/plosbiology/article?id=10.1371/jou... Likewise, a great paper for understanding why a story focused on e.g. glial cells, cell types, and such is unlikely to produce substantial results about e.g. psychedelics: https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1005268 https://journals.plos.org/ploscompbiol/article?id=10.1371/jo... Beyond that, I would suggest the most helpful criticism is specific. If you don't agree with something, good to point out exactly what you disagree with. :)
  - AstralStorm 7y ago
    A definition of how the aggregate would look, and explanation on why lesions have predictable results in every brain, as pure annealing based process would produce amorphous results. Why brains are similar in structure? This model just does not deal well with lesions. After all, a lesion would be even stronger of a change than normal processes or psychedelics, causing the annealing to revert locally while it is healing. We do not observe this "untraining" behavior caused by healed neural lesions as per experiments on quadriplegic rats that get spinal therapy. They need next to no rehabilitation to run again. Likewise, existence of phantom pain phenomenon is counterevidence - and why it just cannot be simply untrained. Yes, I literally poked a hole through this idea. If it actually ever produces working results, then it should be able to fix phantom pain. Likewise, single seizures would cause permanent changes in personality or skill as aftereffects, which is not observed. Whereas repeated ones do, but still not exactly with permanent effects. How does this model explain the difference? The applied "energy" is the same... However epileptics rarely show similar effects, even if they also suffer Grand Mal seizures. Why the difference? Why do mood disorders show up most often in adolescence and say not later? How does this model explain age related dementia? Why only certain anesthetics cause mental problems? (Or sometimes cure, e.g. ketamine.) Etc. The model perhaps works partially in a normally functioning brain, but does not help us understand at all how memory is formed or why is it located mostly in hippocampus. Why cerebellum handles balance and movement, why there is another cortical movement handling? "An attractor forms" begs the question of "why" or "why in this form", which this theory cannot answer. (Especially why these structures are mostly already formed in utero, but still need additional training.) Whereas developmental neuroscience based on proper hormonal basis can. (Up to a point, it's quite new.) The whole thing is a logical (as in high level) explanation based on no biological basis. It lacks the complete causation chain for how psychedelics operate and why every one of them produces different results. Most importantly, an underlying chemical based model would produce identical results as the neurons themselves change permanently in response. (See: neuronal memory allocation mechanisms.) We just do not understand all the neurochemistry related to specific chemicals. Likewise how psychedelics and induced seizures are related. They are, but not entirely and any useful model would explain the difference rather than cover it up. Neuroplasticity is being studied, a simplistic model of "annealing" is not helping in understanding - whereas results relating to say biology of cell given specific inputs are. (E.g. LTP and its relation to CREB and thus other neurochemistry. Synaptogenesis and axonal chemotaxis. More... That get activated during normal and abnormal operation.)
0db532a0 7y ago
Paradigm-shifting ideas. Thanks for posting. Sounds like this could be A New Kind of Science.
- etaioinshrdlu 7y ago
  For those who don't know, A New Kind of Science is when Stephen Wolfram of Mathematica fame wrote something like 1000 pages in layman's terms to convince you of his pet theory that the universe is a cellular automaton. I didn't the hate the book but it was pretty weird and not super scientific.
  - 0db532a0 7y ago
    I regret being so cynical about this post and the book I mentioned. It's a really nice book. Lots of beautiful illustrations and interesting ideas with code snippets for generating those illustrations. I found it fascinating as a kid. The message wasn't too important for me, and I didn't really have the background to draw my own conclusions from it anyway. Someone on IRC, who was a big fan, introduced it to me. I actually started programming with this very book, copying the examples from it into Python and then Scheme (introduced to me by that same guy) and Common Lisp. I respect Wolfram's dedication. It's the type of dedication that brings us things like Mathematica and TempleOS. The guy who introduced me to this book and then Scheme had contacts at Wolfram Research and later sorted me out with a free copy of Mathematica. I could never have afforded that with my pocket money, and I wasn't at university yet. He was a big underdog in the community and had his own problems too. What a great guy. He's still going strong today: http://xahlee.org/ http://xahlee.org/ There's something to be taken even from book like this. It would be good to revisit and re-evaluate the book as someone who now has a semblance of an education in the area. I wish I didn't sell it.
    - 1996 7y ago
      I share your views. TempleOS and Mathematica are monuments to what the human creative spirit can achieve. Let's not be too fast to pass judgement on their authors while having done 1/100000 of what they did
KhoomeiK 7y ago
I like articles like these where you can't tell whether it's about Neuroscience, Artificial Intelligence, Computer Science, or Psychology till you're a significant chunk in.

