Torah and Jewish Idea > Torah and Jewish Idea

Ticket to Heaven Daily Dose

<< < (20/20)

Hrvatski Noahid:
Although there are many complexities to Friston’s theories, one of his key observations is that biological systems, and neurological systems like the human brain, seem to violate the fluctuation theorem – a physics term for a tendency towards complexity instead of disorder, as would be suggested by entropy, or the Second Law of Thermodynamics. Precisely how the brain does this, however, is a matter still debated by philosophers, physicists, and neuroscientists.

Much in the same way that Rene Descartes’ cogito ergo sum postulated that the only thing we can be totally sure of is that we think and therefore exist, Friston’s Free Energy Principle begins with the problem that the brain is a separate system from the world – at least to some extent. Put more simply, there is information that exists “out there” in the world, as well as information contained within the brain’s memory, which remain separate from each other. The senses, which provide information about the external world to the brain by way of electrical signals, are the intermediary by which the brain updates its understanding of what is happening “out there”.

Although some lazy philosophers assert that nothing can be known outside of one’s own mind, physicists and mathematicians represent the brain-environment-senses system using a concept called the Markov Blanket. This concept, although considered tautological by some and a sneaky trick by others, is simply a mathematical formulation of the underlying physical realities governing the brain that also reconciles with common sense. The implications of the Markov Blanket, however, suggest that the brain must be involved in some kind of iterative or cyclical process to match its impression of the external world with the information it gets from its senses. (Ticket to Heaven by Zachary R.J. Strong, PDF version, p 107-108)

Hrvatski Noahid:
Although this arrangement is theoretically elegant, some difficulties arise when applying it to real-world systems like brains and computers, as they have limited memory and limited time to process information. Even with an estimated processing power of trillions of calculations per second, the human brain faces constraints on its capacities that only allow it to process a fraction of what is actually happening in an environment. Additionally, Friston’s Free Energy Principle would imply that the brain is making vastly complex comparisons between its memory and the world at every second. While this is technically true, this principle faces similar mathematical constraints in real-world scenarios, as there are only so many calories that can be burned at one time.

The brilliance of the Free Energy Principle, and the reason it has proven to be so exciting for neuroscience and psychology, is because several pre-existing constructs in mathematics and physics exist that simplify these computations to such a degree that the proposed computations could be plausibly handled by a human brain. Put simply, instead of having to make calculations across all possible environmental states, including extremely unlikely ones like unexpected meteor strikes or ghost apparitions, the human brain instead makes a series of guesses or hypotheses, and then calculates the amount of surprise generated by its sensory inputs.

Once a level of surprise is ascertained, the mental model can then be progressively revised to fit the incoming sensory data, or action can be taken to change the world to match the model. This can be represented visually by two converging graphs, which represent probability distributions, which in turn correspond to expected and actual states as expressed by the brain’s electrical signals.

With its emphasis on an iterative progression towards a sophisticated and expansive perspective on the world, the general trajectory of development implied by the Free Energy Principle would imply that the human brain begins at a low level of sophistication and grows in response to the sensory information it acquires. The Free Energy Principle would also imply that the more data someone has access to, the more sophisticated and expansive their hypotheses about the world should be – and the less often they will be surprised. (Ticket to Heaven by Zachary R.J. Strong, PDF version, p 108-109)

Hrvatski Noahid:
So far, this seems to map to our reality, where people with more experience in a field, like expert pilots, are less likely to experience surprise or confusion when performing related tasks. But what of the physics involved in these processes, and the information that exists within the system? How might the brain change over time as a result of sensory data and an iterative truth-finding process?

The key to understanding these matters lies in Friston’s mathematical formulation of the Free Energy Principle, which represents the comparison the brain makes between its generative model and the sensory information it receives.

Although the Free Energy Principle only involves a simple subtraction operation, what the symbols in the equation actually represent is a matter of great complexity. For example, the presence of logarithms is due to Friston’s work being founded on the physics of information, discovered and developed by Claude Shannon and Warren Weaver in the mid-twentieth century.

Mathematically, unlikely events contain more information, while events that are entirely unsurprising provide observers with little to no data. In addition, the concept of information in physics is very similar to the idea of computer binary, with each element of data providing an answer to a question about the world and thus constraining possibilities. How these dynamics manifest within the brain have been a rather complicated matter for many researchers but can be largely explained through a thought experiment that follows a hypothetical human throughout their lifespan.

Consider a newborn baby that is only beginning to acquire information from their environment. At this point in their lifespan, the information contained within the baby’s body and brain is almost entirely of a genetic nature, and the contents of their mind consist solely of the core “programming” that is a consequence of their particular iteration of the human genotype. Mathematically, this corresponds to logarithms of zero across most of the generative model, meaning that from the baby’s perspective, almost everything is entirely possible and most sensory inputs would yield a great deal of surprise.

This would seem to be the case for observable infant behavior, as it is known that babies are easily disturbed, are crying in anguish a great deal of the time for no discernable reason, and can be soothed by predictably rhythmic sensory inputs like rocking or gentle singing. (Ticket to Heaven by Zachary R.J. Strong, PDF version, p 110)

Hrvatski Noahid:
It is also well-known that infant minds are in a state of constant learning, and more neural connections are formed in the early stages of life than at any other time. From the perspective of Friston’s mathematics, the infant’s brain is acquiring information and developing reasonable expectations of the environment based on that information, which would correspond to more information being stored in the brain and used to constrain the generative model.

So far, this makes sense.

Moving from infancy to early childhood, the Free Energy Principle would suggest that the brain in this stage has the beginnings of a generative model but is still reasonably “unconstrained” and lacks a great deal of information about reality. Here, the phenomena of fantasies and imaginary friends in childhood further corroborates Friston’s model, as it is clear from any sustained interaction with a young child that they live in a world of near-unlimited possibilities and extraordinary fantasy.

This was the general trajectory identified by child psychologist Jean Piaget, famous for his observations of the development of logical and “rational” capacities in children as they approached adolescence. Among the findings of Piaget and those that followed him include children’s inabilities to understand logical puzzles until certain stages of development, a movement away from fantasy and whimsy to more coherence in thinking and self, and a steady progression towards full rational capacity by puberty. Piaget’s work can be summarized in four stages, which, as would be expected from the work so far, represent increasing levels of sophistication and a metaphorical “awakening” or “activation” of the mind’s capacities and a loss of propensities towards animism and other whimsies.

Adulthood and old age are typically associated with the acquisition of knowledge, as well as an elusive capacity known generally as wisdom. However, based on the tendency towards complexity dictated by evolutionary dynamics and Friston’s work on generative models and free energy, it should be expected that people in old age would have fairly expansive and well-developed models for what they think should and should not be true.

Also, from a computational perspective, to “rewrite the code” would take an extraordinary amount of time and energy, making it perhaps unattractive. Indeed, this would seem to be the case – things like fluid intelligence, or the ability to solve novel problems, begin to decline noticeably after early middle age. Conversely, crystallized intelligence, or the ability to draw upon what one already knows, is maintained into elderhood for most. (Ticket to Heaven by Zachary R.J. Strong, PDF version, p 110-111)

Navigation

[0] Message Index

[*] Previous page