86 points by dangirsh 7y ago

> being young is like microdosing on LSD all the time If we invent significant life extension, microdosing may be a required part of the protocol just to keep the brain plastic enough to function. It may already be beneficial as a routine geriatric prescription. "Go play with grandma, she's on her weekly trip."

ngcc_hk 7y ago

Another major area of life. Nice to read.

taneq 7y ago

> But we’ve also mostly been doing it wrong, trying to explain the brain using methods that couldn’t possibly generate insight about the things we care about. Wow, it must suck to work in neuroscience and find out that your entire field is pointless because this guy says so.

loa_in_ 7y ago
This comment is inflammatory, how is it at the top? Statements saying we're doing it wrong are either revolutionary or misinformed, and I don't see that it's obvious either way
- taneq 7y ago
  I thought it was a reasonable response to an inflammatory position. If you think you've discovered a new approach that works better than the current ones, you can announce it without dismissing an entire discipline as hopelessly ineffective.
  - xtiansimon 7y ago
    I read the article (once) and didn’t encounter this passage nor much criticism in general. Most of the article outlines in a positive way the annealing idea and it’s related extents. Inflammatory ‘comment’, sure, but hardly a main theme for the article.
- mannykannot 7y ago
  It will not be obvious for some while whether the author is revolutionary or misinformed, but, historically, revolutions have usually depended on painstaking groundwork of the sort the author disparages in this statement. The discovery of gravity built on painstaking and thorough astrometry, and the discovery of evolution built on painstaking and thorough taxonomy. In neither case could it justifiably be said that these precursors "couldn’t possibly generate insight about the things we care about."

crimsonalucard 7y ago

To hold a concept of certain complexity within the brain the brain itself must have greater complexity than the concept. Thus humans can never understand the brain because the brain has equal complexity to itself and for the brain to understand the brain then it must have greater complexity than itself which is impossible. Like many other things in life, we can only hope to understand a simplified and symbolic representation of brain. The problem is... it's quite possible that human intelligence itself has no higher level abstraction that we are capable of holding in our head. The processes that create consciousness, stripped of all the fluff and irrelevant details, may be of sufficient complexity that our brains will not be able understand what's going on. In short I fear that the ultimate goal of neurology may be a fruitless endeavor.

schmilk 7y ago
This sounds eerily similar to some of the early proofs of God. I think there may be a flaw in this line of thinking. |Thus humans can never understand the brain because the brain has equal complexity to itself and for the brain to understand the brain then it must have greater complexity than itself which is impossible. We don't need to know the entire state of the brain at any given moment to understand the mechanics of the brain, just as we don't need to know the entire state of every molecule on planet Earth in order to understand it's physics.
- crimsonalucard 7y ago
  Sigh. My statement of impossibility was understanding the brain understanding the brain as a whole. This is the only part that is IMPOSSIBLE. Your statement of modularizing portions of understanding into pieces to deal with complexity is isomorphic to my statement that we can only hope to understand A representation of the brain as a simplification. The chemists who understand molecules Understand the atom of the as a symbolic representation. The physicist understands the atom in a more detailed manner and sees the molecule as a symbolic representation. In short each human as individuals CAN ONLY understand the BRAIN as a symbolic simplification. This is the BEST possible outcome. What you are SAYING is the EXACT same thing I am SAYING. The one difference is... I am taking it ONE STEP further with a speculation on the nature of intelligence. There exists concepts in this world that are fundamental and cannot be modularized. It is a very reasonable speculation that the understanding of consciousness itself cannot be further modularized or subdivided. I am saying that in order to understand consciousness it may very well be, that we have to understand consciousness as a complex whole and this complexity may be too big for us to hold in our heads. This IS a realistic possibility. We are already seeing the limits to this with the black box nature of the neural nets we are generating. A possible future simulation of Consciousness may very well be encoded in the same blackbox that our deep fakes are also trapped in. Either way this was a speculation. You (and many others) completely misunderstanding and taking it the wrong way.

ilaksh 7y ago

This is really interesting but I would almost use the word "pseudoscience". I mean a lot of it seems to be on the right track to me in a broad way, but its problematic because it is a mashup of real scientific ideas but the process seems to be more like a philosophical essay. You can't get scientific or engineering progress from philosophy. There are good reasons that most of academia moved on from philosophy.

Sevrene 7y ago
You can't get scientific progress from philosophy? Wait until you try justify your belief of knowledge without philosophy. The fact is philosophy underpins much of science. You can't just 'move on' from that when it literally forms the foundations. That being said, you might be right about the article itself. Just as you can have bad science, you can have bad philosophy too.
normalnorm 7y ago
> There are good reasons that most of academia moved on from philosophy. There are philosophy departments in almost every reputable University I know... Everything that is non-STEM is underfunded, it doesn't matter if it is philosophy, anthropology, sociology, history, etc. This has nothing to do with academia itself, but with the managerial society we live in, where MBA-types decide on the value of everything with simplistic, dumb and short-sighted economic metrics. Science absolutely needs philosophy. Science is philosophy. There is not a lot of encouragement to do real science/philosophy these days, because this requires deep thinking and following all sorts of unknown paths. Everything must be justified in terms of what gadgets can be built with the discoveries. We are going through a profoundly anti-intellectual stage in western culture.
- allovernow 7y ago
  >Science is philosophy No. Philosophy does not speak the language of mathematics. Without that tool it's mostly just ego masturbation. I'm not saying it's not interesting, but it's fairly easy for an unusually intelligent person to build cohesive, rational arguments in favor of any arbitrary point. Without some measure of validity philosophers are condemned to continuously argue over the same ideas and assumptions. Philosophy evolved into science when quantifiable validation metrics were introduced. What's more, the hardest and most profound of sciences outside of applied math, physics, simply cannot be advanced without a skilled application of math, without which any amount of essays or books is worthless.

AstralStorm 7y ago

I do not like the lack of actual support in the data for this idea. Not in developmental science, not in neuroscience. Not in how we understand cellular mechanics to work. Neurons and glial cells are born, grow, die and chemically change all the time. Psychedelic therapy is in the infancy. Go and run trials before building sand castles of theory. Most importantly, the testable assumption is what, adding energy would change belief? But of what form and sort? Neurons organize into attractors? Then what do they look like and what's the process? Without that it is woo.

johnsonmx 7y ago
Hi AstralStorm, You may want to check out some of the background research to get a sense of what this piece is trying to do-- I can recommend: - Atasoy's YouTube talk on CSHW (as linked and transcribed here: https://qualiacomputing.com/2017/06/18/connectome-specific-harmonic-waves-on-lsd/ https://qualiacomputing.com/2017/06/18/connectome-specific-h... ) - Carhart-Harris and Friston's REBUS model: http://pharmrev.aspetjournals.org/content/71/3/316/tab-article-info http://pharmrev.aspetjournals.org/content/71/3/316/tab-artic... - More recently, Friston's Waves of Prediction: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3000426 https://journals.plos.org/plosbiology/article?id=10.1371/jou... Likewise, a great paper for understanding why a story focused on e.g. glial cells, cell types, and such is unlikely to produce substantial results about e.g. psychedelics: https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1005268 https://journals.plos.org/ploscompbiol/article?id=10.1371/jo... Beyond that, I would suggest the most helpful criticism is specific. If you don't agree with something, good to point out exactly what you disagree with. :)
- AstralStorm 7y ago
  A definition of how the aggregate would look, and explanation on why lesions have predictable results in every brain, as pure annealing based process would produce amorphous results. Why brains are similar in structure? This model just does not deal well with lesions. After all, a lesion would be even stronger of a change than normal processes or psychedelics, causing the annealing to revert locally while it is healing. We do not observe this "untraining" behavior caused by healed neural lesions as per experiments on quadriplegic rats that get spinal therapy. They need next to no rehabilitation to run again. Likewise, existence of phantom pain phenomenon is counterevidence - and why it just cannot be simply untrained. Yes, I literally poked a hole through this idea. If it actually ever produces working results, then it should be able to fix phantom pain. Likewise, single seizures would cause permanent changes in personality or skill as aftereffects, which is not observed. Whereas repeated ones do, but still not exactly with permanent effects. How does this model explain the difference? The applied "energy" is the same... However epileptics rarely show similar effects, even if they also suffer Grand Mal seizures. Why the difference? Why do mood disorders show up most often in adolescence and say not later? How does this model explain age related dementia? Why only certain anesthetics cause mental problems? (Or sometimes cure, e.g. ketamine.) Etc. The model perhaps works partially in a normally functioning brain, but does not help us understand at all how memory is formed or why is it located mostly in hippocampus. Why cerebellum handles balance and movement, why there is another cortical movement handling? "An attractor forms" begs the question of "why" or "why in this form", which this theory cannot answer. (Especially why these structures are mostly already formed in utero, but still need additional training.) Whereas developmental neuroscience based on proper hormonal basis can. (Up to a point, it's quite new.) The whole thing is a logical (as in high level) explanation based on no biological basis. It lacks the complete causation chain for how psychedelics operate and why every one of them produces different results. Most importantly, an underlying chemical based model would produce identical results as the neurons themselves change permanently in response. (See: neuronal memory allocation mechanisms.) We just do not understand all the neurochemistry related to specific chemicals. Likewise how psychedelics and induced seizures are related. They are, but not entirely and any useful model would explain the difference rather than cover it up. Neuroplasticity is being studied, a simplistic model of "annealing" is not helping in understanding - whereas results relating to say biology of cell given specific inputs are. (E.g. LTP and its relation to CREB and thus other neurochemistry. Synaptogenesis and axonal chemotaxis. More... That get activated during normal and abnormal operation.)

0db532a0 7y ago

Paradigm-shifting ideas. Thanks for posting. Sounds like this could be A New Kind of Science.

etaioinshrdlu 7y ago
For those who don't know, A New Kind of Science is when Stephen Wolfram of Mathematica fame wrote something like 1000 pages in layman's terms to convince you of his pet theory that the universe is a cellular automaton. I didn't the hate the book but it was pretty weird and not super scientific.
- 0db532a0 7y ago
  I regret being so cynical about this post and the book I mentioned. It's a really nice book. Lots of beautiful illustrations and interesting ideas with code snippets for generating those illustrations. I found it fascinating as a kid. The message wasn't too important for me, and I didn't really have the background to draw my own conclusions from it anyway. Someone on IRC, who was a big fan, introduced it to me. I actually started programming with this very book, copying the examples from it into Python and then Scheme (introduced to me by that same guy) and Common Lisp. I respect Wolfram's dedication. It's the type of dedication that brings us things like Mathematica and TempleOS. The guy who introduced me to this book and then Scheme had contacts at Wolfram Research and later sorted me out with a free copy of Mathematica. I could never have afforded that with my pocket money, and I wasn't at university yet. He was a big underdog in the community and had his own problems too. What a great guy. He's still going strong today: http://xahlee.org/ http://xahlee.org/ There's something to be taken even from book like this. It would be good to revisit and re-evaluate the book as someone who now has a semblance of an education in the area. I wish I didn't sell it.
  - 1996 7y ago
    I share your views. TempleOS and Mathematica are monuments to what the human creative spirit can achieve. Let's not be too fast to pass judgement on their authors while having done 1/100000 of what they did

KhoomeiK 7y ago

I like articles like these where you can't tell whether it's about Neuroscience, Artificial Intelligence, Computer Science, or Psychology till you're a significant chunk in.